Reflections on AI at the End of 2025

Re:Reflections on AI at the End of 2025 by russfink on Saturday December 20, 2025 @12:40PM (#46337939)

Practical question: when getting the AI to teach you something, eg how attention can be focused in LLMs, how do you know it’s teaching you correct theory? Can I use a metric of internal consistency, repeatedly querying it and other models with a summary of my understanding? What do you all do?

[ Parent | on HN ]

Re:Reflections on AI at the End of 2025 by layer8 on Saturday December 20, 2025 @02:43PM (#46338919)

> What do you all do?

Google for non-AI sources. Ask several models to get a wider range of opinions. Apply one’s own reasoning capabilities where applicable. Remain skeptical in the absence of substantive evidence.

Basically, do what you did before LLMs existed, and treat LLM output like you would have a random anonymous blog post you found.

[ Parent | on HN ]

Re:Reflections on AI at the End of 2025 by akomtu on Saturday December 20, 2025 @10:08PM (#46341870)

In that case, LLMs must be written off as very knowledgeable crackpots because of their tendency to make things up. That's how we would treat a scientist who's caught making things up.

[ Parent | on HN ]

Re:Reflections on AI at the End of 2025 by jennyholzer2 on Saturday December 20, 2025 @01:01PM (#46338103)

You don't

AI is for idiots

[ Parent | on HN ]