top | item 44420602

Anthropic's Claude AI became a terrible business owner in an experiment

3 points| gdrift | 8 months ago |techcrunch.com

1 comment

order

leakycap|8 months ago

I read this earlier on Anthropic's blog, and their wording was a breath of fresh air

They definitively said it didn't work and the fix was more prompting, which at some point is just pseudo-programming in a LLM