Brendan Eich: JavaScript, Firefox, Mozilla, and Brave
Guests: Brendan Eich
Full transcript · Read the full conversation in order, with its original context.
Contents23 chapters
§01Transcript · 00:00–08:37§02Transcript · 08:40–17:06§03Transcript · 17:07–25:07§04Transcript · 25:08–32:39§05Transcript · 32:40–40:20§06Transcript · 40:23–48:18§07Transcript · 48:21–55:33§08Transcript · 55:37–1:02:50§09Transcript · 1:02:53–1:10:45§10Transcript · 1:10:48–1:19:12§11Transcript · 1:19:14–1:27:21§12Transcript · 1:27:22–1:35:40§13Transcript · 1:35:43–1:43:45§14Transcript · 1:43:47–1:51:23§15Transcript · 1:51:26–1:59:28§16Transcript · 1:59:29–2:08:07§17Transcript · 2:08:08–2:15:34§18Transcript · 2:15:36–2:23:05§19Transcript · 2:23:06–2:30:55§20Transcript · 2:30:57–2:38:12§21Transcript · 2:38:13–2:45:39§22Transcript · 2:45:43–2:53:41§23Transcript · 2:53:43–2:59:12§01Transcript · 00:00–08:37
The following is a conversation with Brendan Eich,
creator of the JavaScript programming language,
co-founder of Mozilla, which created the Firefox browser,
and now co-founder and CEO of Brave Software,
which has created the Brave browser.
Each of these are revolutionary technologies.
JavaScript is one of the most widely used
and impactful programming languages in the world.
Firefox pioneered many browser ideas that we love today,
or even take for granted today.
And Brave is looking to revolutionize
not only the browser, but content creation online
and the nature of the internet to make it fundamentally
about respecting people's control over their data.
Quick mention of our sponsors,
the Jordan Harbinger Show,
Sun Basket Meal Delivery Service,
BetterHelp Online Therapy,
and 8sleep Self-Cleaning Mattress.
Click the sponsor links to get a discount
and to support this podcast.
As a side note, let me say that there's a tension
between theory and engineering
that I've been thinking a lot about.
I tweeted something like,
"Good execution is more important than a good idea,
"but one helps the other."
I think the wording of that sucks,
but what I mean is a good idea is a must.
But in my experience, good ideas are in abundance.
Good execution, on the other hand, is rare.
I think some mix of good timing, good idea,
and good execution is essential.
Getting that mix right is tough,
and Brendan somehow, multiple times in his career,
did just that.
I'm starting to believe it's more art than science,
like most interesting things in life.
If you enjoy this thing, subscribe on YouTube,
review on Apple Podcasts, follow on Spotify,
support on Patreon, or connect with me on Twitter,
@lexfriedman.
And now, here's my conversation with Brendan Eich.
When did you first fall in love with programming?
- I didn't program a lot when I was in high school,
but I had a friend who had a Commodore PET.
And after we saw Star Wars, he said,
"Hey, let's make a basic program
"that does the Death Star Trench run."
And it was just simple 2D graphics.
And I didn't know what I was doing,
so I just helped him out on the math and stuff like that.
I was a math and science kid.
I was really into the HP calculators of the early mid '70s.
These were the RPM.
They were really strongly built,
and are at Goldfinger, instead of gold divinely heavy.
There's probably some gold in them, too, gold metalization.
But they were awesome calculators,
and they had all the scientific functions,
so I was really into that.
So I aimed toward physics.
I was a little late for the, I think,
the 20th century golden age,
and I read a lot of science fiction,
so I was like, yeah, it's on the hyperdrives and warp drives.
And physics was not gonna get there quickly,
and I started hacking on computers
while I was studying physics as an undergraduate
at Santa Clara University.
And I dodged the Fortran bullet
because I was in the science department
instead of the engineering department,
where they still did Fortran card decks.
I think they had an auto collator.
But we were using Pascal,
and I got one of the first portable C compilers,
ports to the deck minicomputers we were using.
And I fell in love with programming
just based on procedural abstraction, Pascal,
just what now would be considered old school,
like structured programming from the '70s.
Niklaus Wirth, the creator of Pascal,
was a good writer and a good pedagogue.
He always, at ETH, would do these courses
where it's like build your own computer,
build your own compiler, build your own operating system.
- From scratch. - Yeah, kind of.
And I know some people who are grad students under him
said he would torture the students
with things like this custom email system
that had 25 word limit. (laughs)
And things like that.
- I unfortunately dodged both the Pascal
and the Fortran bullets.
Could you maybe linger on the Pascal?
What kind of programming language was it?
What is it reminiscent of today?
Because it sounds like it may have had
an impact on your own trajectory.
- Yeah, it was in the Algol family.
And Algol was the big successful
language design and compiler project in the '60s.
It had a successor called Algol 68,
which was ambitious but not as successful.
But Pascal was kind of a wordy procedures
and functions language.
It distinguished between functions,
which return a value, and procedures, which don't,
which just compute.
And you could say that whole Algol family went into Ada.
Pascal had a second life thanks to Borland
with Turbo Pascal, which was hugely successful.
I think in large part due to Anders Helsberg,
who then went to Microsoft and did C# and .NET
with his team there and has done really well.
Doing TypeScript, Type JavaScript.
So yeah, there's a lineage here.
But I was also interested in C and Unix
by the time I was an undergrad.
Because people were bringing Unix up
on all sorts of hardware.
I had some friends who were doing
their own wire-wrapped computers, 6820 maybe.
And I was wire-wrapping for my engineering course,
1609 or something simpler, building a computer on a board.
And I wanted to build a more ambitious one
and port Unix to it, but I picked the wrong processor.
I picked the National Semiconductor NS16032,
which was this amazing CISC,
Complex Instruction Set Computer.
And not the reduced instruction set computers
that were just being contemplated into the mid '80s.
- And RISC ultimately won out.
- RISC-I in some ways, it dissolved into,
you have both now.
You have these super scalar architectures
where Intel has kept probably too much backward compatibility
at the instruction level.
But there's a front end that parses that
into these wide internal instructions.
So the very long instruction word research
that was also interesting at the time
kind of became the microarchitecture
inside the backward compatible Intel.
But I picked the National Semichip
and it never got made successfully.
It was full of bugs and I never could have brought it up.
But I went on out of physics after three years
into math computer science.
And like I said, I did it because I saw,
I was being sort of childlike and naive about physics.
And I thought, meanwhile, the valley is go-go for computers.
The Apple II, the PC, the Intel.
8086, 8088 based PC, the IBM.
Gave Microsoft the future for a somewhat fishy deal.
- So it was wide open in the computing space,
but in physics, you were as optimistic about physics as--
- No, I mean, I was one of three brothers
who were all in the same grade.
I have a twin and a younger brother
who skipped second grade and was with us
the whole time after that.
And he went on.
He actually studied under Kip Thorne at Caltech.
But he also didn't, he ended up in software.
(laughing)
- Does it make you sad that theoretical physics,
even with string theory, hasn't really had
any foundational breakthroughs in the latter part
of the 20th century?
- Yeah, in fact, I'd say the problem is theory
over experiments.
I would say we need more Aristotle and less Plato.
Mathematics is not all physical.
There are lots of mathematics that cannot be realized
as far as I know in this world.
So to understand the world, you need to do experiments.
You need to not just dream up inductive theories
that could have lots of alternative theories
competing with them, with no way to decide between them,
except aesthetics, which is not a good guide, in my opinion.
- I don't know if you are friends with
or have a relationship with Elon Musk.
Where's the, in terms of what you would love to see
our society investing in building up,
is it closer to Elon or is it closer to
Feynman and Einstein and those?
- Well, those gentlemen are no longer with us,
and I think that's noticed.
So like I said, the real glory days of physics,
the famous pictures from Germany before the Second War
were just a fantastic assembly of brains,
Schrodinger and Einstein.
Physics, I think, took a wrong turn
that maybe all of, I would say, Western science took
in going for models over reality, right?
You see this in all sorts of fields.
Now, we can build models that are very predictive
and generative and then we build actual devices
§02Transcript · 08:40–17:06
or semiconductors, things like that.
That's good, I'm not dismissing that.
We need good models, we need to experiment
and prove them and test them.
But the problem I've seen in physics,
which you see certainly in economics, the dismal science,
and you see, surprisingly, in other so-called hard sciences,
is models that don't really have to be tested
against reality.
They can instead become policy tools
or they can become, like I said,
one of a large family of alternate theories
that could be as predictive,
but nobody's doing the winnowing out.
- That's such an interesting tension in society.
You see this in even the softer sciences,
which I have a deep love for, like psychology.
You see this in epidemiology, not with the virus.
- Absolutely.
- There's this tension of how much of the world
can we understand through just a beautifully fit model?
And then at the same time, my main work
is in machine learning, where it's like
there is no provable thing usually.
It's just kind of, it's all about just getting
the right data set and getting tricks and so on.
And there's this tension, even in my own soul,
of like I grew up in theoretical computer science.
I loved approximation algorithms,
like all of that, like different complexity classes,
just those little puzzles.
I mean, I don't know, to you as somebody
who was in math and computer science
and then ended up going into places
where you engineered some of the most impactful things
in this world, do you see the P versus NP,
all that whole space as interesting at all?
- Yeah, it's not that useful in practice, right?
People are using it with sort of crypto analysis
or asymptotic arguments about can we have
a quantum-resistant crypto algorithm, things like that,
which may not be practical, right?
If you follow Mikhail Diakonov or Gil Kalai,
there are big questions about how quantum computing
will scale up, how practical it will be.
- Is that something that you think about,
quantum computing?
- Not for a spare time.
Like you said, I'm not using this kind of computer science
in practice because almost everything now is engineering
and finding ways to get computers
to be more useful for people,
which goes from design problems,
which are really kind of an art.
Like Knuth said, anything you can't automate is an art.
Well, we can have machine learning compose music
and it can imitate, you can train it,
and it can sound kind of decent,
but maybe lacking that je ne sais quoi.
But user interface still, I think, requires human art.
- So speaking of things that didn't follow
a perfect theory and model, JavaScript.
So there's two things.
One, it had an impact on the world at a huge scale,
obviously, and it also still is one of,
probably the most popular programming language in the world.
So can we go back to the origin story?
Can you tell the story of how JavaScript was created?
- Yeah, I was at Silicon Graphics after graduate school
for seven years, and it got to be big and successful
and divisionalized and political,
and I thought, kind of boring.
And a friend who'd been there went to
one of the last of the super companies,
super startups in the early '90s.
There were several.
I suppose General Magic was a little after that
or around the same time.
But MicroUnity was that company that I went to,
and it was because my friend, Jeff Weinstein,
had gone there from Silicon Graphics.
He recruited me, and MicroUnity was doing everything.
So this was like the ultimate sort of pretend grad school.
It was doing a new fab, new semiconductor process.
It was doing new analog and digital circuits
on the same very large but not wafer-scale chip.
Originally, it was five centimeters on a side.
It was really hot, too, so it needed a water cooler.
It was a Craykiller, and then they shrunk it,
and they tried to do a home sort of media processor
that was essentially a barrel processor,
but you could think of trying to do all the things
that we now see in modern architectures
with short vector instructions
and sort of wide instructions for multiple issue,
and doing a lot of the stuff in software
because the second iteration, the set-top box,
was really for avoiding the cost to the cable company
of rolling the trucks out
to replace your garbage General Atlantic set-top box
with a totally newer, less garbagey one.
So if you could have software-gradeable set-top boxes,
the cable companies thought they could save a lot of money
and add features.
- Is this assembly, or which level of the software?
- It was like, we were writing in, we were using GCC.
We were writing C++ in C.
Somebody I worked with there, really very smart guy,
hired from a sort of Wall Street
hotshot programming consultancy,
did his own hardware design as well as software.
We were working on how to make
not only short vector units,
but general bit shufflers and permuters
so you could do things like crypto algorithms efficiently,
and you could do demodulation of the cable,
complex quadrature amplitude modulated signal.
So you're basically taking A to D converters,
dumping things in buffers,
and then doing the rest in software.
All the framing and the Reed Solomon
and Viterbi and all that error correction.
So that was really great learning experience,
but it was not gonna work.
It was doing too many risky things at once.
As Jim Clark said to me when I hopped to Netscape
after three years at MicroUnity,
he said, "Oh yeah, you do 10 things,
"each one in 10 odds, it's gonna be one in 10 billion."
(laughing)
The multiplication principle.
So Netscape was already a rocket,
and I passed the chance to go there in 1994.
I knew the founders, 'cause I worked at SGI,
Clark's company.
- Could you pause for a second?
Netscape, when was the launch of this rocket?
- '94.
- '94 was the launch of Netscape?
- And I went there in early '95, in April.
- Okay, so you said you missed the launch.
- Well, I missed the first floor employment opportunity,
but the IPO was August 1995, so I was there for that.
- How obvious was it that Netscape was world changing?
What was the layout?
Was Netscape one of the first big browsers?
- Yes, so when I was at MicroUnity still in '93,
we saw a browser called Mosaic.
And up 'til then, we'd used email,
and we'd used Usenet, the NNTP protocol.
We'd used Newsreaders, we used FTP,
we used all these old internet protocols,
all relying on the DNS and TCP/IP and UDP,
for that matter.
When I was at Silicon Graphics,
we brought up the whole stack, right?
We had to discover how to find the ethernet addresses
on your network, and then find IP addresses for them,
ARP protocol, all that stuff.
And it was great, because nobody knew in the '80s
what was gonna win.
All the proprietary stacks like IBM, SNA, and DECnet,
and all these other protocols were saying,
we're gonna do it, or it's gonna be heterogeneous future.
Instead, it was Berkeley Unix and the TCP/IP stack
that dated back to the ARPANET that won.
And I think we knew it.
We all knew it at SGI, but the salespeople didn't.
And so they kept trying to get multiple network stacks
interoperating, but in the end, it won.
And so that was the internet.
And it was email and texty, and it was used then,
very texty.
And then Tim Berners-Lee did his thing,
but I don't think I was paying attention.
And I think the date when he first did it,
or when he wrote the famous email,
it's been pushed back to '89.
But I noticed a mosaic in '93,
because one of the things that Mark Andreessen
and Eric Bina did at NCSA was they innovated
on the early HTML standard.
They, in particular, Mark sent this email saying,
hey, everybody, we think you should be able
to put an image in a page.
And you know when he sent that,
Eric Bina had already written the code.
And I talked to Tim Berners-Lee more recently,
just a few years ago, and he was like,
oh, we had another way of doing it.
And it didn't work out, 'cause Mark shipped his in mosaic.
And this convinced me of several things.
One, the internet meant there was a huge
first mover advantage, and being fast,
getting on first, mattered a lot.
And so Richard Gabriel of Scheme and Poetry fame
has written about this, the famous essay.
- Poetry, what's poetry?
- Well, he's a poet.
- Oh, actual poetry.
§03Transcript · 17:07–25:07
He's like talking about some kind of subject.
- No, no, I mean, he's the founder of Lucid,
which is where Jamie Zawinski worked before Netscape.
And Lucid was doing compilers, and Lucid Emacs,
which was a fork of Emacs, famously Jamie fighting
against Richard Stallman, Stallmax.
And so Richard Gabriel, very brainy computer guy,
but also a poet, but he wrote a nice essay
that gets abused all the time.
In fact, Jamie's put a kind of warning
in front of his version of it on his site, JWZ.org,
called Worse is Better.
And this is about survival advantage of software
in the network world, in my opinion.
It's about Unix, it started out being framed
as Unix and Lisp, good news, bad news,
because all the Lisp people, the MIT people,
were like, oh, the crown jewel scheme,
this Fabergé egg, or Common Lisp, this giant cathedral.
Of course we're gonna win, this is civilization.
And those farmers in New Jersey, to borrow from Sopranos,
those hicks down at Bell Labs, they're just,
there's nothing sound there, it's all hacking.
- Yeah.
- Well, guess what won?
- Wow, so you're saying this is a fundamental principle
of the internet is moving fast winds.
- You could say in almost any network system,
like in biological evolution, you see successful alleles
sweep populations, and they don't always have,
they aren't free of flaws, they,
heterozygous advantage, right?
You can get, both parents give you the gene variant
and you get sickle cell anemia, right?
But if one of them does, you're more resistant to malaria.
And so this isn't a beautiful process,
except at large scale, and then you realize that
because it moves fast and can adapt, it can win.
And people still struggle with this.
I used to struggle with this, because JavaScript was done
in such a hurry, and the force of web compatibility
meant early mistakes couldn't be fixed.
And even the standards process injected new mistakes,
as it will, but often standards bodies go back
and making compatible changes.
You can't do that with the web.
It's more like, again, like biology,
you preserve what still works.
You don't wanna break ATP metabolism or whatever.
So you have to kind of resign yourself to the reality
of worse is better being enshrined
in actual design points you might not like.
And that happened with JavaScript, and I'm way over it,
but it also, I think, was a huge advantage,
which is why JavaScript has kind of swept
a lot of programming domains.
People will say, "Oh, it's not because of merit."
Well, you're right, but we also improved it over time
in the standards body.
I spent 20 years doing that.
And you don't get that choice.
It's like, I'm not saying that that was the best language.
I'm just saying that was the right time to do it.
And I like to say the alternative was not to do it.
I could have told Netscape, "I can't do this.
"It's too rushed."
And it would have been Visual Basic Script,
and it would have been bad.
- So that's a good way to present the alternative.
But so it was a Netscape, and you have written it
in how many days, and why was it only that many days,
and what was the goal and the underlying principles
in your mind at the time?
- So the whole, I'm sort of describing worse is better
in a frenetic way because it fit the model of Netscape.
When it was known that Jim Clark and Mark Inveas
were founding Netscape, and they did the first release
in 1994, that browser took over from Mosaic.
In fact, that's why Mozilla is called that.
It's the Mosaic killer.
It's like the giant monster that kills Mosaic.
- That's awesome.
- And they knew that, it wasn't that, again,
it's not like you're doing advanced scientific research
that is changing the world.
You're more like taking down the last iteration
of the browser Mark did, which had images,
and other importances before he stopped working on it,
and you're making Netscape the new thing that has images,
plugins, which was the way to do video back in the day.
It had something that's kind of died now
for tiled windows called frames and framesets.
- Oh, yeah, yeah, yeah, yeah.
- HTML tables, that was new.
Eric Bina did tables in Netscape 1.1.
So when I got there, they were heading toward IPO.
Clark wanted to IPO early.
I think his instinct was right,
and that kicked off the whole dot-com era.
There was a recession in the US in '91.
You can see old Law & Order reruns
where they talk about the recession,
how hard it's hitting New Yorkers.
And after that, Greenspan really goosed things
at the Federal Reserve, and technology had been
sort of fermenting in a way that came together
with the internet, and Netscape made it possible
to do pets.com, to do eBay, to get people to recognize
a URL on a billboard and then type it in when they get home.
And that was huge.
That was so fast moving a rocket that Mark
and the engineering team there thought,
we need to make this a programmable browser,
not just a document viewer, not just a video.
- So it was all HTML with images and tables,
and also, like you said, frames.
- Early plugins.
- There's no dynamic element at all.
- Yeah, the most dynamism you get was from a plugin,
which there were a few of them then.
Flash didn't exist at that point.
It was, I think--
- Java Applets yet, or no?
- Well, that's the thing.
We did the deal with Sun.
In fact, I was recruited to go do Scheme in the browser.
Remember Guy Steele and Gerald Sussman's
beautiful Lisp variant?
I was gonna do it in the browser
because my friends from SGI thought,
hey, we like Scheme, you like Scheme.
And I'm like, I hardly ever use Scheme.
It's not really used in industry,
except in sort of silos.
But I like it.
Okay, I'll come do Scheme in the browser.
I have a slide from my 2017 talk
where I have Bruce Willis crawling through the duct
in Die Hard.
He's like, come out to the coast, have a lot of fun.
Come on, do Scheme in the browser.
But when I got there, there was no Scheme in the browser
because they'd started a deal with Sun Microsystems.
And my best contact there was Bill Joy,
who I admired as a Berkeley Unix founder and Sun founder.
And Bill got the idea of making the browser programmable too.
And so the main idea was to put the Java VM,
which at that point was not really easy to embed,
into Netscape, including the Netscape version on Windows
that was still most popular,
which was the 16-bit Windows 3.1, which was going away.
Microsoft was coming out with Windows 95
and everyone was afraid they were gonna do
Internet Explorer, I guess, two at that point,
three the next year.
They already bought or invested in somehow Spyglass,
this other company that shot out from NCSA
at University of Illinois.
And in fact, Microsoft had tried to buy Netscape
in late '94, before I got there.
And I heard about this later.
I heard they offered way too little money.
And so Jim Barksdale and Jim Clark said,
"Get out of here," in a pound of sand.
But then they realized, oh, this is going to hurt us
because now they're gonna copy us.
Didn't happen right away.
I'm not sure when Gates' Internet Tidal Wave memo
was written.
That's the famous memo he wrote when Bill Gates realized
that Microsoft was going down this old copy AOL path
or copy CompuServe path, a project called Blackbird,
presumably after the SR-71, I don't know.
But they were gonna make a dial-up service
with a custom content language stack and custom rendering.
It wasn't the web.
They could have content partners.
They have a lot of money.
But it still wasn't to scale the web.
It wasn't gonna be compelling.
And Gates realized this, and he turned the company
on a dime, and they couldn't buy Netscape.
Again, I'm not sure of the timing.
So they decided to copy it.
And once we realized that, everybody inside Netscape
felt even more urgency and more of a frenetic mood.
And so my chance to do Scheme disappeared
when the Java deal started brewing.
But there was still a chance to do a companion language
to Java because Java is a compiled language.
It's evolved and improved quite a lot since then too.
But it was for serious advanced programmers
that cost a certain salary or hourly rate.
And people observed, Bill Joy observed,
and Mark Andreessen and I observed,
that in a mature stack like Microsoft,
you really benefit from having a scripting language
like Visual Basic, which became Visual Basic Script
in IE3, but didn't take over and kill JavaScript,
that you need two languages.
§04Transcript · 25:08–32:39
One is for the component writers
who are higher priced and more expert.
And the other is for scripters,
certified public accountants, designers,
graphic designers with some programming inclination.
Anybody, amateurs, doesn't matter.
There's a much more demotic approach there
for programming the components together,
gluing them together.
Some people say duct tape language, which I don't really like.
But we saw, Bill Joy and Mark Andreessen and I,
we saw the need for a companion language.
And the gleam in our eye was to call it JavaScript.
I didn't like it, that was marketing's plan.
Mark called it Mocha, which I liked.
And Netscape Marketing, I think, didn't like that.
So they said, "Oh, there's some trademark
"in some software somewhere that uses Mocha,
"so we can't use that."
And they tried LiveScript in August and that didn't last.
And then finally, we got the trademark license
in December, 1995.
But the work I did to prove that it could be done
was important because I came in in April
and even then Netscape was growing so fast
that they couldn't find an open hiring requisition
in the client team for me.
So they hired me into the server team.
And I worked for a month on server team
on what became HTTP/1/1.
So I was actually, I had done protocol work
at Silicon Graphics with Greg Chesson,
former Bell Labs intern, grad student intern,
who knew all the Unix founders.
And Greg was very interested in taking protocols
to the next level with VLSI
'cause he thought that CPUs wouldn't scale up.
He was mistaken in that, unfortunately.
Moore's law more than kept up.
And you have gigabit ethernet running
with conventional processors.
But I worked on protocols at SGI
as well as Unix kernel hacking and NFS and things like that.
So I came into Netscape to work on the server side
for a month, but I was thinking the whole time,
what should this language be like?
Should it be easy to use?
Might it syntax even be more like natural language
like HyperTalk, which is Bill Atkinson's language
in HyperCard, if you have ever used HyperCard
on an early Mac.
And I thought, well, I'd like to do that,
but my management is saying make it look like Java,
which looks like C from a distance.
- What does that mean?
Is it braces we're talking about visually?
Does it mean like, what, management,
do they understand what--
- They don't, marketing didn't know,
but management did, like the rickshaw of engineering knew.
And we had a plan even that was,
if you have this companion language,
you're going to glue things together
between Java and JavaScript.
So you're going to have commerce in memory,
in the heap with data types.
So you're going to want some of the data types in Java
to reflect in the JavaScript.
You're going to want the primitive types that Java,
unfortunately, severed from objects.
So at least some of them, double, let's call it
in Java's terms, from the C term for double precision,
floating point, or strings, or Booleans, and objects.
And so right away, there was this constraint
that looking like Java meant kind of the C curly brace
syntax, but also some of the data types and objects.
- Like objects and so on, all that kind of stuff.
- Dotted columns.
- Comparison operator.
- Garbage collection, all that stuff.
Even the bitwise operators and the shift operators,
including the unsigned right shift,
which Java had because it didn't have unsigned integer types.
It said if you want to do unsigned operations,
use an operator.
And that turned out to be important much later.
I'll tell that story five times.
But JavaScript inherited a set of,
operators, the expression grammar,
the statement grammar up to a point from Java.
But I wanted a functional language.
I wanted a little bit of Scheme,
even though it wasn't as clean as Scheme.
- So you had a love, sorry to interrupt,
you had a love for Scheme and Lisp,
that functional language landscape.
- Yes, I wanted first class functions.
Because I saw the need for callbacks in the browser,
where it's a single threaded program.
All the early browsers were single threaded.
And it's the right model for users.
Most users weren't ready for mutual exclusion
and threading.
So in a single threaded world,
you cannot block the user interface.
So you have to use a callback and run later.
And without getting too fancy
and trying to capture the continuation,
like call CC does in Scheme,
I thought I'll just make it easy to have fun arcs,
first class functions you pass downward
and it can be called back.
And Java didn't have that at the time.
It took forever to get proper first class functions,
lambdas now into Java, Java seven or eight, I think.
- It did have concurrency, right?
- Yes.
- From the very beginning.
But you were thinking that the JavaScript in the browser
would not have the luxury of being concurrent.
- That's right.
And the reason was,
Java was gonna run in the plugins.
We could fork threads and go to town.
But the main action in the browser
was in the single threaded program,
in the single Unix process on Unix or Windows.
And it was where you had to service the event loop
and then go do things.
Respond to the network, lay out some HTML,
render it, turn widths into heights
by filling containers, boxes,
the early, what became the CSS box model.
And run scripts to make the thing livelier,
respond to user input.
And all that event driven programming
was in part like HyperCard
because HyperCard had this on event name syntax.
And so that's why you have in JavaScript on click
run together as the name of the event handler.
And there's some funny ones on mouse over and on mouse out.
People still complain about those.
But there are many more events now
over the years standardized.
But it was a mix of event driven,
single threaded programming
because it had to run in the main thread of the browser
where the action is.
And Java never got there,
which meant Java could not interact easily
or quickly or in a nested way with the document,
with the objects reflected from the HTML document,
with the tables and forms and so on.
And that is one of the reasons I think JavaScript survived
and Java kind of died.
Java was in this plugin prison.
It essentially was confined to a rectangle,
the app with rectangle.
And while we even built a next year,
Nick Thompson, a friend from SGI,
who was an intern grad student at CMU at the time,
built the first version of Live Connect
to glue Java and JavaScript together
to deliver on that vision where you do have commerce
between the data types in the heap.
- Did it work?
- It worked, but Java was in charge.
JavaScript was in charge
and Java was just these components, these helper.
- You might as well do everything in JavaScript.
- What happened over time.
It's like an evolutionary filter.
It just kind of, who needs the plugin?
And in fact, Sun mismanaged Java as a plugin.
They thought, oh, Netscape was giving us
the distribution vehicle and we don't care
about the browser, it's just about getting Java out there.
And that was a big miscalculation.
They then tried, because Microsoft's killing Netscape
after a few years, they tried getting into Microsoft.
And you may remember there was a Sun-Microsoft deal
which famously blew up.
And Microsoft kicked Java out of Windows.
And that's when they really pulled the trigger.
I think they already evaluated it and liked it
on Anders Helsberg's .NET and C#
and decided we're gonna just not have Java.
We don't want any of that Sun stuff.
We don't want the patent risk.
We don't want, I'm not sure what the fights were about.
There was some patent angle to it, I think.
And up till then, Microsoft had been using Java components
like in Outlook Web Access, which had a lot of JavaScript
to be a Hotmail-like user interface.
They had to call the mail server through HTTP
and they used a Java object to do this.
And when they gave the boot to Sun,
they suddenly other, the left-hand gave the boot
and the right-hand said, "We better do something else
in Outlook Web Access.
§05Transcript · 32:40–40:20
What are we gonna do?"
And they said, "Let's just add an ActiveX component,"
which is their own native way
of embedding things in languages.
And we'll make it, it'll be what became XML HTTP request,
which is now a web standard for calling asynchronously.
And it's been replaced by the Fetch API
in HTML5 or HTML living document.
But this whole lineage goes back
to Java being successfully the loser and getting kicked out.
And after Microsoft kicked it out, it was a plugin
and you would find it required for like smart card banking
in the Nordic countries where that was mandated by law
but really didn't get used much.
Or there were pilots who used it for flight information.
But Flash, which Netscape could have bought
but unfortunately didn't.
- Unfortunately didn't?
- Yeah, we would have screwed it up.
- What year are we talking about with Flash?
- I think after the IPO, so it was probably late '95.
- Flash was around, was it Adobe?
No, it wasn't.
- No, it was called Future Splash
and it was these brothers, Jonathan Gay,
I think his name was.
He came knocking and the marketing guy at Netscape
who was screening the technology partners
or wannabe acquisitions was brutal
and just everybody wanted to get in on the Netscape
stock gravy train and he sent them packing
and they ended up selling to Macromedia
and Macromedia was where Flash was created.
And the good thing about Macromedia was it was a tool company
so it invested in the best ideas, I think,
which are still somewhat lost to us of Flash.
The timeline, animation is sort of an immutable function
over time.
- They had the tooling around that too,
like the Dreamweaver, there's a Flash--
- Flash Director, there were a bunch of them.
- Yeah, I mean, yeah, that was a great--
- Flash Builder was one of the last ones.
These tools were used by real artists
and special effects people and designers.
All the restaurant websites around 2005 were done in Flash
which was, we were trying to do HTML5 at the same time
that was the Firefox era.
We were trying to make the web capable enough
you didn't need Flash but if you recall,
you go to a restaurant and it's like,
this is kind of like a game or something.
It's like a Flash, all the font looks small.
- You didn't like Flash from the beginning.
You're like, this doesn't feel right.
- Not really, I actually admire Flash's technology
and I'm pretty pragmatic about these things
and I realized that it doesn't matter
if you're dealt a bad hand like JavaScript was a rush job
or if you have Flash as a plugin
and you can invest in the tools and make it pretty good.
You should make it better for your users
and grow it as best you can.
And what happened with the browser
due to Microsoft's monopoly abuse
for which they were convicted and even after that
until I think Firefox and then Chrome
was people kept saying, oh, the web can't do X,
it can't do Y.
We'll have to have a plugin.
We'll have to have a new approach.
We'll clean the slate and have a new web
and everyone who said that failed
and the reason they failed is because there's too much value
in the web, this huge network
and the worse is better principle
means that you can not only start bad
which they all sneer at but get on first
and get wide distribution, get sort of evolutionary advantage
and priority of place but you can also improve it over time.
And so if you're gonna improve Flash
and for some reason Flash is now out of favor,
Steve Jobs said you can't have Flash on the iPhone.
That was probably the death knell.
Put your energy into JavaScript and that happened, right?
So we did things at Mozilla with Adobe to improve
which bought my Macromedia to improve Flash
and to improve the version of JavaScript that was in Flash.
We tried to standardize that.
- Oh, that's right.
- I'm getting ahead of myself.
- You can program Flash. - ES4, yeah.
- That's right.
Can we just rewind to the magical,
like it's a special moment in the history
of all of computing.
We'll talk about it later but it's arguable,
it's possible that the entirety of the world
will run in JavaScript at some point.
So it's like those days, it would be interesting
if you could just describe actually zooming in
on how the cake was baked from the several days
that you were working on it, what was on your mind,
how much coffee were you drinking,
were you nervous, freaking out?
- I'll try to remember it.
I mean, you're right, there were these pregnant moments
you see in hindsight, maybe they're overrated
but like Hegel sees Napoleon on horseback at Jena
and says, "There's the world spirit on horse."
And I knew that there was a chance to do it, Mark knew,
and he was my executive sponsor
and he was the one sort of brainstorming
how the JavaScript should be right there in the page.
That was important for him to say that
'cause I thought so too but a lot of people were like,
"Oh, you can't write programming language
"in the middle of the markup."
And indeed there are problems, if you did it naively,
you'd see the code laid out as like random gibberish.
So I had to figure out how to hide that,
that was a challenge.
- Is that a breakthrough idea?
I mean, so you and Mark thinking about this idea
that you just inject code in the middle of the markup.
- Of the webpage, yeah, it was considered kind of heretical.
There was an SGML guru, I forget his name,
but he corresponded with me and at first he was angry.
He's like, "You should have used a marked section,
"why didn't you use a marked section?"
And I said, "Well, SGML marked sections
"are not part of HTML by the way,
"and they're not supported in the browser."
And so I did some hack that was equivalent
and over time you could do the proper SGML thing.
But eventually he came around and it was again
sort of evolutionary necessity,
it was almost like introgression,
like the idea which Lin Margulies I think helped get across
that we have to consider mutualism biology
that maybe mitochondria were ancient prokaryotes
that got into the cell and became beneficial.
Somehow the same sort of thinking applies,
you have to embed JavaScript in HTML.
It's gonna be a good virus.
- It won't hurt you.
- So the code becomes data in the sense
it just gets carried along.
But is there the side of the,
so you were focusing on the Netscape at that time,
doesn't the browser have to support,
interpret correctly this mix of HTML and whatever code?
- I had to hide it from old browsers including Netscape 1.1
which was predominant then.
So I used an HTML comment,
but inside the container that comment lived in,
the script tag, which is a new element,
I could make different semantics in Netscape 2
where those HTML comment delimiters
instead of being multi-line brackets
became one line or essentially one line.
- So you wrote, so JavaScript was written,
the programming language was written as a comment.
- Yeah, comment for old browsers
and a set of brackets that were ignored
with real code for new.
And it was this two-way comment hiding hack, as I called it,
that was absolutely necessary for us to get off the ground.
We couldn't have bootstrapped JavaScript without it.
We didn't have scripts that were loaded
from a separate file.
The only scripts in Netscape 2 were inline in the document.
- What were the challenges here?
Like what, you know, typing,
what were the choices you were thinking about?
- Garbage collection. - Was garbage collection.
- I didn't have time to write a garbage collector,
so I just, I didn't at first.
So the thing was using essentially arenas
or what GNU calls obj pools
and just would run out of memory eventually.
And I added reference counting in a hurry
after the 10 days in which I hacked.
So after I was in the server team doing HTTP 1.1
and thinking about the language,
I finally got transferred to the client team in early May.
And that's when I, you know, I got the go sign from Mark
and it was like, we can't wait
because people inside Netscape are doubting,
even people inside Sun are definitely doubting.
Bill Joy was the champion, but he was like alone in that
in seeing there was a role for JavaScript
as the, as I call it, the sidekick language,
robbing the boy hostage.
(laughing)
Frank Miller put it in the Dark Knight Returns
that there was this silly little language
that would be the glue language
and it could become important over time.
And you were better off having that complementarity,
§06Transcript · 40:23–48:18
that pairing of languages,
just like Microsoft stacked it
with Visual C++ and Visual Basic.
- So what was the big moment of I'm done?
- So I had to do a demo.
I forget the dates.
I think I, for a history of programming languages paper
that Alan Weissbrock did with my help,
he did a lot of the writing.
I think it was the 10 days from like Thursday evening
through to the following weeks, you know,
the whole of that week and then into the Monday.
- Did you get sleep?
- Not, not enough.
And I was really going fast
because I'd already used a lot of C compiler
and front-end compiler knowledge
that I'd gained from undergraduate school.
When I started getting into computing
as a renegade physics major,
people were formalizing more efficient bottom-up grammars,
parsers for bottom-up languages.
Really, LALR1 was the big thing.
And I studied all this and learned how to parse them.
And in the end, if you're doing C languages,
you often do what Dennis Ritchie did anyway,
which is a recursive descent parser.
You can hand code it.
And I did that for JavaScript in a blazing hurry.
Mostly got it right.
Didn't have precedence inversion problems or other bugs,
but I copied a lot from Java and C.
And I tried to keep things simple.
Like the equality operator in those 10 days sprint
between two objects of different dynamic type said,
"No, they're not equal.
Their types are different."
And then after that, I had internal early adopters
and they were using JavaScript to match a number
against a database field that had been stringized.
And they said, "Oh, can't we just have implicit conversion?"
And like an idiot, I agreed.
I gave them what they wanted.
I was trying to please them and get adoption.
And that broke what equivalence relation nature
there was to the double equal.
There's some edge cases with not a number
that break that at two.
But it really broke it.
Having implicit conversions in the operator
is something that people still roast me over.
- So let's talk about two things.
One, it sounds like the comparison operator,
the equality operator is the thing that you regret.
So maybe can you--
- Making it sloppy.
- Making it sloppy.
So what is the biggest thing you regret in those 10 days?
And what is the biggest thing you're proud of?
- So that making it sloppy came after the 10 days.
And my lesson there, which I've tweeted is,
when people come to you saying,
"Can you please make it sloppy or add this cute feature?"
The answer should be no.
And I should have known that 'cause I think Niklaus Wirth,
one of my heroes, said,
"The essence of design is leaving things out."
But during the 10 days, I also, like I said,
I was in such a hurry, I left out garbage collection.
Came back to haunt me,
but I got reference counting in in time
that people weren't running out of memory right away
on long-lived JavaScript.
- Wait, what happens when you don't have
garbage collection and you have objects?
- Well, you just run out of memory.
(Lex laughs)
And you know, at first, you write a short script
and the page doesn't last long or it doesn't do a lot,
and it's okay.
- Oh, I see, yeah, yeah, yeah.
- But if you're writing a game or something
and you're doing event-based allocation,
you run out of memory.
And this was noticed in the summer of 1995,
and people were like, "What's going on?"
I was, "Oh yeah, I better go back and do reference counting."
And then the problem with reference counting is
you're writing the language in the runtime in C,
an unsafe language, and if you're reference counting
and you overflow the counter, you mismanage it
so it goes high, it gets stuck high,
you leak memory again and you run out.
If you underflow it, you pre-memory that's still in use.
And even then, we knew what all the security hackers
came to know, that you therefore have
potentially a remote code execution vulnerability.
'Cause this was before things like
non-executable heap memory and
stack defenses against taking over memory.
So if you can, from the remote side,
write some HTML and JavaScript that just happens
to exploit a bug in memory safety,
like it causes JavaScript to underflow a reference counter,
and the script still has its hands on that object
and it's trying to call a method on it,
and there's some kind of lookup function table
in the object, but you've managed to stuff the heap
with strings that forwards their own lookalike
for the function table, you can call some other code.
And this was a problem right away.
So security, JavaScript upped the ante.
Java had this problem too, but in its own VM.
And it just was a separate headache for Sun to worry about.
We had this problem in Netscape right away.
So Netscape 2 came out after my 10 days,
and after these follow-on work to embed JavaScript
better in the browser and to add garbage collection
through reference counting, really,
I call it reference counting, and get it shipped.
We had a bunch of dot releases where we fixed
security bugs like maniacs. (laughs)
- But what is the thing you're,
you know, when you sit back on a porch
and just look out into the sunset,
what are you most proud of from those 10 days?
- I think the first class functions, Sean.
I think, especially since Java didn't have it
and it was somewhat unusual.
- Scheme made it in somehow at the end of the day.
- In spirit, in spirit.
I mean, people complain because Scheme has,
you know, minimalism.
It has six or seven special forms.
It has hygienic macros.
It has call CC.
It has sort of a beautiful, complete set of forms
to make the Lambda calculus pleasant to use in practice.
And JavaScript is, you know,
kind of a multi-paradigm or Schembolic. (laughs)
- Just on a small tangent, you mentioned Mark Andreessen.
It sounds like, and Bill Joy, but staying on Mark,
it sounds like he had an impact on you
in that he sort of believed in what you were doing there.
Can you talk about like what role Mark had in your life?
- Yeah, we would meet at the Pinsla Creamery
in downtown Palo Alto.
And Mark was just fresh out of, you know, grad school
or whatever he was doing.
And he was a big dude and he got fitter later.
He had hair.
He would order giant milkshakes and burgers.
And we would meet there and brainstorm about what to do.
And it was very direct because we didn't have much time.
Sort of, we didn't talk about it,
but the implication was Microsoft was coming after us.
Mark was saying things boldly, pre-IPO,
like Netscape plus Java kills Windows, right?
- Wow, it's ambitious.
- Make a browser programmable,
it becomes the new runtime for programs.
It's the meta OS or it's the replacement OS.
- But he still saw value in JavaScript.
- Yes, even though he was saying that
and Java was the big name, hence the trademark license,
he saw JavaScript as important.
And he even thought, what if we got,
I told this in other interviews, I can say it,
he thought, what if we had my friend Kip Hickman,
who'd been at Netscape from the beginning
and who was a kernel hacker at SGI when I joined,
he started writing his own JVM
before we consummated the Sun deal
and got our hands on their code.
And the Java compiler, Java C,
which Arthur Van Hoff had written, very nice code,
was all written in Java, it was self-hosted
or so-called bootstrap.
And so we could use that as soon as Kip's Java VM
could run the bytecode from the Sun JVM
running the self-hosted compiler to emit the bytecode.
So once we could bootstrap into Kip's VM,
we wouldn't need Sun.
And Mark was like, well, maybe we can just ditch Sun,
we'll have Kip's Java VM, we'll have your JavaScript VM.
Now we need graphics.
So Mark was thinking far ahead
because he knew you could do things with HTML and images,
but at some point you really want--
- Like dynamo graphics or three-dimensional?
- Like even SGI had already started its downfall
because the first four VLSI team there
had gone off to do 3DFX and all these other companies
that made the graphics card on your PC, right?
Doom was big and Quake.
And so we were all playing Quake.
I was old, so I was terrible.
But why not put that graphics capability on the web?
And in fact, it finally happened at Mozilla
with Firefox era with Vlad Vukicevic
§07Transcript · 48:21–55:33
taking OpenGL ES and reflecting it as WebGL.
But OpenGL ES is the mobile version of OpenGL,
which is a standard based on SGI GL.
So this whole lineage of graphics libraries
or really graphics languages for what became the GPU.
And Mark was thinking ahead.
He's like, "We need graphics too."
And I thought, "Okay, I can try to get somebody
I knew at SGI, but he's a grad student at MIT.
He was studying under Barbara Liskov.
He laughed when he heard about this later.
Andrew Myers, he's at Cornell, long time.
I think he's a full professor.
And Mark said, "Great, we'll get him."
And I'm like, "I'm not sure he's gonna come.
We'll throw money, we'll stock options."
We never did it.
And they did the Sun deal.
So Kip Nobly put aside his own JVM
and we used the Sun JVM.
So that was an ambitious period.
And Mark was very generative because he was pushing hard.
He was ambitious and he wanted to have Netscape
possibly be in control of the ball.
- Maybe you can speak to this dance
of Netscape versus Internet Explorer.
You've thrown some loving words towards Microsoft
throughout this conversation,
but that's a theme with, I mean,
Steve Jobs had a similar sort of commentary.
From a big sort of philosophical principle perspective,
can you comment on the approach that Microsoft has taken
with Internet Explorer from IE1 to Edge today?
Is there something that you see as valuable
that they're doing in the occasional copying
and that kind of stuff?
Or is the world worse off because Internet Explorer exists?
- So I'm gonna segment this into historical eras
because I think Microsofters today with Satya
is quite a different company
and what they're doing with Edge is different.
But back then, Gates, aggressive character,
not really original in my view, not an originator.
Steve Jobs famously said once, "He doesn't have any taste."
And I don't mean this in a small way.
He has no taste.
You can see this.
Apple at the time had beautiful typography
and ligatures and kerning and the fonts looked great.
And Windows had this sort of ugly system font
that was carefully aligned with pixels
so it didn't get anti-aliased.
- What is it?
I'm sorry to keep interrupting.
Why was Internet Explorer winning
throughout the history of these competitions?
- Distribution.
Distribution matters more than anything.
And this is why even now we're seeing
in the browser wars, Edge doing better
because it's being foisted on people of Windows.
We have Windows 10 boxes at home.
We have some Windows 7 boxes or laptops we keep running too
because we don't connect them to the internet generally.
But once you have that operating system to hold,
you can force Edge.
And Apple did it with Safari too.
It's not unique to Microsoft.
- That's sad.
- But distribution matters.
And that's why I think IE was going to win.
That's why everybody at Netscape felt we're doomed.
This was something Michael Toy and Jamie Woodson were doomed.
But for a while there, we had a chance
and we innovated in Netscape too.
We did a big platform push.
Java and JavaScript and plugins, more plugins,
and more HTML table features.
And really started making a pre-arable stack
out of what were pretty static web languages.
And even in the beta releases of Netscape,
two people were using JavaScript
to build what you would call single page applications
like Gmail.
And they were using JavaScript locally to compute things
and to call the server on a hidden frame in the background.
So it was prefiguring a lot of what came later as AJAX
or dynamic JavaScript, dynamic HTML.
- So people saw that, I mean--
- Even then they saw it, yeah.
- That's kind of, I don't know, from my perspective,
that seems quite brilliant.
It seems like really innovative
that you would have code run in the browser.
- It did impress me with something,
which I learned later about from Eric von Hippel of MIT,
which is user innovation networks, lead user effects.
That throwing out JavaScript,
even though we weren't doing open source,
we were doing beta releases early
and permissively with Netscape.
Getting early developer feedback, absolutely critical.
I loved it.
I did some of that with SGI,
with some of the products I worked on,
but it really came to the fore in Netscape.
And that culminated in Mozilla,
where you're dealing with developers all the time
and early adopters, lead users.
But the lead users helped improve JavaScript,
even in those last few betas,
where I could hardly change things.
I was under pretty rigid change control.
- So we're talking about just a small collection
of individuals that are just like up front--
- A guy named Bill Dortch,
you can find his work in the web archive,
still from 1996.
It's a single page application.
It's an artist gallery of mountain art.
- He used JavaScript.
- It doesn't quite work.
He uses JavaScript locally.
He uses a local database.
What you would think of now is JSON,
but it's all pure JavaScript code,
a bunch of objects being constructed.
- That's so cool.
So how is, if you can do sort of a big,
sweeping progress of JavaScript,
how has JavaScript changed over the years,
in your view, from those early 10 days,
with a quick addition of garbage collection
and fixes around security?
How has this evolution, that now it's taken over the world?
- It's been a bumpy ride,
because the standards body got shut down after Microsoft,
I think, took over the web,
and then felt punished by the US v. Microsoft
antitrust case.
- Can you speak to the standard body?
- That was a fun ride, too,
because Netscape had taken the lead with the web
and HTML innovations like frames and framesets tables.
And the W3C was sort of off,
even then, sort of in SGML land,
heading toward XML la-la land.
I'm gonna be a little harsh on it.
- What's SGML?
I'm sorry.
- SGML was the precursor markup language to HTML,
or it was sort of the more extensible
standard generalized markup language.
It was a-- - XML-like.
- Pointy brackets, but it had all sorts of elaborate syntax
for doing different semantics.
And this is why I think TBL and others
who wanted to do the semantic web
then took XML forward.
But they had this, or some of them anyway,
had this strange idea that they could replace the web
with XML, or that they would upgrade the web to be XML.
And it couldn't be done.
Worse is better had concrete meaning.
The web was very forgiving of HTML,
including sort of minor syntax errors
that could be error corrected.
Like error correction isn't generally done
in programming languages because--
- Right.
That's another amazing thing about HTML,
is like it's more like biology than programming.
- Yeah, exactly.
And so XML was, in its standard form,
super strict and could never have admitted
the kind of users who were committing these errors.
And the funniest part was Microsoft said,
"Hey, we're doing XML."
But the way they put it in Internet Explorer
under the default media type,
put it through the HTML error corrector.
(laughing)
- Oh, wow.
- So they kind of bastardized it
to make it popular and usable and accessible.
And so XML as a pure thing was never gonna take over.
And W3C was kind of not fully functional
because Netscape wasn't cooperating with them.
We thought about where to take JavaScript
and we realized, our standards guru, Carl Cargill,
realized there was a European standards body
that had already given Microsoft fits
by standardizing parts of the Windows 3.1 API,
which European governments insisted on.
They said, "Microsoft, we can't use your operating system
"without some standards."
And Microsoft said, "Here's our docs."
And the government said, "No, we need a European standard."
So this body called
the European Computer Manufacturers Association, ECMA,
§08Transcript · 55:37–1:02:50
which eventually became global
and became a proper noun instead of an acronym.
- Right, it's just one capital E now
with a lowercase CMA.
- Right.
And as one of the early Microsoft guys I met
when we first convened a working group
to talk about JavaScript said,
"It sounds like a skin disease."
- But it gave, I mean, maybe you'll speak to that,
but it gave the name to JavaScript of ECMAScript.
- That was the standard name
because Java was a trademark of Suns.
They were so aggressive,
they were sending cease and desist letters
to people whose middle European heritage
meant their surname was Javanko
and they called their website javanko.com
and Sun would send them a letter saying,
"You're using J-A-V-A at the start of your domain name.
"You must cease and desist."
- I love marketing more than anything else in this world.
So ECMAScript and now is popularly named as ES plus version.
- I would say people use JS more than anything.
People still say JavaScript.
JavaScript is in all the books.
- So I mean, when you're referring to it,
it's usually JavaScript
and when you wanna refer to a version of JavaScript,
you'll say ES6, ES5.
- Yes, or now they've gone to years,
which is kind of confusing
'cause it's an offset of 2009.
ES6 is ES2016.
- Yeah, it doesn't match the years perfectly.
Yeah, so what were the choices made
and how did JavaScript evolve here?
- So we took this new standards body,
which we thought sort of a proven record
of standing up to Microsoft,
but Microsoft sent a lot of people.
They sent some people who were pretty good
and then when they realized that I was there
and Netscape was not gonna just bend over
and do whatever they wanted,
they sent somebody really good.
He was a smart guy.
He did a lot of the work on the first draft of the spec,
Sean Katzenberger, he's left Microsoft.
He even did what I sort of did.
He told his bosses, stop bugging me to do other things.
I'm focused on this
'cause it took a lot of focused work
to create the first draft of the spec.
And I was still holding,
I was spending almost all the plates.
I had like part-time help in certain areas
and on the front end integrations,
I had the front end guys.
But I couldn't take as much time as Sean was
to write the draft spec,
but I had to participate
because I was essentially helping write down
what the language did.
And in areas where we didn't like what it did
and Microsoft didn't agree,
we sometimes got away with slight changes.
And that's the story of standards.
You have different implementations
and depending on their market power,
they interoperate where you have agreement
and where they don't,
the dominant one usually sets the de facto standard.
And then you should probably reflect that
into the de jure standard.
And this happened with JavaScript.
Over time as Netscape went down and Microsoft went up,
we did the first edition of the standard codified in 1997.
In France, we had a trip to Nice,
which was very memorable.
- For any interesting reason or just because it's Nice?
- And ECMA's European and IBM and others were there.
Mike Kalashaw, an IBM fellow was a British.
And the guy who ran ECMA at the time,
Jan van den Bel was quite a raconteur
and a very fun guy.
And he had us out for the great,
"Fui de mer," the bouillabaisse.
- Was the standardization process beautiful or painful
that those early days, you as a designer of the language?
- It was painful because it was rushed.
Now, Guy Steele was contributed by Sun.
So even more than Sean,
you had this giant brain, Guy Steele helping,
bringing some of that scheme magic.
He even brought Richard Gabriel for fun.
Richard wrote the fourth clause of the ECMA standard,
which was kind of an intro to what JavaScript's all about.
So we had some really good people.
And we didn't fight too much.
There was some tension where I was fixing bugs
and I was late to a meeting.
And Sean Katzenberger of Microsoft was actually mad.
Like, "Where is he? We need him."
And when I got there,
I saw that only he saw this sort of off by one bug
and somewhere in the spec.
And then I saw it too.
And I said, "There's a fence post bug there."
And then we kind of locked eyes
and we realized we were on the same page.
And we kind of, he wasn't mad anymore.
- What were the features that are being struggled over
and debated and thought about?
- It was mainly writing down what worked
and what we thought should work in the edge cases
that didn't interoperate or that seemed wrong.
But we were already laying the groundwork
for the future editions that I was already implementing.
I was still trying to lead the standard
by using the dominant market power
to write the code that actually shipped.
So the de facto standard would lead the de jure standard.
And I was putting in the missing function forms
that I didn't have time for in the 10 days.
- So this is the engineering mindset
versus the theoretician.
So you didn't want to create the perfect language,
but one that was popular and shipped
and all that kind of stuff.
- And you could say there was,
I was standing on the shoulders of giants.
So there was a staged process where I had to hold back
things that were well designed by others
in other languages that I could imitate.
But I couldn't do them all in the 10 days.
So they came in in 1996 and '97.
And they came into the third edition of the standard,
which was finalized in 1999.
But at that point, Netscape had been sold to AOL
and was, which was a decent exit considering.
And had previously been mercilessly crushed.
Netscape was selling the browser along with server software
that it had acquired after its IPO.
And Microsoft was just underpricing it.
So there was no way to compete with that.
Microsoft was also making Internet Explorer
the default browser in Windows,
which is called tying in antitrust law.
And they were doing even more brutal things.
There's a famous investor, he did very well on Google.
So he's a billionaire, Ram Shriram.
He was sales guy or head of sales at Netscape.
And he got off the phone looking ashen faced
after Compaq called and said,
"Microsoft just told us they're gonna pull
"our Windows license if we ship Netscape
"as the default browser."
- Wow.
So there is some bullying going on.
- It was totally immaterial in the antitrust case.
But JavaScript escaped into the standard setting
where there was fairly good cooperation.
Microsoft had a really good guy on it.
And Guy Steele was there for a time.
And there was some good work.
But after the antitrust case and Netscape
kind of dissolving into AOL
and not really going anywhere quickly,
Mozilla took years to really bring up,
the standard froze.
And by 2003, even though they'd been sort of noodling around
with advanced versions, JavaScript 2,
I'd given the keys to the kingdom
to another MIT grad, Baltimore Horwatt.
Very big brain, still at Google, I think.
He won the Putnam in '86.
So he's, yeah, very mathematical.
- Legit.
- He designed the successor language, JavaScript 2,
but it only showed up in mutated form
in Microsoft's ASP.NET server side.
And it didn't last there.
And it showed up in Flash.
And that's what became ActionScript 3.
- Ah, ActionScript.
Interesting.
And then Flash was declined.
And so how did we arrive at ES6,
where it's like, there's so many,
where everyone, okay, there's this history of JavaScript
that people were, it was just like, cool,
when you're having beers to talk crap about JavaScript.
Everyone loves to hate, like people who are married say,
ah, marriage sucks, is they just wanna let off some steam,
even though everyone uses the language.
But ES6, it's become this,
it fixed major pain points, I think.
- It added things to the language
and added something that was already ES5 strict mode,
§09Transcript · 1:02:53–1:10:45
but made it implicit in class bodies and module bodies.
It was a big jump, but it accumulated
some of the ES4 designs that we'd done with Adobe
for what we hoped would be the fourth edition of ECMAScript
that were supposed to fold in some of these old JavaScript,
two ideas that had come into ActionScript 3.
So you look at the family tree and you see these forks,
and the main ones are the ones that go into Adobe Flash,
acquired from Acromedia,
and the one that went into the server side
of Microsoft's stack, which kind of died.
And then trying to bring them back into the standard,
and not quite succeeding, ES4 was mothballed.
But all the good parts that everyone liked
made it into ES6.
And so that was a success.
And I said earlier, I had the wrong year,
I think it's 2015, so it's off by--
- For ES6.
- Yeah, it was done, finalized in 2015.
It took a little longer than we hoped,
but 'cause ES5 was 2009,
and that was a smaller increment from ES3.
We skipped four again, we mothballed it.
And we had a split in the committee
where some people said, ES4 is too big,
we're gonna work on incremental improvements,
no new syntax in particular, they promised.
Not quite true, but they added a bunch of interesting APIs,
Alan Weir-Sprock, my co-author of the Hubble paper.
And he was at Microsoft at the time,
I ended up hiring him at Mozilla,
he wanted to get to Mozilla and keep doing
the sort of editor job of the JavaScript standard,
ECMAScript.
And when we got ES6 done, it was a little late,
2015, and we switched to year numbers.
So people still call it ES6, I call it ES6.
But if you remember, off by nine, plus 2000.
- Yeah, I mean, ES6 is such a big job.
I mean, like you said, there's a thread
that connects all of it, but ES6 is one
that's like, became this language
that almost feels ready to take over the world completely.
- More programming in the large features,
more features you need for larger teams.
- Software engineering.
- Microsoft did something smart too,
Anders and company, Luke Hoban, who's left Microsoft,
also did TypeScript.
And they realized something, I think,
that Gilad Brak has also popularized,
and he was involved in Dart at Google.
If you, don't worry about soundness in the type system,
you don't try to enforce type checks
at runtime in particular, just use it
as sort of a warning system, a tool time type system,
you can still have a lot of value for developers,
especially in large projects.
So TypeScript's been a roaring success for Microsoft.
- What do you think about TypeScript?
Is it adding confusion or is it ultimately beneficial?
- I think it's beneficial.
Now, it's technically a superset of JavaScript,
so of course I love it.
(laughs)
The shortest JavaScript program
is still a TypeScript program.
Any JavaScript program is a TypeScript program,
which is brilliant 'cause then you can start
incrementally adding type annotations,
getting warnings, learning how to use them.
Microsoft's had to kind of look around corners
at the standards body and guess how their version
of modules or decorators should work.
And the standards body then may change things a bit.
So I think they're obligated with TypeScript
either to carry their own version
or to bring it back with incompatible changes
towards the standard over time.
And I think they've played generally fair there.
There's some sentiment that,
why don't they standardize TypeScript?
Well, they've been clear they don't want to.
They have a proprietary investment, it's valuable,
they have control of the ball.
And in some ways you can say the same thing
to any of the other big companies in the standards body,
why doesn't Google standardize its stuff?
- So you think it'll continue being
like a kind of a dance partner to JavaScript,
to the base JavaScript?
- There's a hope that at some point,
if they keep reconverging it and the standard
doesn't break them and goes in a good direction,
we will get at least the annotation syntax
and some semantics around them.
Because when you're talking about type annotations,
they're generally on parameters and return values
and variable declarations, they're cast operators.
You want that syntax to be reserved
and you want it to work the same in all engines.
And this is where ideas like GILAD's
pluggable type systems might be good,
though then you could create the same problem
you have with Lisp and Scheme,
where there's a bunch of macro libraries
and they don't agree and you have conflicts between them.
But pluggable type systems could be one way to standardize.
- What do you think about the giant ecosystem
of frameworks in JavaScript?
It feels like, because, I mean,
this is a side effect of how many people use JavaScript,
a lot of entrepreneurial spirit,
like create their own JavaScript frameworks.
And they're actually awesome in all different ways.
And this is an interesting question
about almost like philosophically
about biological system and evolution,
all that kind of stuff.
Do you see that as good or should it,
should some of them die out quicker?
- I think that maybe they should.
Now jQuery was a very clever thing.
John Resig made this library that was sort of query and do
and blended sort of CSS selector syntax
with JavaScript sort of object graph or DOM querying
and made it very easy for people to do things
almost like they were learning jQuery as its own language,
domain-specific language.
And that I think reflected in part
the difficulty of using the document object model,
these APIs that were originally designed in the '90s
for Java as well as JavaScript.
They were very object-oriented or even procedural.
They were very kind of verbose.
And it took like a constructor call
and three different hokey-pokey dances to do something,
whereas in jQuery, it's just one line.
So that fed back finally into the standards.
It didn't mean we standardized jQuery.
It wasn't quite that concise.
But you find now with the modern standards
that we were working on in the HTML5 sort of effort,
that things became simpler,
the fetch API and the query selector API,
document.querySelector.
A lot of things can be done now in raw JavaScript
that you would make more concise and terse in jQuery,
but it's not bad.
It's pretty good.
Whereas in the old DOM of 15 years ago,
it was just too verbose.
- So maybe the frameworks were born
because JavaScript lacked some of the features of jQuery.
And so now that JavaScript is swallowing what jQuery was,
then the frameworks will,
only the ones that truly add value will stick around
and the other ones will die out.
- And that highlights also this division
between the core language JavaScript,
which can show up in other places
like Node.js on the server side
and the browser specific APIs
or the document object model APIs,
which are even managed by the W3C,
the standards body that was off in XML
while I land on we were doing
real JavaScript standards in ECMA.
And you have this division of labor,
division of responsibility and division of style
and sort of aesthetics and also speed.
So the document object model really stagnated
after Microsoft kind of de-invested in the web.
And Microsoft did something in their haste
in the spirit of Netscape,
doing things quickly and getting on first called DHTML.
And some of their innovations
that were like an alternative document object model
didn't really get standardized until HTML5
when we pragmatists set up for at the time,
Ian Hickson, who went to Google, Apple and Mozilla said,
"XML is not gonna replace HTML.
HTML4 is too old.
Let's standardize HTML5 based on all this good stuff,
including that DHTML variant, dynamic HTML.
- HTML5, it feels like to me,
maybe you can correct me,
like a beautiful piece of design work.
It's not often with web stuff,
you have this breath of just like,
oh, whoever did this, this just feels good.
What are your thoughts about HTML?
Am I being too romantic?
- A little bit, a little bit.
- Are there flaws, fundamental flaws to it
that I'm just not aware of?
- My old friend Hicksey did a great job.
He was another renegade physics student.
And he was basically a QA guy at opera,
§10Transcript · 1:10:48–1:19:12
but he obviously trained physics student
and someone who could write, a Britisher.
He developed test suites
and he started thinking about them more axiomatically.
Now this can be good because you can sort of systematize
in a way that makes a better HTML,
or you can get caught in the pragmatism of saying,
well, we have to handle all of these edge cases,
so we're just gonna have sort of a test matrix.
And if the matrix is large,
it will not be beautiful by many people's lights.
Everyone likes to minimize along their preferred dimensions,
the seven special forms and scheme or whatever.
But reality is HTML needs to be big.
It's kind of shambolic, it's a creative multi-paradigm.
And Hicksey did a good job, I would say, with a bunch of it.
Other people came in in the spirit of Ian Hickson
to do HTML5 work and they've carried on that effort.
And so it's a mix of pragmatism,
de facto standards from the past being sort of combined
or written down for the first time,
and then rethought in a way that has a simpler syntax,
like the fetch API instead of XMLHTTP request.
- This video too as well,
it ultimately, it feels like, maybe you can correct me,
it feels like it was the nail in the coffin of Flash.
- Steve Jobs saying no Flash on the iPhone,
in my opinion, was the actual stake to the heart.
But, well, I'm not sure what trope you wanna use.
Flash was a zombie until just this year, right,
or last year, I think last year was the end of Flash
in main browsers.
But Jobs really did the death blow.
And yet, you're right, we had to make HTML5 competitive.
I still don't think we got that beautiful
timeline animation.
- The timeline thing, so you like the time,
I mean, me from, I used to animate all kinds of stuff
inside Flash, plus there's a programming element.
It was a little bit, I don't know if you can comment
on that, but to me, it was a little bit like
go-to statement, like in a sense that
it was a little bit too chaotic.
Like it didn't, that OCD part of me as a programmer
wasn't satisfied by Flash.
It feels like there was bugs that were introduced
through the animation process that I couldn't debug easily.
- Yes, I heard that too.
I didn't use it, so I'm doing the grass is greener thing
here.
The thing I liked about the animation model
was that it was this immutable function of time,
so you could time warp and you could,
if you dodged these bugs or worked carefully,
you could really make it sing in ways that I think
still a little challenging with web animation standards,
but, or just using raw canvas and WebGL.
But there's so many tools now that maybe it doesn't matter.
And yet we had to do video, we had to do WebGL
and then evolve it.
We had to do web audio.
But once we did all these things that helped Flash die,
thanks to Steve Goves, we had something that
people didn't realize.
We had that vision that Mark and Jason had,
this graphics capable to the metal, portable runtime.
And we at Mozilla realized this and we saw JavaScript
was something that you could compile to.
Adobe had somebody in the Adobe labs doing this too.
He had a project called Alchemy.
We had somebody who's now at Google,
Alon Zakai, who did his own LLVM based compiler
that would take C or C++ and it would emit JavaScript.
And you would think this is crazy.
You're going from this sort of machine types,
low level, controlled memory allocation language
to this garbage collected dynamically typed,
high level, higher level language.
But Alon sort of just phenomenologically carved nature
of the joint and found the forms
that were fast in JavaScript.
And then with Dave Herman, who I'd recruited
from Northeastern University, who was a type theorist,
and Luke Wagner, who's still at Mozilla,
who was the compiler guy and the JIT guy,
they figured out how to codify what Alon had done
into a typed subset of JavaScript called Asm.js.
And this is a strange thing to think about
because it doesn't have new syntax.
The types are casts that occur in dominator positions
in the control flow graph.
So it's like a hack on JavaScript and it's a subset.
And it uses those bitwise operators
that I talked about copying from Java
to basically cast numeric types,
which are double precision point into integers.
And so inside JavaScript,
in the kernel semantics are integers.
And if you use these operators,
if a compiler emits them in the right places,
you can then treat them as typed values,
typed memory locations, and you can type check your program.
You can not only type check it, you can compile it.
This is all in sort of linear time.
Oh, and you can compile it
to have deterministic performance.
It doesn't touch the garbage collector.
It calls a bunch of functions that come from the C functions
or C++ code that you're compiling.
And you can make the epic Unreal Engine
go in 30 frames a second.
And when we did this in 2013 in the fall,
Tim Sweeney, I bet, didn't think it could be done quickly.
He thought it would take years.
And the team went to Raleigh to Epic,
and in four days they had Unreal Engine ported
by pressing a compile button.
But they had to have WebGL, which came from OpenGL,
ES came to OpenGL, which came from Silicon Graphics GL.
They had to have Web Audio, so they could map OpenAL,
which was another audio library standard,
to Web Audio, which was kind of a Chrome idiosyncratic thing.
But they could make it work.
And they had to have Asm.js for fast C++ to JavaScript.
And if you didn't have that fast compiler step,
the JavaScript you'd write by hand
trying to do an Unreal game would be too big and too slow.
It would touch the garbage collector.
It would not keep up with 30 frames a second
on the 2013 hardware.
So we demoed that at,
this must have been fall 2012 now that I think about it,
'cause we demoed it at GDC, Game Developer Conference 2013,
and people were stunned.
That's like Unreal Engine, Unreal Tournament,
running in my browser window.
No plugin, no Flash, no Java, no--
- So were those the early days of,
because JavaScript now is able to run basically on par
with a lot of the C++.
- Yeah, and even before then,
you had the fast JavaScript VMs in 2008 when Chrome came out.
Just before it came out, Mozilla,
my friend Andreas Gal and I and others
hacked out TraceMonkey, our trace-based JIT.
The SquirrelFishExtreme team at Apple did their JIT.
And we were all competing
on these crazy performance benchmarks.
It was a little bit too much tuning of the benchmark.
But JavaScript started getting fast,
and developers started noticing it.
But it was still kind of its own high-level language
with garbage collection.
The Asm.js step helped us go further,
because until we really proved the concept,
people were still saying, "Well, JavaScript's okay.
"It's getting faster thanks to V8."
Everybody gave Google credit, especially Google.
But we need something to kill Flash.
Let's use a portable native client code
that Google had acquired, Native Client,
which is a separate lineage for taking basically C code,
compiling it into a software fault isolated container
of some sort, using some kind of virtualization technique.
And maybe it can even be in process
and still be memory-safe.
That would be awesome.
But they ended up using process isolation, too,
and that kind of weakened it.
And in the end, it was like portable native client,
okay, meet the new boss, same as the old boss.
This is the Google Flash, right?
But when we did Asm.js
and we showed Unreal Engine working,
I think it was only a matter of time
before Google threw in the towel.
And in fact, everybody agreed in spring of 2015,
we're gonna take what was proven by Asm.js
and make a new syntax, a binary syntax that's efficient,
that loads into the same JavaScript VM
that JavaScript loads into.
So there'll be two source languages,
one VM, very important, one garbage collector,
one memory manager, one set of compiler stages.
And that's called WebAssembly.
And that's the successor to Asm.js.
And it's important that it have binary syntax
because at the end of the day, especially on mobile,
if you're downloading JavaScript,
even if you're using LZ compression on the wire,
that's cool, but you've got to blow it out into memory
and then parse the silly eight character function keyword
that I picked when I should have used something shorter.
I picked it because of awk, the Unix tool.
So anyways.
- I'm not following, I want to,
but I'm not following the awk thread.
- Yeah, don't worry about it.
- Is it surprising to you that
how damn fast JavaScript is these days?
§11Transcript · 1:19:14–1:27:21
I mean, 'cause you've been through the whole journey.
I know every step of the way, but is it like,
I mean, it feels incredible.
- It does, but I knew, so the funny thing is,
computer science is this big karmic wheel, right?
Wheel of Fortuna.
And in the, it was about the '97,
I was loaned by Netscape to do due diligence for Sun
in their acquisition of Animorphic,
which was David Unger and friends,
people, Craig, I'm forgetting his name,
he went to Microsoft.
These Stanford language buffs who had taken Smalltalk
and then David created Self
as a simpler sort of Smalltalk language
and made really fast, just-in-time compiling VMs for them.
And they, well ahead of Java Hotspot or JavaScript V8
or any of these modern VMs,
figured out how to make dynamic code fast
because Smalltalk is dynamic language, right?
It has classes, it has, I think,
more lockdown declarative syntax than JavaScript,
but it's fundamentally dynamic, you don't declare the types.
But you could infer the types as the program runs
and you start to form these ideas
about what types are actually flowing
through key operations and you form little
so-called polymorphic inline caches
that are optimized machine code.
The cache is the machine code that assumes,
does a quick check to make sure the type is right,
and if it's not right, it bails to the interpreter.
If it is right, you go pretty fast.
And that short test is a predicted branch,
so things are pretty quick.
All that amazing stuff I knew about in the '90s
and I didn't have time to do it
and Anamorphic got bought by Sun and they did Hotspot.
And you needed that even in Java because at scale,
Java has some dynamic aspects due to invoke interface.
You can have basically collections of Java code
where you don't know at the time each module
or package is compiled exactly what's being called,
what subclass or what implementation
of an interface is being called.
And so you want to optimize using this sort
of dynamic polymorphic caching there too.
And they did that and Hotspot is an amazing beast.
I've met like 13 people who all claim they created it.
I think one of them may deserve credit more than others.
But I didn't get to do that in JavaScript
and when we knew that Google was going
to do their own browser, which we knew at Mozilla around 2006
I also met the team that did V8.
And it turns out it was Lars Bach
who was one of the young engineers from Anamorphic
who got acquired by Sun.
And so Lars is like one of the world's expert
on these kinds of virtual machines
and he picked my brains about JavaScript.
I could tell he didn't like it at the time
but he had to do it.
- Oh, really, interesting.
- Yeah, in 2006 launch at Google's campus.
And then I had another friend who was DevRel at Chrome
and he said, yeah, we don't know what they're doing.
This is getting 2007 to fall, getting toward 2008.
We're trying to get Chrome out
and we don't know what's going on with the V8 team.
They're off in Aarhus, Denmark,
rewriting their engine four times, which is good.
That's the right way to do this kind of development.
They were learning JavaScript, including all its quirks
which they came to hate, the fire of a thousand suns
which is one of the reasons that Lars and company
did Dart, their own language.
But they also made the language fast.
And meanwhile, we knew this was happening.
So we got our act together with TraceMonkey,
our tracing JIT at Mozilla.
And Apple, I think, was also aware.
And so they were doing their own JIT.
So the era of JIT-ed fast JavaScript in 2008
had this prehistory going back to Smalltalk itself
and anamorphic.
And again, the lineage is interesting
'cause you had Lars at anamorphic
and then he ends up at Google.
- Yeah, and today we have an incredibly fast language
that like you said, still, without hate, you can't have love.
So I think there's both love and hate for this dance,
this rich, complex dance of JavaScript
throughout its history.
- There's a dialectic for sure.
- Today, JavaScript is the most popular language
in the world.
Why, by many measures, why do you think that is?
Is there some fundamental ideas
that you've already spoke to a little bit
but sort of broader that you think
is the most popular language in the world?
- So I think I did, by doing first class functions
and taking the good parts of the C operator hierarchy
and just keeping things simple enough,
maybe it could have been simpler,
but I had to make it look like Java
and interoperate with Java,
that there was inherent goodness,
Aristotelian quality there.
And people perceive that
even through all the quirks and warts.
And then over time, working on it with the standards body,
working on it not only as a core language,
but in the context of HTML5 and making the browser better,
listening to developers, thinking about,
this is something that Nick Thompson wrote nicely about
on the Hacker News, I was very flattered.
He said, "Java was this thing
"where the experts were writing the code
"and it was compiled and you had to declare all your types.
"And Sun didn't really give a damn about
"the average programmer who wanted to build real web apps,
"dynamic things."
And I was in there meanwhile doing a bunch of people's jobs,
making JavaScript survive those early years
when it was kind of touch and go, right?
JavaScript was considered a Mickey Mouse language.
It was for annoyances like the scrolling text
at the bottom of the browser in the status bar.
But I kept listening to developers,
working with them and trying to make it run
in that single threaded event loop in a useful way.
And I think that forged something
that people have come to love.
Now you don't always love the best thing, right?
I talked about, Shakespeare,
sunnet about, "I'm Mr. Sizer, nothing like the sun."
Or the scene from Josh Whedon's film,
Serenity at the end where the actual piece in the score
by David Newman is called Love,
where Captain Mal is teaching River Tam
about how to pilot a ship.
And she's a super genius, super soldier.
She knows how to do it already.
And he's basically talking about how you have to love
the ship because if you don't, it's gonna kill you.
And then the piece falls off the ship.
It's kind of like JavaScript.
You have to love it.
You have to love it because now people say,
"Ah, we're stuck with it
'cause it got this priority of place."
- But there's love underpinning that.
And actually, listening to developers,
that's kind of beautiful.
There's most successful products in this world
with all the messes, with all the flaws.
Perhaps the flaws themselves are actual features,
but that's a whole nother, that's a discussion about love.
But underneath it, there's something
that just connects with people.
- And it has to keep connecting.
If JavaScript kind of went off in this,
people sometimes complain about ES6.
Oh, you put classes in JavaScript.
I hate classes.
You've ruined it.
But it's not true.
It's a dynamic language.
Smalltalk had classes.
Python has classes.
There are lots of Lisp variants that had classy systems.
Common Lisp.
So people who don't reject it based on some sort
of fashion judgment do use it
and do interact with the standards body.
The standards body is competing browser vendors mainly,
but also now big companies that use JavaScript heavily
the PayPal's and other such companies, Salesforce.
And they have to cater to web developers.
They have to hire developers who know JavaScript.
They have to keep their engines up to the latest standard.
And this creates all this sort of social structure
around JavaScript that is unusual.
I mean, you get C++ buffs that follow the inner workings
of C++, what is it now?
21 something.
I don't know.
I've lost track.
But it's a more rarified group.
It's more like the old language, gray hairs.
Whereas JavaScript is a younger
and more vibrant and large crowd.
- There's a community feel to it.
There's a echoes, perhaps I don't wanna draw
too many similarities.
Maybe you can comment on it.
There's a C++ is like Wall Street
and the JavaScript is like Wall Street bets
from the recent events.
§12Transcript · 1:27:22–1:35:40
It's like, there's a chaotic community of all.
And there's some power from that distributed crowd
of people that ultimately--
- It's more demotic.
It's more of the people.
It lets people in without requiring these credentials.
I remember in the late 90s into the noughties,
people were all getting Java credentials.
And I knew people and friends knew people
who became Java programmers.
And you knew they really should have been
like nature guides or pilots.
They hated programming, but they thought,
I gotta make money.
I'm gonna become a Java programmer.
- Do you have some, because it's such a monumental moment
in our current history as a quick aside,
do you have thoughts about this huge distributed,
crowdsourced financial happenings with Wall Street bets?
- Yeah.
- That's like nobody could have,
well, you could have predicted,
but the scale and the impact of this kind
of emergent behavior from independent parties
that could happen.
- Like I said, my own experience with the dismal science
as with physics led me to reject a lot of bad models.
And economics was always compromised by politics,
political economy.
You could also argue that it was,
it used to be a branch of moral philosophy.
So it was concerned with the good and it became divorced
and became sort of in this quasi Newtonian way,
just about everything's just running by itself.
Don't worry about it.
This monopoly is crushing your Netscape company,
but that's just nature.
- And economics couldn't,
or doesn't really have good models
for the Wall Street bets subreddit.
- They know how to squeeze a short, right?
So the amazing thing is you have Robinhood app,
which was again, supposedly for the demos, for the people,
and eliminated the fee through various kinds of straddles
or some kind of spread operation
that helped them eliminate the fee or eat the fee.
And in fact, as a broker in these days,
because it takes two days to settle,
there's counterparty risk as they found out.
And so the Wall Street bets people,
the memes are like the Terminator robot
with $600 STEMI check and the hedge funds
the little girl hiding under the desk.
There is a problem, which I talked about
in the recent podcast, which I'm conscious of
from the history of the web.
And that is, you could say it's monopoly,
which antitrust wasn't enforced
after US Microsoft for a long time.
And a lot of this was due to the money interests
buying control of politicians.
And in Plato's five regimes, that's oligarchy.
That's where we are.
And now we're seeing a fight against the oligarchs.
I don't know if it'll work,
but you're definitely seeing it.
And it's also kind of hackerish, right?
It's got a hacker ethos.
Hey, Robinhood, no fees.
Oh, interesting.
Hey, I can buy a fraction of a share in this thing,
or I can keep buying with my stimulus check.
So I mentioned Hegel seeing Napoleon on the horse.
Hegel also talked about the cunning of reason,
that you have this sort of God sees history in full,
and if you believe in God, or we don't know the future,
but there's always this sort of fly in the ointment,
this unintended consequence that confounds
the best plans of the powers that be.
And we're living through it.
I'm glad it's not street warfare or mechanized warfare,
'cause it has been in the past.
It's more like soft power.
- Yes.
- And people are fighting back.
- Do you think it's possible,
so JavaScript used to be for the front end of the web.
It's now increasingly so being used for backend,
like running stuff that's like behind the scenes.
And it's also starting to be used quite a bit
for things like TensorFlow.js.
So starting to actually use these heavy duty applications
that are using neural networks, machine learning,
and so on in the browser.
Is it possible in 10, 20, 30 years
that basically most of the world runs on JavaScript?
- This is a dystopia and a nightmare to some people.
(Dave laughs)
When we did Asm.js and WebAssembly,
I would joke and meme people with scenes like
Neo waking up in his pod in the Matrix,
and he's all skinny and weak and hairless.
And you realize in the future that you're living
in some simulation that it's all running on JavaScript,
and you just scream forever.
It's possible.
Gary Bernhardt does these funny talks.
He did Wat.js, and then he did this
Life and Death of JavaScript, I think it's called,
where he took some clever ideas
that actually have a thread of credibility to them.
But I mentioned software fault isolation.
In the old days when we were using computers,
we said we're gonna use the Unix monolithic monitor,
and it's the privileged program.
This is before you even had hardware rings of protection.
Those, some of the early '60s operating systems
used hardware protection zones.
But Unix is privileged, and the program
that runs user code in a process is hosted.
It's the guest in the host, and you get to suspend it.
You get to kill it.
If it crashes, it doesn't take down the whole OS.
It's a wonderful idea.
But the call into the kernel is expensive,
the system call, so-called.
And this has even been optimized now
for things like getting the time of day
so it doesn't actually enter the kernel.
And meanwhile, hardware architectures
and virtualization techniques have gone
in a different direction, even to the point
where you can do software fault isolation very cheaply
without entering the operating system kernel.
And so you get unikernels and exokernels
and very lightweight VMs.
And so Gary took this idea and said,
JavaScript will take over computing
because the system call boundary's too expensive,
so everything ends up in JavaScript
with these lighter-weight isolation enforcement mechanisms.
It's not totally beyond belief.
It'd be WebAssembly, too.
- It's nice to ask you sort of for advice.
There's so many people that are interested
in starting to learning about programming,
getting into this world.
Is there some number of languages,
three to five programming languages
that you would recommend people learn,
or maybe a broader advice
on how to get started in programming?
- Well, so you asked about machine learning,
and JavaScript is a general-purpose language,
and it's a language that's not that great
for doing matrix operations
or doing parallel programming, I would say,
without using some extensions or some libraries
that have some magic in them.
So if someone wanted to learn,
there are amazing languages in sort of the APL family
that are very useful for, I would say, linear algebra,
which gets to a lot of the kernels in machine learning.
And so APL had like J and then K,
and they're K variants
because the guy that did K is still going,
and they're proprietary, but he's still innovating there.
Python is used.
So people talk about TensorFlow.js.
Well, it's not that surprising
'cause Python was heavily used for machine learning,
and Python was always, you know,
they didn't have this fast just-in-time compiler tradition.
There were some projects that tried this,
and some of them were interesting.
PyPy was interesting.
But the philosophy with Python was,
oh, you need to go fast, write a C plugin,
and drop into C code.
So I think people should look at multiple languages
because there are different tools in the belt.
If you're trying to do supervision or rapid prototyping,
you want a dynamic language.
You wanna throw things together and see what works.
If you are trying to go down to the metal very fast,
well, I'm an old C hacker,
but I was also the executive sponsor of Rust at Mozilla,
and Rust has now escaped from that sort of nest
where it was born to be adopted by a bunch of companies
to have a foundation in the works.
Some of the key core team members
are working now at Amazon and other places.
So it looks like Rust has reached escape velocity.
And Rust is an interesting language
because one of our goals there,
one of the reasons I sponsored it was,
we were all tired of seeing
those remote code execution vulnerabilities
due to C and C++.
And we thought, can we have a sort of safety property
§13Transcript · 1:35:43–1:43:45
through a type and effect system or an ownership system?
And Rust has that.
And that ownership system is interesting
because it doesn't just give you memory safety.
There's a sort of theorem for free,
a dual that falls out for protection against data races.
So Rust is better for low-level programming.
You delimit your unsafe code where you do have to be unsafe,
and you can prove certain facts about memory safety
and race condition avoidance.
And so I think people should learn these new languages.
I think Go is a great language.
I admire the Unix people who did that.
Ken Stoll was involved, Rob Pike, of course,
David, what's his name, and other people.
Go is a huge success, really, on the server side,
mainly where you have a lot of networking to do.
And it's garbage collected, but it's also very pragmatic.
It has that sort of C flavor.
As an old C hacker, I can't get used to the fact
that they swapped the type and declarator
in the declaration order.
- I haven't used Rust, but this is one of the most respected
and loved languages currently, so it's interesting.
- Yeah, and it's still young.
You look at these things,
JavaScript is now considered old.
- It's gone through so many versions
that you can fall in love with it all over again.
- 25 plus years, it's an adult.
(laughing)
It should be out of the house.
But it could be around another 25 years.
Cannot rule it out.
So Rust could be around for a long time.
The longer you're around, the more likely you're Lindy
and you're around your wife.
- A lot of people ask me,
I'm often torn between recommending either Python
or JavaScript as the first language to play with,
because it's difficult,
because it's so easy to do JavaScript incorrectly.
It's much easier to do it correctly these days,
or like well, like learn about programming.
But the cool thing about JavaScript
is that you can create stuff
that will put a smile on your face.
Like as a developer, you can create stuff
and it'll visually look like something and it'll do stuff.
And it makes you feel good.
It makes you fall in love with programming.
With Python, you could do the same.
It's a little slower.
And with C++, it takes five to 10 years
to write a program that actually does something.
So like there's that tension between,
is JavaScript the right first step or is it Python?
And I've been going back and forth on those two.
- I have my Python, right?
It came from a lineage of ABC,
which was a pedagogical language in the Netherlands.
And it was a good teaching language too.
I think it is a good teaching language.
And it's a little more restrictive
in that if you misspell something
in a way that JavaScript might let run,
let reach run time, it'll get stopped
at syntax check in Python.
That's good for beginners.
I think the sloppiness that some people object to,
like people were just tweeting at me,
having just learned JavaScript.
They said, "I can take a number and I can index into it
"and get undefined out of it as a property."
And why is that?
A number is not an object.
And I explained why it is.
'Cause like in Java, the primitive types,
which unfortunately are not objects,
can be automatically boxed or wrapped by an object.
And I made that implicit.
In Java, it's typed and you have to declare things
and you'll get type errors.
But there are cases in Java where you get auto-boxing
or auto-wrapping 'cause you've declared that you want it.
In JavaScript, it just happens.
And so once I explained it, I'm like, "Oh, wow, I get it."
But it also means that you can commit a blunder that just--
- You don't get punished for it, you don't detect.
- You get an undefined value
and you don't know where it came from.
- Right.
I've been reading a lot about military history recently.
And one way to paint the picture of browsers,
internet browsers, is through the various wars
throughout its history.
I don't know if that's a useful way to look at it,
but we've already talked a little bit about Netscape
and Internet Explorer in the early days.
Can you tell the story of the different wars,
if that's at all an interesting way to look at it,
of the '90s and to today?
- Yeah, so I mentioned that Microsoft,
which was convicted for it, did abuse its monopoly,
but they had a pretty good team by the time they did IE4.
And Netscape, unfortunately, I was like second floor
and I was friends with all the first floor people,
the front end guys who did the JavaScript event hookup
and things like that, that team was fairly burnt out.
And I think having gone public,
the upper management wanted to buy a bunch of companies
to try to go head to head with Microsoft.
Didn't work, but buying a bunch of companies
usually doesn't work.
I think the modern sort of approach roughly
is like Mark Zuckerberg took,
which is to keep them at arm's length
and let them do their thing.
And now that he's pulling WhatsApp in
and people are fleeing it
because it's tied into the ad surveillance.
But for a while they were keeping it separate.
It really does work 'cause you bought it for its value,
it's complimentary and you're not messing with it.
With Netscape, when they bought a bunch of companies,
they had some of the first floor people
or the founders burned out.
They had newcomers who wanted their turn to do the browser
and they hadn't really done browsers or understood them.
And so Netscape 4 was originally supposed to be 3
and it was so late, they renumbered it.
We did a 3 release, Jamie and a few others
put some extra effort into secure MIME
was supported in the mail, built-in mail program.
And Netscape 4 was late and it was only on Windows at first
and Microsoft had really started doing better.
Like they do, they copy and the first version is trash
and the second one, you're starting to feel threatened.
The third one, you can tell what's gonna happen
and the fourth one's good.
- And plus there's the benefit, like you said,
that it comes as a default browser.
- Yes, and yet Netscape's screwing it up
and Microsoft really putting some quality people on it.
IE 4 was good, on Windows it was good.
And they did the dynamic HTML innovations,
they, Scott Isaacs, my old buddy,
a former accountant who programmed in BASIC
and became what Microsoft calls a program manager,
which is kind of an elevated position.
It's, you can be a programmer or an engineer track
but you switch to it and you sort of lead a lot of design
and standards efforts.
And so Scott Isaac put in a lot of those funky DHTML APIs
that didn't quite have the same flavor
as the stuff that I did.
And neither of them was like the later
sort of verbose Java like DOM W3C standardized.
But IE 4 was pretty darn good.
I remember a friend, Scott Furman and I
got invited by Scott Isaacs to Gordon Beers in San Jose.
They were doing a preview of IE 4, this must have been 1997.
And Scott said, yeah, we've got,
here's the new graphics stuff we're doing.
We've got something like your Netscape layers.
We've got VML, a vector markup language.
We can do like virtual reality
and Scott and I looked at each other
and said, we're doomed, right?
Microsoft was starting to fire on all cylinders.
So I have to give them credit for that
even though they abused their market power.
And maybe I shouldn't give them credit
for having the resource to hire talented people
but they did a credible job on IE 4.
What really was bad was that phase of the browser wars
ended with monopoly and perhaps due to the antitrust case,
perhaps due to regulation in Europe,
perhaps just due to Microsoft
not liking dealing with standardization.
They let it rot.
They just abandoned it, IE 5, 5.5, IE 6 later,
but these were not well-maintained.
They had a lot of security bugs.
- And it was really closed and outdated too,
even though it was getting updated, it's just weird.
- Browsers like Mozilla and then Firefox were adding tabs.
Opera had a version of tabs and they didn't add tabs
and they pop up blocking,
something I should have done from the start.
People realized that you can tell
when the user clicks something
and it goes in JavaScript to open a little window
that you can sort of inspect the stack
and see that the click originated that
and it's probably okay.
Whereas if you're just loading a script
and it opens a new window, that's a spam technique
and you should block it.
- Tabs were a brilliant innovation.
§14Transcript · 1:43:47–1:51:23
Like you said, Opera had it,
but I remember I fully switched to Firefox the moment.
I remember the moments of first using tabs in Firefox
and not liking it for the first few minutes
and then like, wait a minute.
- You get the groove, yeah.
- You get the groove and you understand.
So that timing, what year was this?
Because also as a aspiring web designer, I use table.
So we didn't mention layout or CSS much.
There's also a change in the way
like the frames were going away.
So there's a change in the way websites looked
and behaved and all that kind of stuff.
- CSS finally, which Microsoft embraced with IE4
and Netscape never really did, right?
CSS became a better standard over time
for doing table layout that relieved you of the need
to use what are called spacer GIFs, spacer GIFs, right?
Images you would throw into space out tables.
The typographic power of the web has gotten better,
but it's still not on the level of PDF
and you can't do advanced typography,
but it's gotten really better.
And even then tables were getting better.
If you were using Firefox, that would have been 2004
'cause it was called Firebird until earlier that year.
- No, yeah, I think it wasn't.
- 2003. - Well, I don't remember.
It was a Firebird, which had tabs.
- We had tabs the whole way.
So it started out as Mozilla/browser in 2002,
became Phoenix.
There's a BIOS that has an embedded version of IE
and they said, "We're called Phoenix Technologies.
"You can't use Phoenix."
And so we said, "Okay, we'll call it Firebird."
And then this Australian-centered
open source database project started really
like in the true Mad Max style, just screaming at us,
saying, "You can't use Firebird."
And I had to sort of be the ambassador.
I said, "Okay, we're gonna rename it."
And they're like, "We don't believe you.
"You shouldn't have used it.
"We hate you."
And then we renamed it to Firefox.
And they're like, "Oh, we love you."
And then I haven't heard of them ever since.
But Firefox was a clever name.
We had to think of something distinctive.
We wanted to keep the fire going.
And it turns out there's a red panda, right,
it's a nickname for.
- So that's the second set of browser wars.
- Second browser war.
- So how did you, how was Firefox born?
How's Mozilla born?
Is there a--
- There's a long story there too.
So Netscape got acquired by AOL,
which as I say was a reasonable happy ending
for a lot of people because Netscape otherwise
was gonna go out of business
because Microsoft was just killing its market.
There was no way to charge for a browser.
Windows came with IE.
IE 4 was pretty good and Netscape 4 wasn't that good.
It took a while to get better.
But the Netscape executive said,
"Let's do an open source escape pod."
And like in Star Wars and New Hope,
the gunner won't shoot it 'cause there's no life forms
on board, right?
It's not a threat.
And so we did Mozilla in 1998 and it looked like
it was going to initially just give the world
an open source browser.
But it's really hard to get people to work
on this sort of hairball that had been hacked up
over by that point four years.
It also couldn't have the crypto module
for secure sockets, so-called,
or now transport layer security.
That was an electronic munition.
We were not allowed to release that
in the full 1024-bit key strength.
And yet people, one of whom I happened to meet previously
at SGI when I went on a sales support engineering trip,
Tim Hudson in Brisbane, Australia,
and Eric A. Young did what became OpenSSL.
It was called SSLEAY after Eric's initials.
And Tim and Eric took their OpenSSL
outside of the purview of the NSA
and the Department of Commerce,
and they stuck it into Mozilla's code.
And that was perhaps the best hack that was done
in the first few months after we open sourced the browser.
We had other problems.
The politics inside Netscape were riven
by these acquisitions.
So the one acquisition that kind of messed up Netscape 4
also wanted to keep doing a proprietary mail
and groupware program, not Jamie Zawinski's mail program
that was in Netscape 2 and 3.
And they held it back from open source.
So we didn't have a mail program, it was just a browser.
We didn't know what AOL would do to us.
Turns out they didn't interfere with us for a long time.
But Netscape wasn't the best steward of Mozilla.
We were operating Mozilla as a pirate ship
without a legal entity.
So most of us worked for Netscape
under a separate organization.
And initially the first engineering manager, Tom Paquin
of Netscape was the Mozilla founding manager.
But he left pretty quickly.
And he left me as the acting manager,
which is more like method acting in my case.
And that was my first management stint.
But then someone who'd written the licenses, Mitchell Baker,
she was a lawyer at Netscape.
She was involved in the open source license decision-making
and the actual writing and construction of those licenses.
That was Mitchell's job.
Netscape public license
and the truly open Mozilla public license.
And there were two because Netscape needed,
because of some encumbered code, needed some special rights,
but that went away over time.
Mitchell was always interested in Mozilla
and she came back from maternity leave.
And she said, "I'll be the manager if you want."
And Jamie and I said, "Sure."
And then Jamie quit.
He quit after a year.
He said, "This didn't work.
I'm sorry."
He acted like it was a total failure
because Mozilla didn't restart the browser market.
But there's no way it could have.
Netscape was still shipping variants of Netscape 4,
which was based on the old code.
Mozilla was trying to re-architect code
to make greenfield for developers.
So it was one of my big goals.
It wasn't a technical goal so much as, again, a social goal.
People wanted a more standard spaced browser.
They wanted less of a hairball that had been hacked on
by ex-grad students starting four years prior.
So we said, "We're gonna make a modular code base.
We're gonna use a variant or an open source version
of Microsoft's component object model,
has reference counting and standardized V tables,
virtual calls and C++.
And we're gonna use JavaScript.
We're gonna have a bridge between those two
so you can script those components
just like Java components.
We're going to make a portable front end
with a markup language for the user interface.
Not tables, not HTML, but custom menus
and dropdowns and toolbars."
And that was called Zool, XML user interface language.
And some real talent on the Netscape side
delivered that, Dave Hyatt, who was instrumental in Zool,
Chris Watterson, Joe Hewitt, Blake Ross.
And Blake was an intern.
He was like a high school aged intern at Netscape.
And at some point we were innovating rapidly
in the Mozilla world and Netscape was still caught up
in this management mess from these acquisitions.
And it wasn't delivering.
And every year they were wondering if AOL was gonna come
and start beheading the executives
'cause it didn't do anything useful.
And there was this thought,
you should take the Netscape browser engine
and put it in the Windows AOL client,
which was the dial-up client that all the increasingly aging
users of AOL were using.
Never happened.
It would have been too big a change.
So it wasn't clear why AOL bought Netscape,
but as I said, they left it alone,
but Netscape didn't leave Mozilla alone.
And so in 2001, Mitchell called me up and said,
"I'm no longer employed."
And I was like, "What, you quit?"
And no, no, this wasn't my choice.
And there was a layoff, which maybe accidentally
or on purpose got rid of Mitchell.
But the funny thing was we had an open source project.
We had a lot of the engineers on staff on our side
and we had people we'd hired through the Mozilla community
who were top-notch.
They'd risen, they came in high quality,
they knew the code, and they actually were better
than the average or median hire of Netscape.
And so the funny thing was the executive
§15Transcript · 1:51:26–1:59:28
who thought they'd gotten rid of Mitchell on the layoff,
on the next week's community call around Mozilla
and what to do, there's Mitchell.
And so this showed you can kind of transcend
your boundaries of corporate open source
if you get a project that has enough loyalty,
even among the paid staff.
'Cause we had outside people contributing.
We had people at Red Hat and a few other places,
but the majority of the hackers were employed by Netscape.
But a lot of them at that point had come from the community
and others got the community and wanted to work with it.
And it was really the weakest engineers at Netscape
who didn't like Mozilla and didn't like the crucible
of competing with the better programmers.
- So if the project is good enough, it will rise,
the Phoenix will rise out of the--
- That's exactly right.
And so we had this Mozilla code base
that was getting better.
In fact, I think at some point in 2002,
when we declared Mozilla 1.0, I engineered a roadmap
that successively through similar sort of six week,
five week releases, like we all do
with browser releases nowadays,
Chrome does and Firefox braved us three weeks.
We got to a point where we said, you know what?
It doesn't suck.
This is like the 1.0 that you want to release
because if you hold it back any longer to polish it,
you're denying others the ability to use it.
It's like Pro Engineer, the mechanical CAD tool,
embedded the code, they embedded the layout engine.
And Mozilla 1.0 was like a Netscape communication suite.
We had at that point gotten male people
to reintegrate mail and news,
and we had an editor for HTML.
And it felt like a 90s suite, suite where.
And it felt kind of bloated.
And the people who were taking that Mozilla open source
and then adding Netscape flavor to it
were not calling the shots right.
And they were also under ALS thumb a little bit
in that they said, well, we should probably put
the ALS instant messenger chicklet on the toolbar.
And we should put the ICQ,
the other messaging system that ALS had acquired.
We should put the ICQ button on the toolbar.
And pretty soon Netscape looked like a bit of a NASCAR
badged version of Mozilla.
And that also made Mozilla more popular.
And yet they had contrived to fire or lay off the leader.
And we'd carried on with an open source structure
where Mozilla was still, you know, Mitchell was calling
sort of management or project level shots.
And I was calling technical shots.
And we had a popular suite, but we thought,
why not make it just a browser?
Because it'll be simpler, it'll do one job well.
And even then we can strip it down by having extensions.
So Dave Hyatt and Blake Ross, the high school aged intern,
did the first version, which was called Mozilla/browser.
It was a small group of us, Ian Hicks and Asa Dotsler,
me and Joe Hewitt and Hyatt and Blake.
And Hyatt was really the senior hacker.
He'd done all these things like amazing cross-platform menus
through the user interface, market language.
And he knew how to do tab browsing.
He'd implemented it natively on Mac OS at the time
in Camino, originally called Camara.
He'd written multiple implementations,
which was a thing programmers should do.
It's like the V8 team did for those missing years
when the rest of the Chrome team's like, where's V8?
In fact, Dave's wife, Rebecca, told me a story
about when they were at UIUC,
they were also University of Illinois grad students.
There was an assignment, it was a programming assignment,
it was supposed to be due at the end of the semester.
And Dave's friend was this, I'm gonna go think
and I'm gonna design and I'm gonna make this platonic,
perfect form of the program.
And then I'm gonna write it at the end when it's due.
And Hyatt just went in there and started hacking.
And he wrote one version, he wrote a second version,
a third version, end of the semester comes around,
the friend's not doing too well.
It wasn't perfect and it wasn't written.
I'm not sure how that story ended for him,
but Dave's version was a fifth iteration, it was great.
And so he'd done that with everything you need
in a tabbed browser.
And this really showed well in Phoenix,
what we called Phoenix and had to rename two more times.
And Blake went to Stanford, he became a Stanford student
and couldn't work on it.
Dave Hyatt went to Apple in 2001.
He was one of the founding Safari team members.
- Interesting, wow.
- But he was still blogging about tabbed browsing.
I think Apple at some point said--
- Did Safari have tabbed browsing?
- Yeah, but it was because of Hyatt.
Hyatt was quite a feather in their cap.
Don Melton, who had been the engineering manager
for Safari from the beginning, had been in that scape also.
And so this is a diaspora of talent.
And yet Hyatt was still kind of writing blog posts
about how to do tabs right.
And at some point Apple said, "Don't blog about that.
"That's our proprietary tab technology."
And it's like, "No, it's not.
"It was an opera and I've refined it."
So we had to replace people.
And we had Ben Goodger, a New Zealander,
we hired at Netscape, and he stepped in
to be the Firefox lead.
And we also had this weird circumstance
where AOL finally did notice that Netscape
was kind of an albatross, that they bought it
for no particular benefit.
And even then the AOL politics were also heinous,
sort of East Coast politics.
I remember taking two trips there
'cause I was a principal engineer.
And so us principal engineers got trotted out
to do dog and pony shows in Dulles, Virginia.
And the AOL opera management was very East Coast in flavor.
And they were at that time merging with Time Warner,
which did not go well.
So one of these years we went out there
and we were all doing dog and pony shows.
And there were these characters
that were sort of like marketing guys,
one of them was wearing a cravat,
and one was named Reggie.
And they were very you rather than non-you,
or they were like what's Stillman's Metropolitan film,
UHB, urban haute bourgeoisie.
They were haute bourgeoisie, they were funny,
and they were kind of useless and kind of preppy.
And then the next year we went back and I said,
where's Reggie?
And it's like, oh, Reggie's not here anymore
'cause Time Warner realized that the merger
wasn't in their interest either.
And then the sort of knives came out.
And these mergers rarely work, right?
This is very difficult.
You get these giant companies
and they think there's gonna be synergy.
That was the late '90s with Watchword.
And there wasn't synergy with AOL buying Netscape,
and there wasn't synergy with Time Warner and AOL.
- But did AOL ever really work?
Was it ever really cool?
Like the same kind of fire and excitement
that Firefox eventually created,
was that ever there in AOL?
- AOL was the right time to do a dial-up service
that got distribution by basically
leaflet-bombing compact disks on the country.
And they beat out CompuServe and the other ones, Prodigy,
and then the web happened.
And so you had almost like this isolated continent,
like some of the evolutionary biologists I follow
make fun of the funny large marsupial mammals of Australia,
how silly they are.
- So AOL's like Australia.
(Lex laughing)
- And you saw it over time because they kept aging
and they were using AOL to get online
and they couldn't really use a web browser.
And it became sort of a valued cohort
'cause they still have relatively high socioeconomic status
and they have grandchildren, but it's going away.
It's dying at some point.
- Towards the end of the aughts, that decade,
and then to the decade 2010 plus,
that Firefox became this incredible,
I forget when Chrome came out, but--
- 2008, September.
- 2008, but Firefox was the sexy, cool thing
that represented a lot of the cutting edge technologies
and all that kind of stuff.
- Web 2, it was amazing.
Tim O'Reilly and John Battelle did the first Web 2
conference, which eventually became huge
and they split it, but that was in 2004.
That's right when Firefox was out.
Craigslist was huge, it was killing classified revenue
for newspapers, but there was just this ferment.
People starting--
- Wikipedia along there somewhere?
- Gmail was already done and it was an impressive web mail.
There were others before it like Hotmail,
but Gmail was really impressive from Google.
And Google Maps, people started seeing what could be done.
They thought, how can you drag the map around
and how does that work?
§16Transcript · 1:59:29–2:08:07
And it was all JavaScript and images.
- Gmail was 2003, four?
- Yeah, I think it actually started quite early.
It might have been 2002 or three, but by the time
we started dealing with Google and Firefox
to get the search deal, which was the main revenue source
for Mozilla, and still is, 2004, early,
Sergey Brin's, one of his trusted engineer guys,
Fritz Schneider, made contact with me at Mozilla
and we started talking and we realized search
and browser need each other.
And this is deeply true, right?
This is still true.
This is why a lot of the search engines
have their own browsers.
- Yeah, so in case people don't know,
the main revenue source for the browser
is the default search engine, which is kind of incredible
to think about, that that is a revenue source.
It's a little bit sad.
- Yeah, it leads to this capture or kill effect
where you have the search engine own its own browser
and other browsers may struggle to get distribution,
we talked about it earlier.
- So where, and you said you've figured out
that Google's working on its own browser at some point.
- 2006, yeah.
- 2006, so would you say Firefox versus,
was Internet Explorer part of the war here
or was it Firefox versus Chrome?
- So Firefox didn't quite cause Microsoft
to reconvene, they did do IE7, and I remember
being on a plane back from the standards meeting,
JavaScript standards meeting from Seattle, from Redmond,
and there was some Microsoft guy in front of me,
turns out my wife knew him from her past life
before we married, and he was just this bearded big guy
and he was like, we should have just killed Firefox
in the cradle, all we needed to do was add popup blocking
and tabs and we could have made Internet Explorer
kill Firefox, and it's like, shoulda, coulda, woulda, pal,
and I was right behind him overhearing this.
But they didn't, they were slow and IE7 wasn't that great,
and what really got them started, I think, was Chrome.
And I talked to Larry Page in 2005, I think I said,
we're talking about the Firefox relationship,
but he was also saying, what about WebKit?
This was Apple's version of the old KHTML engine
from Linux, the KDE side of Linux, that was used
in the Conqueror browser, also with Ks,
that Apple had forked, and in 2005 was when Apple's
principals, including Dave Hyatt, Maciej Stachowiak,
some of my friends who are still there, said,
we must stop patch bombing this poor KHTML project,
we should make a proper Mozilla-like organization,
WebKit.org, now it wasn't a separate non-profit
or anything, it was still Apple, it was Apple-controlled,
but they made their fork first class and they made it
be something that they all worked in and lived in,
and that was before Chrome, and then Chrome,
Larry Page said, what about WebKit?
I said, yeah, it's nice, I have friends who work on it,
you might use that if you do your own browser,
why don't you do your own browser?
Don't worry about Firefox, you should do your own browser,
you can have your own opinion of how it should work,
and sure enough, they did, so by 2006, we knew
they'd been working on it, some of my friends
who'd been at Netscape did the original demo,
and the demo wasn't what you thought,
it didn't have the fast JavaScript yet,
that was still off in Denmark on a farm.
- Did it have tabs?
- It had tabs, 'cause all browsers had tabs at this point,
and it had this software fault isolation I mentioned,
it was through process isolation, so in theory,
each tab has its own operating system process,
and so what's gonna take your tab down?
Well, WebKit has bugs that can crash it,
but Flash was still big then, all the restaurant sites,
remember, and Flash crashed a lot, so the demo
that I heard about, my friends at Netscape,
as a lot of people did, inside Google was the sad tab,
they showed an early version of Chrome,
which is just this bare bones tab browser,
they loaded a site with a known Flash vuln,
and then suddenly, Flash crashes, and everyone expected
the whole browser to go down, but instead,
you got this little sad face in the tab,
and you could reload it, and there it is again,
so this was an improvement, it was a real move for security,
it was based on a company they acquired called Green Border,
they had some really big brains, like Olfar Erlingson,
I think was involved, and they had done some exotic
security stuff, but they ended up simplifying it
to this process isolation, and it was good,
and Firefox didn't have it at the time,
so we were still struggling with security bugs.
So we knew Chrome was coming, but it took two more years
to come out, and we were still getting the Google search
revenue, and we were still making Google the default engine,
and Firefox was still growing, Firefox grew,
I think until 2011, that was when it peaked,
and as it started falling, it was because of Chrome.
Chrome came out in 2008, and it was a comic book
that leaked accidentally, that showed some of the people
who worked on it, Lars Bach was in there, and so on.
It was kind of soft launch, because they didn't market
it heavily, they didn't push distribution,
but Google had reason to worry about distribution,
because Microsoft was doing a search engine,
Bing, since 2007, in fact, when they came out with Bing,
Google was worried that Microsoft would just brute force
switch the default browser in everyone's Internet Explorer,
or even Firefox on Windows, to Bing from Google,
and Microsoft wasn't, I think, ready to dare
the antitrust cops that way, even though
they'd gone to sleep, and I don't think
Bing was ready, either, but just in case it happened,
Sundar Pichai, who rose very well based on this work,
was sort of in charge of getting distribution deals,
and he got Google Toolbar, and Google Desktop Search
distribution, and if you remember those pieces of software,
those were like desktop extensions, toolbars,
or operating system extensions for doing desktop search
in your local files, kind of like--
- Oh, that's right. - Mac OS Spotlight.
- Sadly died out. - It all died,
and there were some features that we still missed
that didn't make it into Chrome,
but Sundar got OEMs to bundle those,
and then he got enough of those deals
that by 2007 or '08, Google felt, well,
if Microsoft does the worst and tries to force Bing,
we can reach in and reset it with that point of presence.
So that was good for Sundar's career,
and it was good for Google, but it never came to pass
that they had to defend.
Microsoft was still slow, and by the time they saw
Chrome come out, then they did what would have been IE9,
and then they said, we're gonna have a fast
JavaScript engine to Chakra, Chakra Core,
and they did okay.
They were another process-isolated,
fast JavaScript browser, tab browser.
- So it sounds like there's a deep,
fundamental coupling of search engine and browser
that's mixing this whole thing up,
and obviously Firefox doesn't have a search engine
that's like, I mean, you're partnering with somebody
with a search engine, with Yahoo or with Google or so on.
- They tried Yahoo.
That was unfortunate because I think,
even though Marissa Mayer talked about it,
she never pulled it off.
They never restored the search team
that had been laid off.
I believe Carol Bartz was running Yahoo
when Carol said, I've gotta get rid of
one of the three expensive things.
I'm gonna get rid of search.
And those researchers went to Google and Microsoft,
and there was no way to put Yahoo search back together.
So when Firefox tried switching all their users
who'd stuck with a default from Google to Yahoo,
like mid-December 2014, a bunch of users said,
what just happened to my Firefox?
And others didn't notice right away,
but over time they did, and so over the next year,
the traffic just went away for Yahoo,
and yet they were obliged, I understand it,
I don't have inside knowledge, but this is leaked out,
and Danny Sullivan's written about it,
search engine, and I think the deal
was like fixed payments to Mozilla.
So Mozilla was getting a bunch of money for traffic
that wasn't staying because users were resetting
their default, and this shows how defaults are important,
but they have to be good enough
that the user doesn't override them.
And a lot of the commercial value in popular apps
is what are the default settings?
What is the default search?
- But oftentimes there's something, just like you said,
I mean, if there's something compelling,
that also can beat out the default,
like tab browsing and so on.
And that's where, I mean, we'll talk about Brave browser.
It feels like now we're in this third stage
where there's Chrome, Firefox, Edge, I guess it's called,
and Brave, and these all seem like really exciting,
I don't know, innovative browsers.
They're all kind of copying off of each other,
picking up the good stuff.
- There's evolution again,
especially on tracking protection.
So privacy is this sort of global wave that's rising.
I like to call it a wave
because it's a large, somewhat chaotic structure.
It's not a unitary good.
§17Transcript · 2:08:08–2:15:34
You can't say I'm buying privacy for $3.
I'm paying $3 for privacy.
Some people think a VPN does this
and are disappointed when it fails them,
but often people use VPNs for region unlocking video
or getting the US Netflix catalog.
- Exactly.
- But privacy is not a unitary good.
It's complex and people are understanding it
only over time and as they get burned,
but there's a genie
that's not going back in the bottle there.
People are fed up.
Apple has responded to this.
Apple was always making Safari, I think,
more of a privacy-branded browser from the very beginning.
I think this was probably Steve Jobs.
Safari had private windows, private tabs,
before Firefox did.
And these are only private in the sense
that they don't leave local traces
if you don't want them to.
Turns out Safari does keep them around between shutdown,
but the canonical model is no local traces
after you close the private window.
No leftover traces that you went to some site
that you were embarrassed by
or bought a gift for somebody you wanted to keep secret.
- But there's still some level of tracking.
- There's network tracking.
Network privacy is not guaranteed at all
because you're using the same internet and ISP
as a public window, a non-private window.
But Safari had that early on.
They also had a cookie-blocking policy
that might take a little explaining.
If you know what a cookie is,
it's a little bit of storage in the browser
indexed by the name of the site.
And it's really only the main name of the site,
like bofa.com or something like npr.org.
Every site can store some information in a cookie.
Every time it's contacted by the browser,
the previous version is sent back.
And in the response from the server, the cookie's updated.
So it's this little bit of storage in the browser
that the site can keep updating,
and it can store an encrypted version
of your login credentials with a timestamp
so you can stay logged in
without having to retype your password
every time you navigate,
which is how it would be if you didn't have cookies.
The web protocols, especially in the '90s,
are so-called stateless protocols.
So you go to your bank, you log in,
you go from your login confirmed page to your account view.
If you didn't have a cookie, you'd be logging in again.
- Every time you type in the password.
- So that was a great thing about cookies.
Lou Montuoli did it in a hurry in 1994
before I joined Escape,
and he did it for really holding that kind of credential.
But even then, there was the image element
embedded in the page,
and the image gets fetched possibly from a different server,
and that request carries the last cookie,
which could be empty at first,
and the response carries the updated cookie.
So just by having images and cookies,
you got tracking because that image server
can be serving a little one-by-one pixel,
and they still use the word pixel in ad tech,
and that pixel can be served from the same server
embedded differently with different URL spellings
in the New York Times and ESPN.
And as you go from one to the other,
the image server can say,
I haven't got a cookie for you, it's empty initially.
I'm gonna assign you user number 1234,
I'm gonna put a database entry in,
and I see, by the way, I always fetch
the name of the path part of the URL
that I was in the New York Times.
So you're a New York Times reader.
And then you hit ESPN, same thing,
and the database gets updated,
and the number, user 1234 indexes in the database
to a profile of you, you've been tracked.
This was not intended,
and it was too late to undo by the time I got to Netscape.
I think Lou wanted to do Twinkies, he called them,
and he was trying to solve several problems.
He wanted them to be bigger
'cause initially cookies had a short size limit.
I think he wanted to solve the third-party problem,
but Tom Paquin, the engineering manager said,
"Nope, no Twinkies, just cookies.
"We're done, you're done, son."
And that's how a lot of that stuff was.
That's how JavaScript got frozen like a flying amber
in some ways with that sloppy equality operator
that I made 'cause of the early adopters.
And the cookie got stuck with this tracking hazard,
and then, because JavaScripts can be like images,
they're embedded in the page.
By the time of Netscape 3, I made that work.
You can get a request with the last cookie value,
and the response updates it.
That's a tracking mechanism,
and that's why you don't even need images to track.
Now you just use scripts.
So this whole tracking economy evolved,
and it depended on these accidents of the '90s,
these unintended consequences.
- Well, it created some of the richest companies
in the world, right?
I mean, it's the social media--
- All I got was T-shirts.
(laughing)
- All I got is this crappy T-shirt.
Yeah, I mean, so that's the fundamental problem
the world is facing now.
They're looking at what social media has created,
and they're looking at,
and like a world is looking at itself in the mirror,
and seeing that privacy is actually something,
it's supposed to be like a nice thing to have.
It's something that is, actually should be fundamental
to the way we interact with the world,
as part of our tooling,
and that's where the Brave browser comes in.
And I suppose others as well are playing with this idea,
but Brave is at the forefront of that.
So maybe can you describe what Brave is,
and what are its key principles,
and what's broken, and what is it Brave trying to fix?
- So when I realized that these accidents,
like the third-party cookie,
the image or script that's tracking you,
or the JavaScripts that can do it invisibly now,
that all this stuff wasn't intended,
and that Firefox had supported extensions
that block some of these things,
I thought, probably we should have browsers
just block some of these things by default.
These were not intended, and they're now unsafe.
They're tracking you, there could be data breaches,
malware distribution, bullying, and PsyOps,
and other attacks on people.
Block that stuff, block that JavaScript.
I'm Dr. Frankenstein, I've got to deal with a monster here.
But obviously you go to Gmail,
there's so much a script there
to make that amazing web client,
that's okay, that's first-party JavaScript.
So how do you tell the first from the third party?
And it's not easy, it's not a matter of just
what's embedded from a different server,
because a lot of publishers use benign scripts
from unrelated domains, or apparently unrelated domains.
So you end up having to develop a sort of
human and machine learning practice around blocking.
And at Brave, we did that from the start,
and built a research team to help drive it and automate it.
We realized that protecting people needed machine learning,
and around 2017, in spring,
I talked to my friends at Apple about this too,
and they were also doing what they call
intelligent tracking prevention,
which uses local machine learning in the browser.
And the funny thing is, you know,
great minds think alike,
they were taking their third-party cookie blocker
that was in Safari from the old days,
and making it not have a big loophole.
Because what they did was, in 2003,
when Safari came out, they said,
we're gonna block cookies that are from
those third-party embedded elements,
where you've never visited that site before.
So I'm gonna pick an ad company that got sold to AT&T,
so I'm not picking on anybody unfairly,
appnexus.com.
Have you ever been to appnexus.com?
- Nope. - I've never been there.
But I guarantee you, 10 years ago,
you probably had, if you were using Firefox,
you had a cookie, third-party cookie,
'cause you were being tracked by them.
And they were using that cookie
to build up a profile of you.
In Safari, as long as the user never went to AppNexus,
that cookie would not be set.
And that was a real move for privacy early on,
when jobs were still around, in Safari.
But it had this loophole that if you do go to AppNexus,
then why, it's okay to be a third-party cookie.
So AppNexus did something very naughty.
They took their ad partners,
§18Transcript · 2:15:36–2:23:05
that put the actual ad you click on,
and they said, "Hey, add a little script
"so that when somebody clicks on the ad,
"before it goes to your landing page,
"redirect to AppNexus,
"and we'll redirect to the landing page."
And by doing that, they set a first-party cookie,
and they got whitelisted.
So it was a loophole they exploited.
Intelligent tracking prevention in Safari
was sophisticated enough to counteract this,
and it did other things,
and it's evolved since they did it.
And we've evolved brave, too.
And so when I say machine and human learning,
there's a real set of techniques here.
They have to fight--
- This is a fascinating problem, actually.
- Fingerprinting, right?
Anytime you have a little bit of storage in the browser
associated with a website,
if the bad guy can get 32 websites,
each one has a bit of storage that's 32 bits.
You can turn the bit on or off.
You can make four billion numbers.
You can make an identifier.
It's called a super cookie sometimes.
There are weaker ways that are statistical.
They're called fingerprinting.
You have to block all of them,
and you have to not only automate,
you want to work in the web standards body
to put privacy in by default, by design,
from the get-go, not add it as an afterthought
or go hog-wild with new web APIs
that add a bunch more local storage
or fingerprint surface area.
And that's been a struggle, too,
because guess who's the new Microsoft
in the standards body?
It's Google.
And they're not in favor of privacy first.
They wanna do privacy their way,
only under, I would say, market pressure.
But with Apple and with Brave leading the way,
we block third-party cookies almost without exception.
So we just block them,
and that gives us a very strong privacy benefit,
but it also means some sites just don't work right.
Embedded YouTube videos might not work right.
So we're adapting in a similar way to Apple's done with ITP
to make third-party cookies blocked,
but to sort of simulate what looks like
a working third-party cookie for the site.
It essentially tries to partition each site
and its third parties into its own sort of cookie jar.
- Got it.
And so, like you said,
is this both a human fine-tuning issue
and a machine learning problem?
- Yes, and as the humans learn,
then they train the machine learning.
- But maybe Google aside, or including Google,
there's millions of dollars, if not billions of dollars,
to be made from fighting the ways of Brave.
- That's right, and it's been an interesting change
from when we started in 2015.
When we started, ad-blocking extensions,
AdBlock Plus was one of the big ones
that started on Firefox in 2006, I believe,
had gotten to a certain level of use around the world,
and browsers like UCWeb, UC Browser in Asia,
had some amount of ad-blocking built in and on by default.
So PageFare was a startup,
and they measured ad-blocking adoption,
and they tried to say, "Hey, publishers,
"30% of the visitors to Pitchfork or Wire,
"to Conda NAS properties, are using ad-blockers.
"If we can somehow convince them
"to lower their ad-blocking for your site,
"that could be like a 43% lift, right?"
And 3/7ths, well, that's easier said than done,
and PageFare and others, SourcePoint, many others,
tried to either smuggle ads through
or cajole the user into letting ads appear,
and it didn't really work.
Meanwhile, the ad-blocking adoption
has just continued.
Intelligent Tracking Prevention, Safari in 2017,
Brave from 2016 on with very strong cookie-blocking
and other protections, and this is not going away.
The publishers used to rage against it.
We would try to say, "We can help you.
"You're dealing with users
"who are already blocking all your ads.
"We can try to put back some economics
"that help the user and you,"
that led to the basic attention token
that we started with Bitcoin.
"We can be your friend.
"Don't just fingerprint us as an ad-blocker
"and treat us as an enemy."
But in 2015 or '16, it was like,
"Nah, you're an ad-blocker.
"Get out of here.
"I hate you."
And by 2017 or '18, it's like something's happening.
The ad-blocking's not stopping,
and we're all getting sort of pulled
on the Google's plantation through AMP, A-M-P,
or we're getting killed by the Google ad system we use
because it's taking all the revenue,
or it's permitting, or some other vendors we use
are permitting ad fraud.
And so a fake New York Times is getting paid
by the marketer running an ad that a bot clicks on,
and the real New York Times
that's supposed to get the ad doesn't get it.
- And there's something really broken
about that kind of system.
- And that fraud is mediated through Google's ad exchange,
which is the biggest of them all,
and Google takes a fee.
There's a flip side of that,
which is malware distribution, malvertising,
where fake advertisers put malware payloads
in exploit kit loaders in JavaScript,
and they smuggle them in ads onto real publisher pages.
The ad exchange takes the fee.
Now, I'm not a lawyer.
I'm not gonna say this is a RICO predicate,
but why is the ad exchange facilitating fraud
and malware distribution and taking a fee?
It's not right.
- As opposed to just fighting,
this is the really interesting thing about Brave,
is as opposed to just fighting
and then being treated like an ad blocker,
you're providing an alternate.
There's a philosophical idea here
that might change the nature of the internet
with a basic attention token.
- Yes.
- Maybe what is basic attention token, BAT,
and how does it work?
- Okay, I'll tell the story first by saying how I came to it.
I realized for a long time at Firefox,
we were dependent on this Google search deal,
and I thought, now that Chrome's out,
maybe that's gonna go away.
At some point, Google will say,
Firefox, like old yeller, you saved me from the rabid beast.
Now I have to shoot you in the head.
Done your job, sad but true, goodbye.
And what can we do?
And I think Mozilla doesn't know what to do.
This is something that I couldn't solve there,
and I don't think they can solve.
But I thought, why is the browser
the sort of passive servant of these big tech companies?
Why is it a blind runtime for ad tech JavaScripts,
including from Google?
Why doesn't it block some?
And if it blocks some, why can't it reconnect users,
readers, fans, with publishers, creators, websites?
Why can't it help people make direct payments
or even possibly get an ad revenue share
for private ads that are placed in the browser?
The ads are all placed in the browser.
Some people have this sort of model
that the server's painting the ad into some flash,
combined package or into some giant image,
and then it all gets sent down.
That's not how it works.
All the ads you see on the web are placed in your browser
by it calling out to various ad tech partners,
and Google's among them.
And so if you block those scripts,
you break the advertising flow of money
from the brands and their agencies to the publishers.
And if you want to reconnect it directly with the user,
you have limited choices.
The user generally isn't gonna sign up
with a ACH bank connection or a credit card.
The publisher isn't gonna sign up the user,
except as a subscriber,
and then they're gonna overcharge you
because they want you to cross subsidize all the content
and buy more than you read and all that stuff.
And how many, you know, people are doing great
who are big names like New York Times and The Washington Post
but how many subscriptions are you as a user gonna pay for?
This is why startups like Tony Hale's Scroll
are trying to do a portable subscription system.
- By the way, just on a small tangent there,
even the New York Times is really annoying
how difficult it is just to subscribe.
There's way too many clicks.
- They don't make it easy.
And I had friends a few years ago, I think they fixed this,
§19Transcript · 2:23:06–2:30:55
who would pay for the paper and then they'd go online
and they'd get up charged for the digital
and there was no break, there was no connection between them.
But publishers are not that technical
and they can't all get you to subscribe,
you can't have a thousand subscriptions.
So for a long time people talked about micropayments,
there was Blendle and the other ones which came to the US
but didn't grow.
And I thought if you have just a browser
and it's protecting you
by blocking all this ad tech tracking junk,
it can provide you an option that uses cryptocurrency
to let you support your favorite sites
and even YouTube channels.
And that we prototyped with Bitcoin.
And that meant the user had to be of means to contribute
and willing to contribute.
But it could be done on the Bitcoin blockchain
and it could be fairly efficient
even though Bitcoin went through a period
when we had this prototype running in 2016 into 2017
where Bitcoin was very congested and very slow to confirm
and the fees got very high.
And a lot of users who were not Bitcoin maximalists
or even experienced, we helped them out
by embedding a Coinbase buy widget
and they had the income to buy but it was hard.
It was like, do I buy $5 a month?
But the fee is like 450.
I better buy in larger batches, right?
And they're like, I don't wanna own that much Bitcoin.
So it became this painful thing.
And the real idea that I had of private ads
that pay the user a rev share couldn't be realized
alone in that kind of system.
In these cryptocurrency systems,
especially with the blockchain we switched to Ethereum,
you can have smart contracts.
The Bitcoin system is not Turing complete
so what you can do with the script is more limited.
But you can still do sort of clever things
even with Bitcoin script.
What we wanted to do was sort of a three-sided ecosystem.
We wanted users, creators or publishers and advertisers.
And we wanted the advertisers to put money in
just like they do today but without going through
the Googles and the app nexuses
and all these other ad tech companies.
Because those companies take out a huge cut.
The Guardian in the UK once did an experiment for a month.
They bought out their own ad space.
They put in a pound and they were paid 30 pence.
70% was coming out to the intermediary vendors
they were using.
- Wow.
- And that's like the opposite of what the app store does.
The app store takes 30% and gives the publisher 70%.
So pretty broken.
In the old days of the Superstation TBS,
the media owner would get 85%.
So these splits have become really unbalanced.
And the middle players, the ad tech vendors
are taking out way too much money.
And they're doing something worse which has been noticed.
They're letting not just the malware vendors
but also the ad fraud side which fakes the publishers
and clickbait merchants come in
and steal traffic from good sites.
Because once you have a certain audience
identified at one site.
Jason Kalkanis told me this about his experience
with I guess it was in Gadget.
Which site he was running.
Once he started using an ad partner
that was sharing his audience information
across multiple sites,
he saw his competitors stealing all his traffic.
And then what's worse is the clickbait sites
that just have much cheaper rates steal all that traffic.
And that facilitates fraud, it facilitates fake news,
all sorts of problems.
So Grape blocks it and then we give users
the ability to give back.
And because we invented the basic attention token
in Ethereum, we can do this three way split.
And we can give users a share of the revenue.
And if they want to take it out, they can.
Now unfortunately for us and for all of blockchain,
the regulators are saying,
we're gonna have to know who you are.
There's the Treasury Department's FinCEN agency,
there's the Office of Foreign Asset Control's OFAC,
there's the other regulators in the federal government
that take a very dark look at things like money laundering
and sending money to someone named Osama Bin Laden.
So compliance starts to come in.
And even now they're threatening for pure Bitcoin
sending to some address.
If you're a Coinbase, you're gonna have to know
who's at that address.
- Like the actual identities of people involved.
- Now with Coinbase members, you sign up and they know you
and they comply with the regulations.
They're a regulated money services business.
But if somebody's using their own self-custodial wallet
where they have the hardware private key
and they're not named and they want to send to that address,
our friends in the federal government are talking about
requiring at some threshold knowing who that is.
- Some threshold that's unreasonable?
- It's not that big.
Yeah, I don't know how this will play out.
I think crypto is here to stay.
I think the beauty of being able to send peer to peer
without any bank in the middle,
without any huge wire charge and two day delay
and all that nonsense, it's beautiful.
And I've used it and I love it.
But we're pragmatistic brave about crypto
and we realize that anything like a revenue split,
we can't facilitate without being licensed in a certain way
and it requires knowing who the user is.
So our default mode doesn't know who the user is.
It instead imputes to the user's browser
some of the revenue and allows that browser
to steer it back to the creators.
And we do have to identify the creators.
But as things improve, and who knows how it'll play out,
there should be a day when this full vision can be done
more fully on a blockchain.
But regulations and the practicalities
of today's blockchains, which are not that fast
and not anonymous over time,
you fingerprint yourself over time,
we do some of this with the browser.
So one of the ideas of the basic attention token
is to make a hybrid system that's stronger
than blockchain alone.
It's the browser and the blockchain.
And the browser is this trusted endpoint software.
It's this universal app.
Everyone uses browsers.
The bigger the screen, the more you're in the browser
unless you install fat clients for things.
I use Slack on Mac OS and it's like a browser.
It's based on an electron framework we used to use.
And it's just, it's not that great.
Some of the people at Brave use Slack in Brave as a--
- In the browser, yeah. - In the browser, yeah.
- I use that often, yeah.
- And I noticed on the iPad, I use apps less.
The smaller the screen, the browser got handicapped
by Apple and Android both.
And it also can be slower or not have the right affordances
to interface with the security limited APIs.
But in principle, with the right permissioning,
you can make the web browser just as good as any app.
You make it be a super app.
And that's part of our mission at Brave.
So we want to have the economics that got captured
by these big tech companies through tracking
and through social networks.
We wanna block that for your own safety
and then let you opt into a cleaner world
where you keep your data defended in your browser
and you can actually realize value from it.
So the way our ad system works,
I mentioned it being private,
but how does that work?
We don't see your data at all.
All browsers are sort of the mother of all data feeds.
Your history, all your searches at all engines.
Each engine sees the queries you send to it,
but it doesn't see the others.
But the browser sees them all.
Machine learning in the browser that you can opt into
can study all that in a very complete way
and do a better job than Google does.
Google has cookie and scripts across the web
from acquiring DoubleClick.
They have YouTube, they have Android,
they have Search, which is still their big revenue lane.
But they don't see everything.
The browser sees everything.
And if it can do a good job locally,
and this is not advanced machine learning,
this is not TensorFlow, this is like SVMs now,
and Naive Bayes,
then you can match intense signals,
intense signals from those data feeds,
the searches, the queries, the history,
how much you're scrolling down a page,
how much you redid a search.
It's all blind browser algorithm.
We don't see that data.
And then pick the best ad from a fixed catalog per day.
And the catalog is fixed across a large population per day.
And it only updates once a day
'cause new offers come in and old ones expire,
§20Transcript · 2:30:57–2:38:12
sometimes every week or every month.
And that catalog, and there can be many such catalogs,
is sold by our direct sales team.
And so we're making an anonymous audience
available to advertisers
without the advertisers tracking them.
Instead, each browser is a little machine learning system
that's picking the best catalog entry.
Now, the catalog is not the ads.
Those are big, right?
It's a video or a webpage.
It's just the link to an edge cache.
And there are many such edge caches.
We're not trying to protect them from seeing your IP address.
It's not really feasible.
We could use Tor, but we don't yet.
And then some keywords about the ad.
So it's basically like metadata and a link.
And that's what the catalog consists of,
and that's what the machine learning picks.
- And the machine learning is learning
about you specifically, locally,
in order to choose from the catalog of different ads.
Couldn't this possibly be like a multi-billion dollar,
isn't this taking on the Google ad?
- Could be.
- So like what, I mean, one question to ask,
there seems to be some really profound ideas here
that are different than what the internet
has grown up to be.
If Brave or something like Brave,
the ideas, the fundamental philosophical ideas
underlying Brave went out and runs 95% of the internet,
how does that change the,
what are the major things that changes about the internet?
So social networks and then the creatives,
like YouTube creators and all that kind of stuff.
- So let's talk about that.
First of all, if Brave gets 95%,
I'm gonna demand a recount 'cause I won't believe it.
I don't know.
I think we're trying to put things into web standards
that can be standardized across browsers.
So the main value of Brave will be
the trust users have in us
and our ability to give the best deal to users.
So 70% of the gross ad revenue we give to the user,
and if they go through that KYC process I mentioned,
they can take it out.
They can also give it back.
They can take some out, give the rest back.
They can add basic attention tokens to give back.
Some of them turn off the ads 'cause they just don't like ads
but they put in $20 a month.
I believe Zucco of Zcashfame does that.
And that's very generous
because the browser is just anonymously,
based on his browsing,
sort of keeping score on how much time he spent
on this video, on that website.
And if those sites verify in sort of a,
like getting a domain certificate fashion,
they can get paid.
They can get part of his $20 a month.
So that vision could go big.
And if it does, I hope it's across multiple browsers.
I don't know that they'll all compete well
on the quality of the ads,
the quality of the ad blocking and tracking protection.
Those are subject to competition.
It'll take a while to standardize them.
But I think that would be a better world.
It would have less counterparty risk,
fewer fee takers in the middle, really just the browser.
We're taking 30%.
It's sort of the app store split.
And if we get bigger, maybe we can take even less.
Social networks, creators.
If you look at YouTubers,
a lot of them are the indies that are getting some size,
are getting sponsorship deals.
They're using Patreon.
They're encouraging people to subscribe
and give them regular money through Patreon.
But that's centralized through Patreon.
So there's censorship hazards.
There's a 5% fee.
What if that were a web standard?
What if Brave pioneered it first
and we took 3%?
And we did it in a way that was through your browser
so we couldn't censor it.
- That's brilliant.
Do you think it could be standardized across browsers?
Can Internet Explorer come in again?
- Yeah, protocols are easy to copy
and they're meant to be interoperable.
So there's a risk there.
And the loyal users might be tricked into leaving you
or they might, because of that distribution power,
you might end up getting stomped.
I don't know.
I can't predict the future.
I think antitrust is back on the case finally in the US
and certainly in Europe, DG Comp is doing its thing.
So I'm hopeful that we'll have a period of innovation.
People were talking, like Elizabeth Warren was talking
about breaking up the tech companies very clearly.
Now she didn't win and I suspect that won't happen.
But I also suspect that Google might be smart enough
to see they should do something more
than just put privacy perfume on Chrome.
They should maybe get rid of DoubleClick or something,
divest something.
I don't know, it might happen.
- So Brave might inspire Google to completely change
the way they're doing things in the future.
- They're already doing something you may have read about
called the privacy sandbox or Flock,
which they have this bird metaphor going,
turtle dove, fledge.
But these systems have been very Googly,
kind of over-engineered and yet,
depending on differential privacy,
which has weakness over time, if you know how that works,
it's kind of injecting noise to hide you in a crowd.
But over time, an adversary can pull you out of the crowd.
This doesn't look like it's gonna become a standard.
Like Apple, Brave, Mozilla, we're not gonna just say,
oh, Google, you saved us.
You've invented the privacy sandbox,
so we'll all just adopt it.
Not gonna be that easy.
It's gonna be more like pieces of what we do in Brave,
this anonymous ad matching
or the blind signature cryptography
we use to confirm the ad impressions.
That's David Choum's invention.
That could get standardized.
In fact, some of that is being standardized.
Even Google's in favor of so-called trust tokens,
which are Choumian blind signature certs.
But they're not using them for ad confirmations
'cause they don't wanna blow up their own business.
And they need to let some of the publishers they serve
have other ad tech scripts on the page.
And so they're kind of caught.
And this is something I realized doing Brave.
I thought, what's Google's innovators dilemma,
apart from just being mature and having trouble innovating?
It's that they have come to depend on this ad tech system
that has all these vendors that publishers rely on,
because publishers aren't technical enough.
And I feel for the publishers,
but I realized the users have to come first.
And if you give the users a better browser that's faster,
then you'll get enough users to give back
or support publishers.
The speed and the battery savings
and the data plan savings are significant.
There's so much bad JavaScript involved in ad tech
that if you block it,
you sort of chop off what's called the programmatic waterfall
which chains a bunch of requests.
- Yeah, that's one of the incredible things about Brave.
I guess you're saying you should attribute it
to the fact that the messy JavaScript, no offense.
- No, it's good.
It's not my excuse.
(laughing)
Is, I mean, Brave just feels faster,
even then, I mean, Chrome was fast.
One of the things that it was like impressive
is it showed that browsers can be really fast.
And Brave is even faster than that, which is incredible.
- We've blocked so much.
And it saves the network, which means data plan,
it saves battery because the radio consumes your battery
when it's running more to do those requests.
And it's just stunning how many there are.
Like some of my Google friends were like,
oh, that's just that bad site, they'll fix it.
And you actually do a survey of web pages
that they're like mostly like that.
I know Google engineers could make everything
super efficient, but they can't,
especially in antitrust court, do it.
They cannot take over all the publishers and do that.
They're trying with accelerated mobile profile, AMP.
They're trying to pull publishers.
They're like, oh, you poor publishers
don't know how to make your pages fast.
Put them on our AMP system,
we'll give you extra placement in the search carousel.
That's an antitrust problem for one.
But it's also, publishers we talk to hate it
because it degrades their brand.
Now they look like a gig writer wrote a piece
that's got Google's framing and AMP URL on top of it.
And they're trying to fix that too.
§21Transcript · 2:38:13–2:45:39
But it just looks like Google's
borgifying all these publishers
and they don't wanna be plugged into the BorgCube.
They wanna build up their own brand and have loyal readers.
So I'm in favor of giving the user's power
to help all the publishers
in this little platoons and the creators.
And so we talked about Patreon.
What about social networks?
Well, they're inherently like search, a global algorithm.
You're trying to find friends of friends.
You're doing the transitive closure of a graph
induced by this friend of relation.
But you should own your friend relation.
You should own your posts.
They shouldn't be owned by somebody else
who can take them down or censor them.
And your friend relations,
you should be able to find those friends on other networks.
And so I've tweeted about this, I haven't built it yet.
What if the browser could keep track of those for you?
What if the browser could maybe combine Facebook and Twitter
and you could find your friends on both
and you could have a sort of multi--
- So that relationship is not owned by Facebook or Twitter.
It's owned by you through the browser.
- They'll have terms of use and they'll say they own it.
But if they zap you on one and you're still on the other,
your friends find you
and the browser could preserve a combined view.
You could resurrect almost across networks.
- It's something I wanna maybe quickly ask you about.
On that front, there's been quite a lot of centralized,
we talked about Wall Street Bets and then Robinhood.
There's been centralized banning of different accounts
and removing like Parler, for example, from AWS
and this kind of overreach of centralized control.
Is your hope that it's possible to,
like what are your thoughts about that in general?
Is it, and is it possible to create tools
that give individual people the power
to fight back against overreach of such control?
- So we're talking about oligarchy, I do think.
And that, if it controls a nation state, that's formidable.
It's the tax and the police power, the military power.
It means that you may have the Great Firewall of China.
You may have people in China who are jailed
because of their tweets, right?
This is a serious threat.
I can't minimize it or say that we'll win.
I don't know how it's gonna go.
But I do think, like I said earlier
about the cunning of reason, people find ways around things.
The internet routes around censorship.
And this is not to endorse any particular bad faction.
One of the things that happens
when you try to wave the free speech flag too much,
you say I'm not gonna censor anything,
and you get colonized by terrible, terrible people.
I don't care if you call them neo-Nazis,
some of them could be doing illegal things.
And you don't want them colonizing
'cause it'll ruin your reputation and destroy your business.
So what you really want is that kind of
user-first subsidiarity, that subjectivity.
I want my social networks to be composited
in some multi-social user interface
where I don't lose track of people across networks.
And if they leave one or they get banned from one,
I can find them on another,
I can still sort of thread them together.
- Yeah, that's brilliant.
- And this didn't happen because browsers
got captured by the central powers.
And why did they get captured?
Mostly 'cause of search, and search is a central algorithm.
So Larry Page said this too many years ago.
He said, with search, you're giving up a little privacy
by handing the query over to us,
and then we'll error correct it.
Alan Eustace used to be a Google executive.
He said, oh yeah, we used to laugh.
They'd all be doing typos,
and they'd be typing the wrong word.
And we're like, no, dummy, type that query.
And it's like, okay, Google,
might wanna dial back that ego a little bit.
But yes, you do see all the queries,
and you can improve them,
and you can find the best results.
And that was Google's forte.
When we did the Firefox deal in 2004,
Google was really good.
And over time, SEO, which is an adversarial game,
and Google itself buying all these companies
and crowding its own results page
with its own tied-in stuff, the YouTube--
- It's the slip or slope that happens
when you have control over these kinds
of really important mechanisms
as a set of data. - Yeah, monopoly capitalism
or cartel, you get this with the Robin Hoods
and the hedge funds.
You get sort of the money interests take over
and kind of abuse their power and wear out their welcome.
So how do you get around that?
You have to have either new land to go to,
which some people's ancestors, not mine,
did to found the country.
I'm mostly Irish-German.
You have new virtual space people go to,
and that requires an ISP or a colo center
or Amazon to host you.
It requires domain name registrar who will not strike you.
And so when Parler was taken down,
I thought that was egregious.
Parler, it was not well-designed,
and I tried it out 'cause I tried all these things,
but I didn't use it.
And I also felt they were being unfairly scored
for not moderating, because you can find tweets
to this day that are horrendous
and threaten all sorts of violence.
Whereas Twitter, why isn't Twitter being taken down?
But so it was very selective.
It was the insiders who have the power
are gonna take out the newcomer.
And it looked bad, sort of like the hedge funds
shorting GameStop.
Looked bad.
You're seeing a piece in Time Magazine this week
that's basically saying,
"Yeah, we interfere with the election,
"but it was great, aren't we good?"
I don't know if you've seen this piece yet.
If you tried to say that as a Trump supporter in November
after the election, you'd get banned from Twitter.
But now Time, in its Twitter account,
is saying, "We saved the day.
"It's AFL-CIO and big business,
"the Better Business Bureau got together
"and kept Trump from spreading fake news."
So the country's kind of broken.
I don't know how to fix that.
The oligarchs have run wild, in my opinion.
And big tech is in the antitrust dock.
What's gonna happen?
I don't think they get out.
I think some of the DOJ, and certainly the state cases,
'cause they're separate cases,
are not gonna go away just 'cause
somebody got elected differently.
And these are career prosecutors,
and they have a strong case.
And Google's smart.
Microsoft almost got split up, right?
The judge, Thomas Penfield Jackson, he overreached.
He didn't hold a hearing about the remedy.
He just said, "I'm gonna break you up."
And Microsoft appealed, and the higher-level court said,
"Go back and figure this out.
"You're not breaking them up.
"You didn't even hold a hearing."
And when they got back, Microsoft said,
"Let's settle, let's settle.
"We don't wanna get broken up."
Because Jackson was gonna make the OpsCo,
the operating system company,
and the AppsCo, Office, Word and Excel.
And that would've been a huge blow to Microsoft.
- But ultimately, I don't know if you're optimistic
or cynical about the possibility of breaking up big tech.
To me, I'm optimistic that tools like Brave,
I love the idea of owning your friendships.
Users more and more owning the stuff is the only real way.
Unfortunately, it's like the Wall Street Bets subreddit
is the only real way to fight the centralized power.
You can't break 'em up through regulation.
It's very difficult.
- Certainly, I don't wanna wait for the law.
Netscape was long dead or acquired by AOL
and effectively dead.
It was only Mozilla that returned Firefox to the market
by the time that the US v. Microsoft case
was finally settled and the penalties were put in place.
And yet, antitrust has a role to play.
Those penalties caused Microsoft
to kind of turn away from the web.
They did Windows Vista, and they thought,
"The web's too painful.
"We got punished in court, and we had to standardize things
"with those icky standards people."
So they ran back to proprietary lock-in,
and Windows Vista flopped.
It was late, it was bloated, longhorn, remember?
Now, what I was gonna say, but Google's smart enough,
they won't get split up.
They'll split something out to get off the hook, I think.
- This is a complicated subject, but I, myself, was so,
I decided to journey out from the world
§22Transcript · 2:45:43–2:53:41
of being a researcher at MIT,
and potentially doing a startup myself,
and I've been thinking of,
I wanted to come to Silicon Valley to do so.
It's the land of the entrepreneur.
And there's a lot of my friends,
a lot of them are successfully,
have been entrepreneurs themselves,
have said, "Do not come to Silicon Valley."
You've started, you ran amazing teams of engineers.
You started a lot of successful businesses.
I wondered if you could comment on
why a lot of people are leaving California.
Is there something that could be fixed about California?
If you were starting a business today,
would you consider somewhere else,
like Austin, or some other place?
Or is Silicon Valley still, is it just a little lull,
everybody's being overdramatic
during this particular year of the coronavirus, and so on?
- I think, you know, even Austin's getting overheated,
I hear, and I've had relatives and friends move to Texas
within the last few months.
So Texas as a whole is a big place.
And, you know, people are moving to Florida,
there's a big movement toward Miami,
Peter Thiel, Keith, all these people.
The mayor has been very business-friendly about it,
which I think is just good politics.
America is fundamentally a commercial republic,
so you would think this would be what's happening.
For a long time, California was the golden state.
I came here in late '76 when I was a teenager.
It's in crushing debt due to the lockdowns.
It's got the highest taxes, that's gotta matter,
people will do high taxes.
It's got, likely fires every year because of the deadfall.
It's not global warming,
it's because the forests weren't managed
like they had been in the first part of the 20th century.
Just, I would say corruption at all levels,
especially up to the governor,
who, you know, famously was eating at the French Laundry,
and claimed that the outside was inside,
and they were out in masks off, and it was great.
Do what I say, not what I do.
Rules for thee, but not for me.
When you see that in leadership,
people either run, or they get rid of the leadership,
so there's a recall drive,
which is about to reach the threshold,
or in the old days, they get their guns, right?
You don't put up with this junk.
- But ultimately, the thing that made Silicon Valley
a special place, it gave freedom to young kids,
entrepreneurs, young minds, brave minds,
to think bold, to try different stuff.
I mean, even if the taxes are high,
so outside of financial stuff, outside of all of that--
- Housing's super expensive.
- Housing's super, so it's hard.
Okay, everything about startups is hard.
- Pencils narrow, and they didn't plan the roads, right?
- Yeah.
- They got rid of public transportation in LA,
like the Who Framed Roger Rabbit cartoon show.
They used to have trolley cars in Portland, too.
The oil companies and the DOD conspired to build highways
and make cars dominant, and the rights of way are long gone.
Like Elon's gonna go underground.
And I rushed him well.
That's probably the only way to do it now.
- But is it still a place, do you think it's possible
that Silicon Valley is still a place where magic happens,
where the next Google's built, where the next,
I mean, Brave is built, where--
- I think all good things come to an end.
I think the problem is Silicon Valley
had strong network effects through Stanford,
through the angel investor networks and the wealth effect,
and originally, you have to give
the federal government credit.
Like the ARPANET was a government project.
Let's not kid ourselves.
This wasn't wild free market, libertarian capitalism.
This was all Cold War stuff.
You had, out of the academia, you had Shockley,
and then the Trader SA, and Fairchild, and Intel.
But now, when's the last fab that was built in the Valley?
MicroUnity might've been the last.
I don't know, I haven't followed,
but we built a fab in Sunnyvale,
at MicroUnity, starting in the early '90s.
And now the fabs are overseas.
And the one thing that I will say
that the oligarchs have intentionally done,
in both parties, is sort of labor
and environmental protection law arbitrage
by going where the labor's cheaper,
and the environmental laws aren't as strict.
And that's polluted the hell out of parts of China,
but it's made things, you can make cheaper junk.
And this is not a story that's over yet.
So what is Silicon Valley for now?
It's for the network effect, the brain trust,
the who you know, the parties,
the Stanford sort of network.
That's fragile too, over time, I'm afraid.
Stanford, a lot of good professors I like,
they still filter mainly based on socioeconomic status,
but it's kind of a skate school.
I had a friend hired out of Harvard 20 years ago
at Netscape, and we talked about Harvard,
and he said, "Yeah, there's still professors
"who are great on the curve."
And I said, "Oh yeah, I don't think
"they're any doing that at Stanford anymore."
And he said, "Yo, it was shocking.
"Some of the students got Cs and Ds, and they were crying."
It's like, yes, that's right,
the precious deers can't take that at Stanford,
so they get As and Bs.
Now, you look at China, and people say
what you know about China.
They, true of Russia too, a lot of math science training,
a lot of engineering, a lot of people
who are doing their coursework to get the As and Bs.
So I'm an American, I'm born on the 4th of July.
- Really? - 4th of July.
- Yeah. - Wow.
- And America, as I say, fundamentally
is a commercial republic.
You can try to make it something else.
You can say it's the new Atlantis and mystify it.
You could talk about it in a more, I think, correct way,
which is 13 colonies that grew,
and then there's a lot of local,
or in the original design anyway,
the Federalist Papers talk about this,
there's a lot of subsidiarity.
But that's been eroded over time,
and like I say, a lot of the offshoring has hurt.
So what happened with coronavirus?
People working from home.
At first it was funny, 'cause I have friends at Google
who used to grumble that not only did they have
to come into the office, if they joined a different team
that was centered at a different office, they had to move.
Or if the VA team was reconstituted in Munich,
which it was after Lars Bach just got tired of JavaScript,
that they hired in Munich, or they hired PhDs
from Germany and moved them to Munich.
With coronavirus, everyone's working from home,
and it's like, what a relief,
I can work for Google from home.
But then the next shoe dropped,
and people started asking Mark Zuckerberg,
hey, can I move to my hometown in the Midwest?
And he said, okay.
And they said, oh, can I keep getting my Silicon Valley pay?
No. (laughs)
We're gonna figure out what your cost of living there is,
and we're gonna adjust your pay accordingly.
- And these colonies and these little mini experiments
that all combine to the big giant experiment,
I have a, I don't know, I have this vision of America,
which the country, so I was born in Russia,
like I said, here, and this is truly a wonderful country.
I wasn't born on the Fourth of July, but I might as well be.
People still flee.
- I still, and I'm red-blooded American at this point,
and I have a sense that we've figured it out somehow.
If Silicon Valley burns,
another place will come up in its place
that even more innovation, and people will move.
And the remote work might change fundamentally
how we work, or it might not.
It might just give you the freedom to then create
many other small Silicon Valleys throughout the place,
like Austin included, but other places as well.
And we've somehow figured it out.
- I think that's true, that there will be more mobility
and maybe new places that come up.
I don't know if Silicon Valley has passed some sell-by date
because it did hurt.
The coronavirus hurt, the lockdowns hurt,
in the sense that part of what keeps things going is social.
And so a lot of young people,
even before coronavirus, moved to San Francisco.
It was very strange to watch,
'cause in the '80s, we all lived in the valley,
and it was less populated, and San Francisco was grungier.
It was more like Dirty Harry in the '70s.
But by the '90s, and Jamie runs a nightclub there,
and he's talked about this,
you had sort of wealthy tech people moving in,
South of Market, fancy townhouses being built.
And that's continued in such a point
that it's almost like, what's the movie
by the South African director, Nels Jodie Foster,
§23Transcript · 2:53:43–2:59:12
up in the space colony?
Matt Damon is the guy on the Earth who has to go up.
Anyway, it's about the stratification.
It's about the great inequality.
The people in the space station have amazing medical
auto-docs that can extend their life
or save them, cure cancer.
People on Earth are all suffering, ground down in poverty.
And that sort of happened while I was here.
You saw a lot of money drive prices up
along the narrow peninsula,
and the single people wanted nightlife,
so they were in the city,
and the condos in the city got super expensive.
And I know even Google friends
who are socially responsible say,
we should have more housing built.
We should have, yes, in my backyard, not in my backyard.
But that's not happening, as far as I can tell.
And from the government to the incumbent landowners
and renters, it's just not happening.
And that has to drive people away.
I appreciate that people come here,
and you should wait for the prices to monitor.
They will, but a lot of people are gonna go
where the prices are lower.
- You, and sorry for silly questions here,
but just looking back, you have created things,
have been part of creating things
that have transformed this world,
the world of technology,
perhaps more than almost anything else.
But you're still a human being,
and unfortunately, this ride ends.
Do you ever think about your own mortality?
- Not too much.
I mean, I'm Roman Catholic, so I am not afraid of death.
I think a lot of people who have problems with death
are suffering from some lack of either faith
in their transcending death,
or maybe they don't have children,
or they feel like, they get later in life,
and they feel like they've missed opportunities
to do something that endures.
And I sympathize with a lot, because I'm old.
I got married fairly old, so I understand all that.
Nothing human is alien to me, as Terrence said.
But I don't fear it, no.
- What do you hope your legacy is?
- Yeah, it's gonna be JavaScript.
I think, no, I think my legacy has more to do
with my children and their children.
I think it also has to do with web standards.
It has to do with things like Brave.
The things we did with Firefox, when we did,
you know, I'm not gonna over-celebrate,
but I think Brave is important,
and we'll continue to prove this in a way
that counts for many decades to come.
But even Firefox, whatever its future fortune,
showed you can restart the browser market.
This thing you said about people opting out
and routing around, you don't need everybody to do that.
It's more like Taleb's stubborn minorities that do that.
It's the lead users, Eric Van Hippel's lead users.
You can be a few percent, you can tilt the market.
And that can be done in spite of the incumbents,
the moneyed interests, not being in favor
of what you're doing.
So I think what we did with Firefox won't be forgotten,
and it needs to be done more, and we're doing it with Brave.
And you could argue that other projects are doing it.
In some ways, blockchain is doing it.
The Robinhood take down the use of Robinhood
by the Wall Street Bets kids, similar.
So yeah, that kind of spirit endures,
and I think it, in some ways, it's American, right?
It's not hard revolutionary.
It's not trying to burn the past and destroy everything.
It's more like we have these certain, let's say, rights.
We have duties, too, so there's some debate
about which comes first in American jurisprudence
and the founding documents.
But as long as things are working,
we'll be like pragmatic Americans,
like de Tocqueville described in his writings.
But if things get too out of whack,
for one reason or another, too unequal,
too oligarchic and abusive, we're gonna assert our rights.
And even a few of us can do it.
And even in the American Revolution,
it was the minority who fought and put their lives,
treasure and sacred honor at stake.
It was a bunch of people went to Upper Canada,
I think it was called, Ontario.
- Yeah, that's a beautiful, I mean, that is,
at the core, what your work stands for,
is that a few people can have the power to transform society
with just a few radical ideas,
with just a little bit of code, change the world.
- Gotta do it. - And that's empowering,
and that is the American way.
That's why this country is, I believe,
the greatest country on Earth.
That's not over-rheumatized too much,
but I think some special things have already happened
in this country and will continue to happen.
And that's-- - And that spirit
can continue no matter who comes here.
They can adopt those folk ways and that spirit.
- Brandon, I can't tell you how much,
I was freaking out how much of an honor it is
to talk to you.
You're an incredible human being.
This is one of my favorite conversations ever.
Thank you so much for wasting all this time with me.
I really appreciate it.
- Oh, it seems like a breeze.
My pleasure.
- Thank you for listening to this conversation
with Brandon Eich, and thank you to our sponsors,
Jordan and Harberger Show,
Sun Basket Meal Delivery Service,
BetterHelp Online Therapy,
and 8 Sleep Self-Cooling Mattress.
Click the sponsor links to get a discount
and to support this podcast.
And now, let me leave you with some words from Jeff Atwood.
"Any app that can be written in JavaScript
"will eventually be written in JavaScript."
Thank you for listening, and hope to see you next time.
(upbeat music)
(upbeat music)