Jul 31, 2025 · 36m · no-priors
No Priors Ep. 125 | With Senior White House Policy Advisor on AI Sriram Krishnan
0:00 / 0:00
▶ Watch on YouTube →
gold bands on the timeline = statements, start to end. Hover to read, click to jump. CC turns on captions
In this episode of No Priors, hosts Sarah Guo and Elad Gil interview Senior White House Policy Advisor Sriram Krishnan to explore the American AI Action Plan, US-China technological competition, energy scaling bottlenecks, and the strategic importance of open-source AI.
How this conversation actually went
Every chapter scored 0–10 on four independent dynamics. Hover any point for the reasoning behind the score. The hosts hold 27.1% of the talking time here. How this is scored →
The hosts as informed peer 5.7 Guest teaching 2.5 Guest disagreement 1.3 The hosts pushing back 1.6
speaking balance: gold is the hosts, purple is the guest (3 minute bins)
Sharpest disagreement ▶ 29:49 Sriram Dismisses Technocracy Label
Sriram directly rejects Sarah's framing that the administration seeks a technocracy, refocusing the agenda on serving the American worker.
Hardest push from the hosts ▶ 29:27 Sarah Challenges on TechnocracySarah intentionally pushes Sriram with a provocative framing on whether technologists should run national policy.
Biggest teaching moment ▶ 4:01 DeepSeek Wake-up Call BriefingSriram educates the hosts on the White House's immediate reaction to DeepSeek and the realization of how narrow the American lead actually was.
The host holds their own ▶ 5:06 Elad Breaks Down DeepSeek Training MathElad injects precise technical industry context, correcting the popular media narrative regarding DeepSeek's supposed low training expenditure.
the scores for every segment, with the reasoning behind each
| Chapter | Topic | The hosts as informed peer | Guest teaching | Guest disagreement | The hosts pushing back | Why |
|---|---|---|---|---|---|---|
| Sriram Krishnan’s Background and Path to Government | 4 | 2 | 1 | 1 | Friendly opening banter between longtime peers. Elad frames Sriram's background across Silicon Valley firms and VC, while Sriram outlines his transition to White House policy work. | |
| The DeepSeek Wake-Up Call and Global Competition | 6 | 3 | 1 | 1 | Sriram discusses the catalytic impact of DeepSeek's release on White House policy. Elad jumps in with technical domain expertise explaining how DeepSeek's reported training costs only reflected the final run rather than overall R&D investment. | |
| Defining Victory and Measuring Market Share in AI | 6 | 3 | 1 | 1 | Sarah asks how victory in the AI race is defined, prompting Sriram to propose market share of inferenced tokens as a key metric. Elad adds an essential complementary dimension regarding cultural exportation and geopolitical soft power through AI models. | |
| Algorithmic Bias and Government AI Procurement Standards | 5 | 2 | 2 | 2 | Sriram explains the administration's executive order targeting ideological bias in federal AI procurement. Sarah asks a perceptive question drawing parallels to social media platform versus publisher debates. | |
| Action Plan Pillars: Infrastructure and Open Innovation | 7 | 2 | 1 | 1 | Sriram outlines the action plan's infrastructure and innovation pillars. Elad reinforces the argument for open source by comparing it to the internet and crypto protocols, noting how closed models enable centralized government control. | |
| Open Source as Strategic Soft Power and Security | 6 | 3 | 1 | 1 | Sarah and Sriram discuss the security and geopolitical risks of relying on Chinese open-source weights. Sarah then presses on energy bottlenecks, and Elad explores the trade-offs between laying fiber near cheap power versus upgrading the electrical grid. | |
| Supply Chains, Workforce, and Global Stack Export | 6 | 3 | 1 | 1 | Elad asks about physical AI, supply chains, and China's strengths in robotics and autonomy. Sriram describes the plan's focus on global standardization around the American tech stack and domestic manufacturing capacity. | |
| Implementation Strategy and Tech Talent in Government | 5 | 2 | 1 | 2 | Elad asks how agency actions will be implemented and what fallback plans exist. Sriram emphasizes speed and administrative alignment, while Elad praises the technical depth of the personnel brought into this administration. | |
| National Technological Leadership vs. Technocracy | 6 | 3 | 2 | 4 | Sarah poses a deliberately provocative question asking whether the U.S. should be a technocracy. Sriram politely reframes the premise around broad workforce benefits and national economic leadership rather than technologist rule. | |
| Addressing AI Safety Concerns and the Open Source Defense | 6 | 2 | 2 | 2 | Sarah asks for the strongest counterargument against AI safety doomerism. Elad identifies regulatory capture incentives from closed-source incumbents, and Sriram cites Linus's Law to argue that open source provides superior decentralized security. |
Statements from this episode (26)
Krishnan: Senior Global Officials Lacked Basic Competence in AI Policy
“A lot of people who are in very senior roles in governments in the United States back then and other parts of the world didn't know what they were talking about when it came to AI. I was convinced that they were doing the absolute wrong thing on many topics, f…”
Krishnan: America Currently Holds Only a Very Small Lead in AI
“It told us that America doesn't have a huge lead. On AI. It actually has a very, very small lead.”
Krishnan: DeepSeek Preceded Anthropic and Google in Releasing Reasoning Models
“DeepSeq was the only reasoning model, which was not open AI. So I don't think Claude had come out of the reasoning model yet. I don't think Google had yet. It was the only non-open AI reasoning model.”
Krishnan: Chinese Models DeepSeek and Qwen Are the Best Open Source
“I think to even today, I would probably say the Chinese models, Deepsea, Quan are the best open source models”
Krishnan: AI Leadership Will Create an Accelerating Economic Flywheel Effect
“AI might be the most transformational economic cultural force of our lifetime. And I believe that if the country or the ecosystem Which winds up getting ahead is going to have these cyclic effects, right? Like you're going to, you know, you're going to power p…”
Krishnan: Measure AI Dominance by American Hardware and Model Inference Share
“And one of the measures I've been thinking about is let's say the world inferences, I don't know, maybe it's called 10 quadrillion tokens a month. We don't know what the number is, right? What share of those tokens are being inferenced on American hardware on …”
Gil: AI Models Are the Next Global Vehicle for Cultural Export
“And so if you look at prior waves of sort of culture spread, it was the movie industry, it was social media, and then now it's these models, because a lot of people go to these models as a trust, a source of truth for history, for information, for other things…”
Krishnan: Twitter Algorithms Easily Injected Cultural Bias Into Mainstream News Cycles
“One of the things I saw when I was at Twitter was how easily you could inject cultural bias in your algorithms. I have so many stories about how if you pick the right kind of Twitter accounts, which then feed into the trending algorithm, which then feed into T…”
Krishnan: Passing SB 1047 Would Have Ended American Open-Source AI
“You know, like until last a year and a half ago, I was in California, along with all of you, California almost passed SB, which if that had happened, it would have been the end of open source, by the way, in the United States, we would not have a Lama. We woul…”
Krishnan: Single-State Tech Regulations Often Become De Facto National Law
“If you have a small state set rules, It can often become the de facto law for the country, because if you're a company, you're like, well, I have to operate in this state, or I have an office, so let me just do that for everybody. It's kind of just like the EU…”
Krishnan: Biden Administration Attempted to Demonize and Ban Open-Source AI
“The Biden administration really, really tried to scare people about open source, talk about how unsafe it was. SB. And obviously tried to kind of basically ban it in many ways.”
Krishnan: Biden AI Policy Sought Centralized Control Across Four Companies
“One, very fundamental difference I think we have with the Biden administration is the Biden team really looked at AI as something to be centralized and controlled. Everything was about how do we make sure that we regulate these three or four companies and only…”
Guo: Western Companies Are Broadly Using Chinese Open-Source AI Models
“They're happening and Western companies are using Chinese open source models very broadly already.”
Krishnan: Global Usage of DeepSeek and Qwen Is Geopolitical Soft Power
“When somebody is using DeepSeq or Quinn, that's an expression of soft power.”
Krishnan: Undetectable Backdoors in Coding Assistants Threaten Critical Infrastructure
“Interpretability is still a nascent field and you could very easily see ways where You plug a model into a cursor or windsurf and you generate a piece of code. And then two years down the road, it turns out that code had a little if statement saying, if I'm ru…”
Krishnan: Climate Lobbyists and Doomers Blocked US Nuclear Development for Decades
“Nuclear, that's another case where I think for decades and decades, the climate lobby and the doomers that kind of stop any real efforts over there.”
Krishnan: US AI Dominance Demands Blue-Collar Trade Workers, Not Just Engineers
“America needs not just engineers, but it needs people up and down the stack. It needs electricians, technicians. We need to get construction going and we need to get these jobs and this whole ecosystem back in back in the US.”
Krishnan: Biden Diffusion Rule Effectively Banned US GPU Exports to Allies
“Under the Biden era, there was something called the Biden diffusion rule, which basically was a 200 page document, which basically made it illegal for America to export GPUs. It was really hard for you know, if you're Jensen or if you're Lisa Su to really kind…”
Krishnan: Exporting US GPUs Strongly Incentivizes Allies to Run American Models
“And one of the other things we're doing that is we get our GPUs over, we probably get them to run our models as opposed to models from, you know, another country, and we go from there.”
Krishnan: Physical AI and Robotics Will Become Critical Within Two Years
“The action plan actually has a section in it of making sure, you know, we are set up For robotics. I think that's obviously that's going to get super key within the next 18 to 24 months.”
Krishnan: Robotics Startups Are Heavily Using Distilled DeepSeek and Qwen Models
“When I was talking to a bunch of robotic startups, you're seeing a lot of distilled DeepSeq, a lot of distilled Quen out there.”
Gil: Trump Administration Recruited Significantly Higher-Caliber Tech Talent Than Predecessors
“And so I think one thing that's striking about this administration is the caliber of tech people that they actually got this time around. Is very high relative to prior administration.”
Krishnan: David Sacks Regularly Explains Deep AI Technical Concepts in White House
“Like you'd be shocked at how often I've seen David Sachs in a meeting explain how inferencing works, you know, what high bandwidth memory is, you know, how the world has shifted from a, you know, pre-training context to a post-training context.”
Krishnan: Chinese AI Powering Global Devices Will Have Catastrophic Geopolitical Consequences
“There are going to be catastrophic consequences if Chinese models are running on every robot, Every camera, every car, every device around the world.”
Gil: Closed-Source AI Labs Use Existential Risk Narratives for Regulatory Capture
“It's a very traditional form of regulatory capture. If you have a big pharma company, they work with the government to prevent other entrants into the industry. And this is exactly what it felt like is at least partially happening. In the AI world. Now that ma…”
Krishnan: Independent Open-Source Scrutiny Detects AI Vulnerabilities Better Than Internal Teams
“If you have a model up on hugging face and somebody downloads a final gigabyte file and the thousands of students and researchers just kind of pounding away on that, I think there's a good chance they're going to find issues a lot better than a very small safe…”