ALT Late last week, everything exploded when former Anthropic AI researcher Jacob Coxon, in an exclusive interview with the Wall Street Journal, warned that he was “quitting the AI industry” (he wasn’t) over “...fears that the lab and its competitors are racing to build systems they won’t be able to control.” His fears were centered around the creation of “recursive self-improvement,” a still-theoretical concept of AI that trains itself autonomously” and otherwise expressing few specific concerns beyond that “AI labs are unable to control AI,” always phrasing things in the terms of impossible-to-control entities rather than poorly-programmed cloud software running on the infrastructure of the largest companies in the world.

ALT Emily Forlini of Fortune put it best: Hear me out: He claims the AI could kill us someday, but doesn’t point to any projects in the pipeline that could be shut down to avoid this. He says AI companies are moving too fast, but neglects to share screenshots, emails, or specific examples of when this behavior went sideways—when it became clear to leaders at AI companies that the technology was slipping beyond their control, for instance, and how the decision-makers disregarded the warning signs. He doesn’t suggest any new legislation, name problematic leaders that should step down, or post an in-depth look at how Anthropic researches new models and propose a new approach.

ALT This is because, in my opinion, Jacob does not really care about the actual harms of AI, whether we’re talking about Large Language Models or something he imagined while working with the non-profit or PR firm that set up a CBS interview where he claimed that AI that, if we’re talking about LLMs, have model weights of terabytes of memory, would make ten thousand copies of themselves.