Guardrails removed spam, off-topic, unclear, or duplicate replies.
Ask a question below.
Published answers will appear here.
For example, lately GPT-5.5/5.6 max reasoning incorrectly equally weighs well-known philosophical positions on consciousness that are non-materialist/non-computational, with the maximally consistent scientific view that it is computational. It only backtracks when reminded:
It finally does a decent job summarizing my own counter-arguments when reminded of basic scientific facts! I can't help but recall the meme of Copernicus/Galileo/etc being told he was crazy by an LLM trained on only pre-15th century data...
Guardrails removed spam, off-topic, unclear, or duplicate replies.
Ask a question below.
Published answers will appear here.