He saved OpenAI, invented the “Like” button, and built Google Maps: Bret Taylor on the future of careers, coding, agents, and more

Taylor's Google Local story begins with a product that had every apparent advantage: a prominent link on Google's homepage and a clear local-search problem to solve. Yet it underperformed because it mainly reproduced the Yellow Pages online. The product worked, but it did not create a compelling reason to choose it over an incumbent. Taylor's next attempt changed the experience itself by making the map the canvas and combining mapping, local search, and driving directions. Google Maps reached about 10 million users on its first day, while later satellite imagery from Keyhole brought about 90 million users in a single day. His lesson is to use new technology to make something the old medium could not be, while distinguishing the initial attraction from the durable value.

Why should I use this thing in the first place?

Taylor explains his movement across engineering, product, executive, and founder roles through a flexible identity. He calls himself an engineer but more broadly a builder, and says a founder must be willing to become whatever the company needs: a salesperson to investors, candidates, and customers, a product thinker, a designer, or an engineer. A conversation with Sheryl Sandberg changed his management approach at Facebook. Instead of organizing the job around the work he personally preferred, he began asking what would most improve the mobile and platform organizations that day. The answer could be recruiting, partnerships, product, engineering, or sales. He found that the common source of satisfaction was impact rather than a particular function.

Taylor qualifies the impact heuristic with a warning about self-deception. The host's question about a major mistake leads into a broader point: leaders can turn an intuition about a product into an organizational fact, and the difficult work is testing whether the story is true. The episode's opening failure shows why this matters. A product with traffic and basic functionality can still lack a compelling reason for users to choose it. Honest diagnosis requires questioning the comfortable explanation rather than treating a failed first version as proof that the whole problem is unsolvable.