Anthropic researcher Nicholas Carlini's long piece asks: are large language models worth it?

Anthropic researcher Nicholas Carlini published a piece questioning whether large language models are worth it, arguing they may be transformative in coming years but have already caused real harm with serious potential risks. The article, expanded from his keynote at last month's language model conference, focuses on harms LLMs and general AI have caused and may soon cause, listing research questions worth pursuing. The author states the views do not represent Anthropic's position.