Jess notes Elon Musk does not control Tesla and owns about a fifth of it. The segment argues the SpaceX-Tesla merger is driven by Musk’s desire for control, his pay package, and avoiding many earnings calls; an audience poll found 83% say it is mostly about Musk control.
Dave argues labs centralize profits while radiating risk to society; if regional banks get hacked, taxpayers clean up a Katrina-style mess. He says labs should carry enormous liability insurance or pay hundreds of billions annually into a contingency fund, rather than forcing Washington to cover the bill.
Sam theorizes SpaceX faces a huge lockup expiry in the spring. If it merges with Tesla first, the unlocked shares become a smaller percentage of a larger combined company, which could make the supply shock easier for public and private investors to absorb. He says nobody else has made this point.
Sam says access to nonpublic frontier models creates a “wild gap” between lab experience and public APIs. Dave recommends only research teams touch the frontier; everyone else must use public versions, because teams lose track of real customer experience within days. He says this disconnect leads to bad product work.
Sam fed Google Workspace’s updated terms to Claude and says the summary told him users are responsible for everything their agents do. Dave calls that insane: if a company runs a centralized service and earns premium revenue, it should own the liability for what its agent does.
Britt says Netflix uses so much data to pick scripts and priorities that it stopped focusing on great stories. Sam calls this “priority slop”: algorithms split content into buckets and optimize pieces, leading to shows that feel assembled rather than written, even before broader AI content slop arrives.
Sam says he gave his bot a single prompt to write his firm’s entire investor update deck, including the cover slide. The bot’s version was “stunningly good” and better than the deck the partners wrote; Sam attributes this to the bot having far more context on every deal and thought.
Dave argues Tesla will win long term because it has route data from all its customer cars, while Waymo only has data from cities it operates in. Sam says Waymos often seem to know the right path but then add ten minutes, though he still uses them.
Dave says it is unclear how to invest in cybersecurity because the next model is what you need to secure it, leaving mostly professional services. Sam’s seed company instead stores important documents and photos as hard copy in a mine near the North Pole, literally cold storage.
Sam says OpenAI’s PHP vulnerability discovery leaked, forcing the company to announce it with a patch. Dave calls these findings the best marketing frontier models could get, while they also discuss private labs sitting on zero-days like a skeleton key that the public cannot see.
