Jacob Helberg: Winning the AI Race Part 1: Michael Kratsios, Kelly Loeffler, Chris Power, Shyam Sankar, Paul Buchheit, Jake Loosararian

The recording opens with an archival-style montage of rocket launches, the moon landing, early computing and a chess-playing computer. The clips frame AI as the latest in a line of technological leaps, with lines about "an age of abundance" and the claim that "the rate of progress is exponential right now." The montage also introduces the event's central framing: those who can "harness and govern" superior technology "will win," driving "economic vibrancy and military supremacy." It includes the statement that America's destiny is "to dominate every industry and be the first in every technology," including being "the world's number one superpower in artificial intelligence."

The hosts describe this as their first event in Washington, D.C., organized "in just a couple of weeks." They present winning the AI race as something America "has to do" and will do through "grit, entrepreneurship, and dogged competition." What they say distinguishes the current administration is that it is "actually engaging with the technology industry." They credit David Sachs, who took on the role of crypto and AI czar, and after some banter about whether he prepared, they ask about his first six months in government.

The reply says the speaker "never expected to go into government at all." It traces the appointment to President Trump appearing on the podcast about a year earlier, which began a relationship that led to the job offer. The speaker calls the role "a once in a lifetime opportunity" and says AI and crypto have been two of the president's agenda issues where "we are getting a lot done." The event itself was assembled over roughly the last 10 days to coincide with the publication of the president's AI action plan that same day, in partnership with Jacob Helberg's Hill and Valley Forum.

The forum is described as "a community of technology, of builders and policymakers" who see technology as "an engine of wealth creation" and "an indispensable pillar for American national security." One speaker, referring to their own recent confirmation hearing, repeats the view that the country is "at an inflection point" in an AI race. The programming is laid out in three parts. The first covers new industries that AI has created, ones "that couldn't have existed a decade ago." The second covers industries that enable AI: "mining, energy, chips, fabs, and data centers." The third covers the government's role in an economic transformation said to be "already underway."

One speaker directly disputes a common media narrative. They call the press claim that AI is destroying jobs "wrong" and say that what they see "on the ground" is "an incredible job creation engine." They say these stories are "not told enough." This is offered as an observation, and no data is given at this point in the discussion. Another participant stresses that the issue "transcends party" and calls it "the issue of our lifetime," while acknowledging "a lot of hard questions and a lot of hard debate."

Asked how the administration brings together differing views, a participant describes a historical "fork in the road" between optimistic and pessimistic views of technology. In the optimistic view, the country that harnesses leading technologies captures most of the gains, then debates how to spread them internally, and turns economic supremacy into military dominance. The speaker argues that historically the U.S. went the other way, toward "mistrust," which let global competitors emerge and create "real fundamental existential risk" to America's place as a superpower. They describe the last six months as "a massive pivot" back toward celebrating American technology and its builders.

Michael Kratsios, director of the Office of Science and Technology Policy, is invited on stage. One participant explains the plan's origin. In his first week in office, President Trump signed an executive order directing the speaker, Michael and the National Security Advisor to produce the plan, with the goal of figuring out how the U.S. would "dominate in AI." The speaker frames AI as a global race in which losing "would just be unthinkable" because of its "huge ramifications" for the economy and national security. The plan released that day contains "90 concrete actions" that the executive branch can take.

Innovation: "There's just no substitute for innovation." The plan includes measures meant to help the private sector, startups and the tech community out-innovate global competitors. The speaker argues that regulation alone cannot win the race. Infrastructure: more and better AI infrastructure in the United States, including data centers, energy and manufacturing. The AI ecosystem: drawing on the Silicon Valley lesson that companies with the biggest ecosystems, meaning the most developers and the most apps, are the ones that dominate, the U.S. should aim to create "the AI stack for the entire world."

The account of the drafting process starts with a request for information (RFI) issued after the executive order. The response surprised the drafters: "We had over 10,000 responses come from all corners of the country," from tech companies to Hollywood actors. The speaker takes the breadth of responses as evidence of how much AI touches every U.S. industry. The comments were combined with input from the agencies that work with AI to produce the plan.

The speaker contrasts the document with the national AI strategies that many countries have published over the last five or six years. The emphasis here is on the words "action plan": items achievable "in the next six to nine months." On deregulation, the argument is that there will probably never be a single law that defines how AI is regulated. Instead, AI will be built into technologies overseen by many different agencies, such as drones, self-driving cars and "FDA-approved AI-powered medical diagnostics." Each of those regulators therefore has to create conditions in which these technologies "can thrive and not to be hindered by the government."

A second innovation thread is using government data for AI-driven scientific discovery. The speaker says the first wave of AI showed "great progress" in coding, and they point to the Department of Energy's national lab data sets as a resource for discoveries in materials science and medicine. On infrastructure, the plan calls for "categorical exclusions for AI-related activities" so that data centers and power generation can be built on federal lands. This is to be paired with other efforts to speed up power buildout.

The moderators bring up recruiting talent from abroad. They say they want to keep the discussion on AI rather than border enforcement, and they note that the administration has sent "different signals" on a "very controversial issue." The answer recasts the talent question beyond elite engineers. Companies building large U.S. infrastructure projects, with Crusoe named as an example (rendered "Caruso" in the transcript), report shortages of electricians and HVAC workers. The plan directs agencies such as the Department of Labor to use re-skilling programs to fill that gap. The stated philosophy has two parts: attract "the greatest scientists and engineers" to the U.S., and train the American workforce for the necessary jobs. The answer does not spell out specific immigration policy changes.

On releasing valuable data held by agencies such as DOE and FDA, the response says government has "generally" taken an open-source approach. It also acknowledges "a lot of lip service" over the years. The practical obstacle is the format of the data: "dirty, nasty data that isn't homogenized in any way" is "not particularly helpful." The speaker points to recent legislation, referred to as BBB, which provides $150 million to the Department of Energy for an AI for Science program meant to work on this problem.

The final question asks whether federal law should preempt state AI regulation, given that "over a thousand state laws" affecting AI have been proposed or passed. The first answer avoids taking a position. It notes that preemption comes up often in technology policy, privacy being one example, but depends largely on what Congress does, whereas the plan focuses on actions the executive branch can take itself. A second participant concedes that the plan does not address the issue. That participant nonetheless calls the state-level activity a "real threat to national security." They warn of "a patchwork of 50 different state regulatory regimes" while China treats AI as a national priority. They say the administration "weren't ready to declare a policy yet" but that the issue will need to be examined "over the next year or so."