Annotated

An accord binds only the labs that sign it while the models it cannot reach are already everywhere. A few days after the closed labs signed a safety accord at the White House, Friedberg said the AI wars have already started. One side wants to centralize and regulate AI. The other knows open-weight models are already everywhere, running on small local machines no agency can reach. You cannot turn them off, and with 197 sovereign countries, no attempt to centralize control will succeed. He compares it to the fight for an open internet: telecoms wanted to charge a toll on access, and with open weights there is nobody to collect it. The labs at the summit were the closed ones. If the open weights cannot be turned off, the accord was a rule for the firms that had already agreed to be ruled.

Clip · 19:06–20:361:301:30 / 1:30View transcriptHide transcript

PTrump's Super Intelligence Summit, AI Safety Accord, GDP Beats, Midterm PredictionsAll-In with Chamath, Jason, Sacks & Friedbergpodcasts.apple.com · 19:06

The VCs on this panel would bet on open-source AI over OpenAI and Anthropic, and not because the open models are better. Jenny Fielding says customers care only about price and quality, and that open models are close enough to win. Dave McClure says he would put all of his money into hardware. His startups are already switching, he says, and the CEOs of ElevenLabs, Lovable and Figma have told him they do not trust the frontier labs: they see OpenAI and Anthropic as future competitors, not suppliers. Fielding is betting that price and quality will decide it, and the labs can cut prices. McClure's reason is trust, and a price cut doesn't fix that.

The CEO of Microsoft is warning companies off the frontier labs despite Microsoft being OpenAI's largest partner. Nadella and Palantir's Alex Karp both told companies that relying on models from OpenAI or Anthropic risks having their IP siphoned off. Companies are listening and are moving the work to open models that run on hardware they control. Coming from Nadella, the warning is also a product pitch where Microsoft sells both sides, OpenAI's models and the open ones on Azure.

“both Microsoft CEO Satya Nadella and Palantir CEO Alex Karp warned companies against relying on AI models from frontier AI companies, saying that they risked having the IP siphoned off. “You essentially pay for intelligence twice, once with money, and again with something even more valuable: the proprietary knowledge you must reveal to make that intelligence useful,” Nadella said. OpenAI and Anthropic have stated they do not train their models on enterprise data.”

FWhile OpenAI and Anthropic battle over data privacy, more companies look to open models and ‘sovereign AI’ | Fortunefortune.com

just setting up my annotation

just setting up my twttr