David, you’ve said open-weight AI is “synonymous with freedom” and America needs to win that category. Maybe we're missing something, but it appears the admin is doing the exact opposite? America’s AI Action Plan certainly gave open-weight the usual lip service about innovation and “geostrategic value.” The real policy moves seem to be massive data center acceleration, deregulation that protects hyperscalers, voluntary frontier model frameworks, and infrastructure spending that is creating massive financial and competitive advantages for closed-source giants. Anthropic now looking down using its model for innovation. We’re treating open-source AI exactly like critical minerals several years ago: we say it’s strategically critical, then do nothing concrete and let the market (and China) run the table. What’s the actual administration plan to make American open models win instead of just pretending the free market, that is largely invested in closed-source, will handle it?

Stay focused on your purpose. Congratulations to @elonmusk and everybody who works, invested in, and believed in @SpaceX and it's companies. Persistent is underrated.

David, you’ve said open-weight AI is “synonymous with freedom” and America needs to win that category. Maybe we're missing something, but it appears the admin is doing the exact opposite? America’s AI Action Plan certainly gave open-weight the usual lip service about innovation and “geostrategic value.” The real policy moves seem to be massive data center acceleration, deregulation that protects hyperscalers, voluntary frontier model frameworks, and infrastructure spending that is creating massive financial and competitive advantages for closed-source giants. Anthropic now looking down using its model for innovation. We’re treating open-source AI exactly like critical minerals several years ago: we say it’s strategically critical, then do nothing concrete and let the market (and China) run the table. What’s the actual administration plan to make American open models win instead of just pretending the free market, that is largely invested in closed-source, will handle it?

OpenAI may have just run the best ad campaign for open-weight AI in history. No warning. Accounts shut down. Builders scrambling. Probably a mistake. Maybe an automated system error. Either way, the lesson is obvious: the intelligence layer of the economy cannot be controlled by a few companies with a kill switch. We need a US stack open-weight AI system so innovation doesn’t require permission from a few companies who will undoubtedly have political motivations.

This is one of the best thing done in American people for years. Congratulations Brad to you and your wife, great work along with @MichaelDell and his wife.

A significant portion of unicorn founders were rejected over 50 times before finding a VC willing to be anti-consensus. We met a unicorn founder who was turned down 100 times during their seed raise because the product sounded like science fiction and the team didn't fit the AI founder archetype. In reality, the company was simply years ahead of its time, and consensus was a lagging indicator of their potential. Rejection is framed as a setback, but for the best founders it acts as a forcing function. Without early capital, teams are forced to prioritise building, testing and finding true customer pull over pitch-deck salesmanship. This scarcity builds a level of resilience and operational sharpness that easy funding cannot teach.

Very important piece on OpenSource (weight) AI v. China by @bgurley. Today’s frontier AI systems making major gains in domains like coding are not the omniscient, standalone AI models the industry often wants people to imagine. They are complex systems: multiple models, tool loops, memory layers, planning steps, retrieval, verification, and software harnesses designed to keep the model stable long enough to solve difficult problems like software bugs. That matters big time. The advantage is not just model weights. The advantage is the entire execution system around the model. And if that execution layer is controlled by only a few large companies, innovation becomes narrow, expensive, and dependent on their priorities. To compete with China and unlock AI across every major industry, the U.S. needs widely distributed open-source models, domain-specific systems, and new execution architectures that enterprises, startups, universities, and government agencies can build on directly. The future of AI should not be controlled by a handful of closed frontier labs. It should be an open, distributed innovation layer across the economy.

amazing post and great timing w.r.t. ant's post yesterday we must build open ai to not get locked in by the vendors who will decide who gets which capabilities and the west has to realize that open models are important and support open model efforts (like @arcee_ai, @NVIDIAAI)

From the article:

Wild. Pay the ticket before you leave. @BretBaier

Happy Mother’s Day to all the wonderful mothers out there. No bigger calling than being a Mom and looking over our precious children.

What’s happening at SpaceX in El Segundo is truly extraordinary. Walking through those buildings, you feel an incredible sense of pride knowing that some of the most advanced technology in the world is being built right here in Southern California. The young men and women working there every day are helping shape the future of space, communication, and American leadership. We should all be proud that the USA is taking the lead in the future of space!

Feel like I'm watching an episode of the Hills. Let's go Feels like I’m watching an episode of The Hills. Let’s go, @spencerpratt. It all starts with a softball question to Bass about being out of the country before the fires and a warning about naming call during the debate.

Well said @secrubio. A long way to go, but also much progress made.