🔥 Big Tech's Race to Open-Source, Private Aviation Addiction, and the Future of AI Policy

📰 AI Frontier Wars: The Fable 5 Release and the New Regulatory Reality

The battle over how to tame AI has officially begun. Fable 5 is now live, marking a pivotal moment in the ongoing debate between open-source accessibility and closed-source control. The Wall Street Journal noted that the release appears to have been "nerfed" in certain ways, raising questions about transparency and the evolving standards for model deployment.

The core tension: how do you prevent a model from building bioweapons or hacking systems without disclosing the limitations? And what happens when someone simply asks the AI to build another AI that can circumvent those restrictions? The debate around unfettered intelligence—defined as "not limited by rules or any controlling influence"—is no longer theoretical. It's playing out in real-time across labs, governments, and regulatory bodies.

"If there's lab nationalization, if there's some sort of distribution of equity to Americans, it happens through the Trump accounts, through social security number, through some sort of identity that the individual can have a stake in."

Meanwhile, OpenAI is reportedly in early conversations with the Trump administration to hand over a 5% equity stake in the company. The Financial Times describes these talks as preliminary, but the plans seem to involve other frontier AI players apportioning stakes in their companies as well. Dean Ball, formerly of the Foundation for American Innovation and now at OpenAI, offered a colorful analogy: giving the government a direct stake is "akin to inviting rats to live and reproduce in the walls of your house."

Joe Weisenthal countered with a tongue-in-cheek proposal: "Rather than equity stakes, why not make companies pay 20% of all pre-tax income to the federal government?" The debate touches on deeper questions of corporate governance, public participation in AI upside, and whether direct equity distribution to households makes more sense than funneling shares into a sovereign fund.

🛩️ Private Aviation Boom: An Addiction Like No Other

Flexjet's Global CEO Andrew Collins joined the show to discuss the company's explosive growth and the structural changes reshaping private aviation. Collins, who came into the company in 2012 through the acquisition of Sentient Jet (which invented the jet card), described the COVID-era surge as a watershed moment for the industry.

"Private aviation is to me one of the most addicting things on the planet. If you get a bunch of people that get introduced to it because there's a pandemic, you can't go back. It's really hard to go back."

Flexjet now operates 350 aircraft and plans to grow to 400 by year-end. The company has shifted from a focus on the 48 contiguous states to a truly global footprint, with transatlantic flying up 60% year-over-year. The fleet is also evolving—moving from light and super-mid jets to large cabin aircraft like the Gulfstream G700 and G500.

Collins emphasized that commercial aviation is probably the best marketing force that private aviation has. Service degradation, government shutdowns, and capacity constraints at major hubs have all contributed to rising demand for private alternatives. Flexjet's model is capital-efficient: most of the fleet is owned by consumers, not carried as debt on the balance sheet.

The company is also leaning into the economy of experiences, partnering with luxury brands like Formula 1 to create "only in Flexjet" moments. Collins noted that partnerships with tech platforms like Starlink are helping deliver high-speed in-flight entertainment and a frictionless experience.

🏭 Nuclear Startups and the Energy Race

Isaiah Taylor, from Valor Atomics, announced a major milestone: the company became the first nuclear startup to ever make electricity. Valor's W-250 reactor went critical and powered an NVIDIA Spark, marking a significant step in the company's mission to provide scalable, clean energy for AI infrastructure.

Taylor highlighted the speed and flexibility advantages of nuclear over natural gas. Uranium fuel is incredibly dense—one truckload delivers an entire year's worth of fuel—which allows for more flexible site placement and scale. Valor's timeline has been aggressive: the executive orders were signed in May of last year, and the company committed to turning on a reactor by July 4th. They delivered ahead of schedule.

"We didn't just go critical, we went critical twice and we went to thermal power operations and we went to electricity production and we did it outside of the National Labs."

The company is now moving from research to commercial deployment, with plans to engage the NRC with empirical test data gathered from the Utah facility. Taylor also emphasized the importance of community engagement, noting that Valor approaches local leaders before breaking ground to discuss cost-benefits and potential trade-offs.

🤖 AI Policy Heats Up: Dean Ball on Nationalization, Export Controls, and AGI Pills

Dean Ball, who recently joined OpenAI, returned to the show to discuss the increasingly personal and fast-moving world of AI policy. He noted that DC is "underrated" in terms of how AGI-pilled it has become, driven in part by highly online staffers in their 20s and early 30s who dominate Capitol Hill and the White House.

"The US government placing export controls on a frontier model was such a costly move and such a crazy and radical move to make. It was clearly made out of a position of fear. Everyone in the world saw that."

Ball argued that the mythos fable export control incident was a turning point. Governments worldwide—many of which look to the US for intellectual leadership—are now waking up to the potential risks and strategic importance of AI.

On nationalization, Ball expressed concerns about the government taking equity stakes and placing them on its balance sheet. He warned of "picking winners" dynamics, corporate governance issues, and the risk that labs could be deemed instrumentalities of the government, subjecting them to public utility-style regulation. He favored a model where equity is directly distributed to households, creating a symbolic stake in the AI future rather than funneling wealth into a sovereign fund.

Ball also predicted that by 2030, the question of whether AIs deserve legal rights and protections will be a mainstream societal and political debate. As AI becomes more deeply integrated into daily life—AI employees, AI friends, AI doctors—society will begin to grapple with questions of sentience, consciousness, and moral status.

🚀 Open-Source AI: The Supply Chain of Intelligence

Vimple, co-founder and CEO of Together AI, joined the show to discuss the state of open-source AI and the company's recent $800 million Series B round. Together AI serves open-weight models and focuses on delivering powerful AI at dramatically cheaper costs. The company's bookings grew from $100 million to $1.2 billion over the past year.

Vimple argued that open-weight models have closed the gap with closed models and are now capable of addressing the largest agentic and long-horizon tasks deployed in enterprises. He emphasized that open models are becoming a strategic necessity for companies with sensitive data, as they provide transparency, auditability, and control.

"These models are software, right? You can inspect them, you can benchmark them, you can study them in a way that you can't do with closed models."

On the question of distillation, Vimple downplayed its role, noting that recent improvements are driven more by reinforcement learning and the maturation of training tools than by on-policy distillation. He also argued that the US government's focus on open-source risks is somewhat misplaced, and that open models are essential for building trust and enabling innovation at scale.

📊 Jersey Mike's Goes Public, OPM Goes Digital, and SpaceX Rumors

In a sign of the times, Jersey Mike's—the sandwich chain founded by a 17-year-old with a $125,000 loan from his high school football coach—is filing for an IPO. Blackstone acquired the company last year for $8 billion and is now seeking a $12 billion valuation, marking a 50% markup in just one year.

In other news, the Office of Personnel Management (OPM) has finally digitized the retirement process for government employees, ending a decades-long reliance on paper files stored in underground limestone vaults. The old system could take up to six months to process a retirement. The new system takes minutes.

"We watched as they keyed data into a black and white terminal, transmitting to the Cobalt mainframe built many decades ago."

Meanwhile, the Wall Street Journal reported that SpaceX showed investors a prototype of an AI-enhanced handset during its recent IPO roadshow. The device is said to be slimmer than the iPhone, runs a proprietary operating system, and uses XAI technology built around Qualcomm's Snapdragon chipset. Elon Musk has repeatedly denied plans to build a phone, but the reporting suggests the company is at least exploring the possibility.

🔮 Five Bold Predictions for 2030

Rob Taves from Radical Ventures shared his latest set of predictions for the year 2030, including:

• Anthropic will be one of the largest and most important life sciences companies in the world. Taves argued that Anthropic's next big focus area will be biology, driven by CEO Dario Amodei's background as a neuroscientist and the company's recent hiring of John Jumper from DeepMind.
• TSMC and ASML's monopolies will be broken. Taves pointed to Elon Musk's Terrafab project as a key challenger, noting that Musk calculated his companies alone will need 50 times more chips than are currently being produced globally.
• Telepathy will be a well-established way to communicate. Taves defined telepathy as "a human being able to communicate with other people using only their thoughts," and pointed to UCSF's work on brain-computer interfaces as evidence that this is already technically feasible.
• AI will consume less energy per task. While total energy consumption will rise due to Jevons' paradox, Taves predicted that AI will be millions of times more energy-efficient per task by 2030.
• The question of whether AIs deserve legal rights will be a mainstream debate. Taves argued that as AI becomes more sophisticated and takes on physical forms, society will begin to grapple with questions of sentience and moral status.

✈️ Final Thoughts: Infrastructure, Intelligence, and the New American Economy

This week's show was a whirlwind tour through the intersection of technology, policy, and infrastructure. From private aviation to nuclear energy to the future of AI governance, the common thread is speed—and the question of whether institutions can keep pace with innovation.

As Dean Ball put it: "The truism in AI policy is that the technology moves so much faster than the policy. I think we might actually be going through the maybe the one period where the policy is actually moving faster than the technology."

Whether that's a feature or a bug remains to be seen. But one thing is clear: the next few years will be wild.

Until next time, stay sharp. 🚀