The only reason to have an LLM emit sentences like “I understand” is to make it more appealing than a search engine and increase the likelihood that a user will return; that is, it’s another way of maximizing customer engagement.
Sci-fi author Ted Chiang has already written a couple of good essays puncturing the illusions behind the AI industry.
His latest is another masterpiece, using the publication of a “constitution” for Claude by Anthropic as its jumping off point to discuss the absurdity of imagining a program like Claude as a conscious being, and the many ways Anthropic – even as it pushes the fantasy of conscious LLMs on us – doesn’t act in the ways it would have to if such a thing were actually true.
#AI #LLMs #TedChiang #Anthropic
