OpenAI, Anthropic, xAI chiefs urge slowing AI over RSI fear

The weekend remarks stand out because they came from fierce business rivals — Anthropic's Dario Amodei, OpenAI's Sam Altman and xAI's Elon Musk — who compete directly in the same market yet converged on the same warning. That shared stance is itself the news: these are not outside critics but the people building the technology.

The article frames the moment as the latest step in a fast arc. In 2022, ChatGPT was still hallucination-prone; today, the field is focused on recursive self-improvement, the idea that a system could improve itself with little to no help from humans, each advance producing the next. The executives' concern is that this milestone could slip beyond human control.

What the warning ultimately amounts to hinges on timing. Some AI executives think the industry is three to five years away from self-improving systems that need little or no human help, so the debate is likely to turn on whether that estimate holds — and on how the labs that signed onto the warning act on it.