Elon Musk vs. Sam Altman, AI Job Loss, and OpenAI’s $852B Valuation | EP #247

Peter and Dave argue that the traditional job market is being disrupted, but entrepreneurship is more accessible. Dave emphasizes that you don't need incredible skills, just desire and purpose. They reference their last podcast where they discussed this. The advice is a direct response to AI job loss fears.

Dave shares that he told his team to target a one-to-one match of payroll to AI cost by year-end. He argues that once people use AI heavily for a month, they never go back. This is a practical way to overcome resistance and accelerate the learning curve. He contrasts it with token leaderboards, which are primitive; a better metric like 'machine leverage per employee' will emerge.

Dave laments that Meta employees pulled out of the Claude leaderboard because they didn't want to expose their prompt history. He argues that sharing is how everyone improves together. He sees it as disheartening but believes it's the right thing to do.

Peter pitched this idea to Michael Kratsios, the science advisor to the president. He suggests it as part of a new social contract. Alex notes that China already has a similar policy, which would be ironic for the US to adopt. This is positioned as an alternative to simple UBI checks.

The Wharton study showed a 5-month coding course with AI tutors was equivalent to 69 months of additional schooling. Dave argues it's cruel to force-feed lectures when AI can adapt. Alex notes that for self-motivated students, AI is already a Diamond Age primer, but for others, we need AI embodiment that holds attention. Peter suggests gamifying education like video games.

Alex argues that just being an AI startup is no longer enough; you must be recursively self-improving. He cites OpenAI, Anthropic, xAI, and Waymo as examples. This is the new standard investors are looking for.

This is a sponsored segment. Dr. Dawn Mussallem explains that 88% of people coming in have detectable coronary artery disease, and 23% have soft plaque. The message is to be the CEO of your own health.

The hosts discuss how Musk's management style, effective for hardware like rockets, may not translate to AI training where bugs can waste hundreds of millions in compute. They note that parameter count doesn't directly correlate to capability, and that other labs have stopped reporting parameters. Alex suggests the 10T model may serve as a teacher for distillation. The reorganization includes eight founding engineers leaving and the president of xAI coming from Starlink. The IPO is expected this summer with a $2T valuation, raising questions about rebuilding while pricing an IPO.

The hosts discuss the geopolitical nature of the fight, the potential for settlement involving Altman stepping down, and the irony that Musk himself pushed for majority control of the for-profit in 2017. They note the New Yorker investigation published simultaneously. Alex expresses sympathy for OpenAI's governance evolution, comparing it to Singularity University's own flip from nonprofit to for-profit. The trial will feature testimony from Altman, Brockman, Satya Nadella, and Musk. The outcome could unlock value in other nonprofits like universities.

The hosts debate the plausibility of the numbers, with Dave suggesting 100B is a good target but 200B is possible, while the trillion the following year is unlikely. Alex ties it to the race to become the default 'Open Claw' provider—hosting 24/7 multimodal agents. He notes that Anthropic's product decisions are shaped by the looming expectation of an Open Claw-like functionality. The discussion also touches on the competitive landscape: OpenAI is tied up in lawsuits, Google is behind, and xAI is rebuilding, giving Anthropic a window.

Dave speculates that Sam Altman moved top talent like Kevin Weihl to this initiative to generate world-changing breakthroughs that could influence the lawsuit's political outcome. Peter notes that breakthroughs from GPT-6 used for science could be worth trillions. Alex jokes that if the foundation cures Alzheimer's, it might have to become a for-profit again. The hosts see this as a strategic move to demonstrate the mission's success.

Dave predicts more such deals, comparing it to Google buying DeepMind for $600M with no revenue. Alex argues this shows the AI boom is not a circular financial bubble but is metastasizing into real sectors like biotech. He ties it to collapsing timelines for curing all diseases, referencing Chan Zuckerberg Initiative's shift from end-of-century to next few years. The hosts see this as how Dario's vision of solving neurological disease by decade's end will be achieved.

Peter explains the old way of drug discovery (digging up plants) vs. AI's targeted approach. The hosts discuss the potential for virtual cell simulations to eliminate clinical trials within 5 years. Alex argues that classical computing, not quantum, will likely solve the virtual cell, as it's a data problem. Dave notes the biotech community's enthusiastic embrace of AI, contrasting with other industries.

The hosts discuss the implications for Bitcoin. Michael Saylor dismisses the threat, saying Bitcoin will harden. Brian Armstrong is forming a $150M coalition for a quantum-proof upgrade (BIP 360). Salim is optimistic that the protocol will evolve. Alex argues the real threat to Bitcoin is not quantum but AI—either breaking hash functions or making Bitcoin irrelevant as AI agents create their own currencies. He also questions the need for a long-term store of wealth in a superintelligent economy.

Alex frames it as a race: China has manufacturing scale, US has AI models. He hopes the competition will spur US robotics. He mentions his own experience controlling a Unitree robot at MIT and the upcoming Professional Robotics League match. Dave notes the supply chain for physical robotics is miles behind, creating entrepreneurial opportunities. The hosts debate Mark Cuban's prediction that humanoid robots will merge into the environment and disappear as distinct entities.