Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

Ask HN: What is one simple thing LLMs are insanely bad at?

news.ycombinator.com

Post reply on HN
Print view View original

51–60 of 85 posts

Re: Ask HN: What is one simple thing LLMs are insanely bad at?

#51

Post by shepherdjerred »

jstrieb wrote: post #42

Giving hints. On math or programming problems, they are overfit to solving the entire thing end to end (presumably for benchmarks). I have had very poor results asking for pointers and hints that don't give away key insights. This has been the case across models I have tested. An architecture with a "judge" that gates responses and ensures a lack of spoilers would probably work better. But this is a simple thing that…

I haven't had this experience at all. I've used Cursor+Opus on homework e.g. to understand algorithms, but I usually prompt it with something like "DO NOT GIVE ME THE ANSWER, I care about understanding and solving this myself".

Re: Ask HN: What is one simple thing LLMs are insanely bad at?

#53

Post by salamandars »

sandcat_ wrote: post #16

Video game tips. Constant mistakes and hallucinations, in my experience. Seen this across a lot of different games. Even in really well documented games, such as OSRS (which has multiple fantastic wikis). Anno 1800 was a recent one I had trouble with, using Claude Opus. Completely made up game mechanics. Rainbow Six Siege, too.

As a noob, how does the end user improve this? What's the best way to make the knowledge from the specialised wiki available to the LLM?

Re: Ask HN: What is one simple thing LLMs are insanely bad at?

#54

Post by krapp »

alexandra_au wrote: post #35

Being able to read and translate Egyptian hieroglyphs. You may think this is silly but a trained LLM to translate hieroglyphs would be amazing.

>You may think this is silly but a trained LLM to translate hieroglyphs would be amazing

Why would it be amazing? We've known how to read hieroglyphs for a long time. It isn't a problem we need computers to solve.

Re: Ask HN: What is one simple thing LLMs are insanely bad at?

#55

Post by GuestFAUniverse »

Context.

At least ChatGPT assumes too much from former conversations (even in unrelated new questions). It always needs a briefing to forget certain assumptions. It rarely asks for clarification instead of assuming too much.

So, it's answer generation is too dependent on tooling, system prompt and cache/memory to really have a guaranteed conversational experience.

Re: Ask HN: What is one simple thing LLMs are insanely bad at?

#56

Post by SpaceNoodled »

salamandars wrote: post #53
sandcat_ wrote: post #16

Video game tips. Constant mistakes and hallucinations, in my experience. Seen this across a lot of different games. Even in really well documented games, such as OSRS (which has multiple fantastic wikis). Anno 1800 was a recent one I had trouble with, using Claude Opus. Completely made up game mechanics. Rainbow Six Siege, too.

As a noob, how does the end user improve this? What's the best way to make the knowledge from the specialised wiki available to the LLM?

Just read the wiki instead?

Re: Ask HN: What is one simple thing LLMs are insanely bad at?

#57

Post by kiernan »

jampa wrote: post #43

Serious answer: no model ever gets close to writing an architectural floor plan that makes sense. They understand all the rules and best practices, they can (sometimes) spot a bad idea in a floor plan, they can describe a good floor plan. But ask them to make one, even if you give it every detail (even a "node graph" of rooms), they will still output nonsense. Same for text and image models. Floor plans should be the…

Could it write a deterministic constraint solving program that at least gives it a head start at narrowing down options?

Re: Ask HN: What is one simple thing LLMs are insanely bad at?

#58

Post by realitysballs »

jampa wrote: post #43

Serious answer: no model ever gets close to writing an architectural floor plan that makes sense. They understand all the rules and best practices, they can (sometimes) spot a bad idea in a floor plan, they can describe a good floor plan. But ask them to make one, even if you give it every detail (even a "node graph" of rooms), they will still output nonsense. Same for text and image models. Floor plans should be the…

10000% , imho opinion core issue is that cd-level architectural plan-sets en masse are overly shielded by design firms and clients. Diffusion/AI vision has a data problem in this regard.

Also, LLMs fundamentally lacks a spatial intuition or comprehension of orthographic /sectional drawings.

Re: Ask HN: What is one simple thing LLMs are insanely bad at?

#59

Post by jampa »

kiernan wrote: post #57
jampa wrote: post #43

Serious answer: no model ever gets close to writing an architectural floor plan that makes sense. They understand all the rules and best practices, they can (sometimes) spot a bad idea in a floor plan, they can describe a good floor plan. But ask them to make one, even if you give it every detail (even a "node graph" of rooms), they will still output nonsense. Same for text and image models. Floor plans should be the…

Could it write a deterministic constraint solving program that at least gives it a head start at narrowing down options?

I tried doing something like this Ox Alpha with Opus advisor, having it work layer by layer (specs -> rooms -> room graphs ...), but each deliverable ended up a mess.

The curious thing is when I pointed out the flaws it fixed them quickly, but it's not something it can do without supervision, and supervising it takes more effort than doing the blueprint myself (to be fair, I'm not an architect, so I'm not the best at steering an LLM for this task).

Post reply on HN