Post

I know "LLMs are incredibly capable tools but not thinking" is a neglected middle position but I'm still on team "don't call it 'thinking'". (It's fine in prompts – I'm talking about when we describe the tool.) Anthropomorphism distorts our perspective.

Ethan Mollick@emollick.bsky.social

If you want to read an interesting AI thinking trace try "I want you to suggest two poems that you think apply very well to the current state of GenAI models like you. Don’t just pick popular poems and back justify. Think hard about options first" in GLM-5.2 & Opus 4.8 (very different personalities)

dave

@daveliepmann.com

LLMs are stunningly useful because they efficiently search a mind-bogglingly large space. Natural language as 2-way API surface unlocks fabulous things: "intuition" through recursive exploration, style transfer, small programs that supercharge the feasibility of many tasks.

26/6/26, 6:19 am0 reposts1 like

dave

@daveliepmann.com

26 June

Being Deep Blue for pattern recognition and NLP brings it into the uncanny valley for "thinking" and being a "mind", but anthropomorphizing just has too much baggage. I find it leads me to wrong conclusions and it's abundantly clear it's a frame that leads a significant minority of users to doom.