transcript
show notes
The public resignation of Jacob Coxon — who told the world his industry is “gambling with our lives” — buried two other chilling warnings from inside the AI industry this week.
The first is “An Alien Mind,” an essay OpenAI’s chief scientist Jakub Pachocki published on September 6, in which he describes his own creation as an intellect nobody fully understands and says no lab has solved alignment well enough to keep scaling at full speed.
The second is Anthropic’s September threat report, which lays out what people have already done with Claude: a Russian espionage group whose AI agents rewrote their own malware until security products stopped catching it; hotel WiFi networks turned into software delivery systems; one person who built a searchable doxxing database of political dissidents; a Wagner-funded radio station in the Central African Republic where the model wrote the loyalty contracts and the scoring rubric used to decide which journalists to fire; and a cell in northern Yemen running three weapons programs, including a multi-stage ballistic missile with a stated range goal above 2,000 kilometers.
In this episode I go through the report, explain why the labor curve that used to limit what a small group could do has collapsed, and lay out the questions I submitted to a CNN producer for the executives running these companies — including whether Dario Amodei would support the Ban Artificial Superintelligence Act, the bill Sen. Bernie Sanders and Rep. Greg Casar announced on September 3, which would carry penalties comparable to those for unlawfully developing nuclear weapons.
This is a public episode. If you'd like to discuss this with other subscribers or get access to bonus episodes, visit www.theripcurrent.com/subscribe
