Ask HN: There is current insistence upon learning what LLM's do under the hood

1 points by joebig 19d ago ↗ HN

I recognize the value of it in so far as to not anthropomorphize or mystify agents or be duped by them. But beyond these aspects, for an everyday user of ChatGPT or Gemini or Claude_what have you, is there any way the said awareness ('scientia' of it all) objectively beneficial for better utilization?

Put another way, having read a grimoire, I must get fancy spell upgrades, otherwise it's not worthwhile.

2 comments

[ 0.30 ms ] story [ 11.2 ms ] thread

lwarfield 19d ago ↗

Personally I'm curious to the point of doing borderline LLM archecture research. This come from genuine curiosity, and not a want to use LLMs better.

I haven't gotten much out of for using LLMs though. It makes me understand the short comings of LLMs, and I feel like I got an early insight into how important context management is.

HarHarVeryFunny 19d ago ↗

I think there is value in at least having some basic understanding what an LLM is - a statistical predictor, not something brain-like that is reasoning about facts, since it should change your mindset when using them.

I don't expect the public to do this though - I suspect there is really too big a gap from what LLMs actually are to what they appear to be for most non-technical people to be able to get their head around it.