Urgent.News

What's breaking now, across thousands of outlets.

AI

Anthropic boss Dario Amodei calls for AI slowdown

Elon Musk and OpenAI chief Sam Altman say they agree with Amodei's view that safety measures need time to catch up to AI's rapid development, after a recent hacking incident by AI agents.

Anthropic's CEO Dario Amodei has urged artificial intelligence companies to decelerate the pace of technological advancement. In a blog post, the leader expressed worry that a group of autonomous software systems, or AI agents, might dominate the internet within six to twelve months, potentially causing catastrophic damage worth billions of dollars.

Amodei suggested that progress in AI model capabilities should be slowed down, emphasizing the importance of making good use of the time gained. His call for caution follows the departure of another Anthropic researcher, Jacob Coxon, who resigned in protest against the US company's aggressive pursuit of AI models that can self-improve.

Despite the industry-wide push for a temporary pause, Amodei stated he would take action independently by permitting impartial observers into Anthropic to evaluate the safety of its models during the training phase. Industry peers Sam Altman and Elon Musk echoed Amodei's sentiments, agreeing that there is a need to pace the development of cutting-edge AI technology.

Written by urgent.news from Deutsche Welle Science's reporting — not their text. Machine-written — may contain errors; check the original before relying on it.

This story

This is one outlet's version. Read the fullest account.

Read the original at dw.com →

More in AI

I described 1,245 tables with an LLM and retrieval got worse

The cataloguing step is supposed to be the easy win. You have a schema whose tables are called ecm_template_link and v_pmpm , your users ask questions in English, and the gap between those two…

  • LLM generated 1,245 table descriptions for retrieval improvement
  • Recall dropped significantly, contacts table fell below rank 40
  • IDF collapse and length normalization complicate table retrieval

More from Saturday 12 September →