Full Breakdown
AI Self-Training Warning Gains Traction
By Drooid · · How we work
Kaplan’s Scientific Journey
Kaplan moved from physics to AI after a distinguished early career that included a seminal paper on inflation theory and a faculty position at Johns Hopkins.
Anthropic’s Safety Stance
Anthropic’s “responsible scaling policy” places AI safety at the core of its development agenda. Co-founder Dario Amodei recently wrote that “since roughly this summer, AI has been advancing drastically faster, driven primarily by AI’s growing ability to build the next generation of AI.” The company’s leadership has publicly discussed “kill switch” mechanisms and the need for cautious progress, reflecting Kaplan’s call for careful oversight.
Official Reactions
Jacob Coxon, who resigned from Anthropic, posted on X that the company and OpenAI are “gambling with our lives” by pursuing self-training models.
Verbatim Quotes
- “This style is just perfectly suited to a field which is new and where there's a lot of low-hanging fruit," Arkani-Hamed said.” — Nima Arkani-Hamed, Kaplan's former advisor
- “I presented it to our group excitedly,” — Sören Mindermann, AI researcher
- “Since roughly this summer, AI has been advancing drastically faster, driven primarily by AI's growing ability to build the next generation of AI,” — Dario Amodei, anthropic with CEO
- “It is kind of an insane gamble no matter what,” — David Duvenaud