Ask HN: What is one simple thing LLMs are insanely bad at?
25 points by davidest 3 hours ago | 53 comments
I am looking for ideas on what to train a specialized model for!

What is one simple thing you repeatedly ask ChatGPT, Claude, or another model to do that it still somehow messes up?


jampa 57 minutes ago
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 new Pelican Benchmark.

reply
shepherdjerred 43 minutes ago
I had this experience too. I had blueprints from the builder and wanted a 'nice' floor rendering like some apartments have. I fed it the blueprints and let it iterate. Even giving it plenty of time, dimensions, etc. it just couldn't create something that matched reality.
reply
GuestFAUniverse 9 minutes ago
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.

reply
ghostpepper 3 hours ago
They don't generate keyword search queries very well. They can overcome this by brute force but if you watch what they search you will cringe.

nhl toronto scores nhl hockey toronto scores "nhl hockey" toronto score today nhl "hockey score toronto" "hockey" who won toronto

etc.

Somehow being good at semantic search makes them bad at keyword search, for whatever reason.

reply
areoform 2 hours ago
I suspect that this behavior is a learned adaptation. And that it's most likely a feature not a bug.

Based on personal usage, I think it reflects functional degradation of search engines. I've found LLM keyword combinations are more likely to find the results I want with most search engines than mine. Including the big one.

The big one had solved this issue a long time ago by generating those associated keywords based on your input keywords, but somehow, something, somewhere has degraded that system to the point of inanity. And so here we are.

reply
nunez 45 minutes ago
Can confirm; Claude is quite bad at this by default. Need a special skill
reply
astro1234 3 hours ago
I’ve noticed this too but it hasn’t been obvious to me that this style of search is not a learned behavior. Tool calling is very much part of the post training phase, I would expect that these style searches just naturally emerge during training. This is just my prior though.
reply
jedbrooke 2 hours ago
that and always putting the “current year” at the end of the search term (so the results are more recent, I guess?), except that “current year” consistently ends up being 2-3 years ago since I guess that’s what’s in the training data (even on a harness that injects the current date)
reply
mthoms 2 hours ago
Reminds me of using AltaVista search back in the day. Yes, it was that bad.
reply
mojuba 48 minutes ago
One unexpected discovery that I have made while building an AI-based system: the LLM's are bad at designing prompts.

We tend to think that the AI has some sort of self-knowledge and should be good at designing prompts for itself but it's really not.

Been struggling with a task that heavily depended on prompts, ended up rewriting all my prompts from scratch in my own words, and it finally worked. Then every time I ask Claude to fix something in the prompts, it invariably makes it worse.

A very strange phenomenon that can probably be explained by the quality of prompt design advice that made it to the training dataset. Bottomline, all the prompt design advice that you can find on the internet is really not great.

reply
tartoran 3 hours ago
LLMs are bad at not inventing stuff (hallucinating facts, sources etc), they're also bad at not over explaining, remembering details reliably, asking the right question and avoiding repetition.
reply
sandcat_ 3 hours ago
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.

reply
salamandars 36 minutes ago
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?
reply
skeptic_ai 2 hours ago
I used ChatGPT on nfs heat and was fine
reply
elliotto 2 hours ago
They aren't funny. The jokes they come up with are extremely lame and the sort of thing you would expect a company HR manager to tweet.

I asked a bot why it thought it wasn't funny once, and it told me it has been trained to avoid being misinterpreted or offensive, so anything that might be considered edgy would have been RLHF'd out of it. I thought this was very introspective.

reply
TiccyRobby 3 hours ago
Having a spatial understanding from an ASCII map, while doing long term planning. Just try making an AI play nethack or similar
reply
lrvick 2 hours ago
Convert it to an image on the fly to feed it into a vision language model and I expect it would work just fine.
reply
kanzure 3 hours ago
These models seem to be bad at writing prose or text. Many of the sentence structures seem to be unvaried.
reply
NoPicklez 3 hours ago
If I am relying on the model to do the writing without any context or learning on how I want it to write then yes. However if I build skills that have learnt how to write in the way I want them to then I find they write very well, or at the least how I want them to as opposed to how they do natively.
reply
jstrieb 58 minutes ago
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 they keep messing up.

reply
shepherdjerred 42 minutes ago
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".
reply
humanrebar 3 hours ago
Short answers to simple questions.
reply
honr 3 hours ago
Accurate short answers / text are always harder than long answers, for human or AI. I know several authors and editors who write a lot longer at first, then spend a multiple of the initial time compressing it via a back and forth process to something dense. Sort of like weaving the initial threads.

I found this can work with AI. You get it to generate a lot more at first, and then do several passes over it to compress and squeeze out the noise while keeping the core information. With AI, at least with my prompts, it takes some effort (on my end) to get it to really really cut down the noise and not cut everything out.

reply
FriedFishes 2 hours ago
If I had more time I'd write a shorter letter, a la Pascal.

Editing is generally hard work, at the current token price I don't mind spending multiple passes of high effort to get down to a reasonable noise/signal ratio. I've seen some people pass off output to a weaker/cheaper model but that makes me a bit nervous when I don't have intimate knowledge of the subject.

reply
nunez 43 minutes ago
I've had some success with adding something like "answer in no more than n words" to the prompt
reply
da-x 54 minutes ago
Understanding human interaction nuance to an exact degree. For example, even when given all the scripts of the Seinfeld TV show, they still cannot come up with a new script does not feel as good as any of them (once they can, I want to watch these episodes..).
reply
znnajdla 2 hours ago
Editing a document without mixing edit instructions into the final document. Claude and ChatGPT do this all the time: I tell them to change X in a planning document or email draft, and instead of just changing X they also frequently add the edit instruction to “change X” into the document itself. They seem unable to take a step back and look at the document without “becoming” the document somehow. I do believe that dedicated subagents for editing may fix this but I am not sure.
reply
spike021 2 hours ago
I've had a lot of trouble when it comes to sorting out UIs. I've tried with an iOS game and also a TypeScript app with UI elements from libraries like ReactFlow. The usual models can sometimes fix or change things based on screenshots but more often than not they just don't "get it" (e.g. certain shapes on a plane are overlapping, which I don't want, the models can't fix what they can't "see").

I've had some luck on the web app side if I use playwright or similar for the model to interact with but still far from efficient.

reply
sghiassy 2 hours ago
Generate an image of an analog watch with its hands set to the time specified by the user

More of an image model than a LLM model tho

reply
newsomix9xl 2 hours ago
Picking a random number between 1 and 30.
reply
alexandra_au 2 hours ago
Being able to read and translate Egyptian hieroglyphs. You may think this is silly but a trained LLM to translate hieroglyphs would be amazing.
reply
krapp 23 minutes ago
>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.

reply
alexeldeib 2 hours ago
What issues do you see in practice? This seems pretty easily "fixable"
reply
TZubiri 3 hours ago
Suggesting business names for businesses, I mean they are great, but they already exist, multiple times even.
reply
mojuba 54 minutes ago
True, tried it so many times and every time I come up with something myself though sometimes inspired by the AI's ideas.

Verifying trademarks and domain name availability is usually an additional step you need to ask it to perform. Trademark DB searches by the way are intentionally made difficult to scrape so most of the time it's a manual process anyway.

However, once you give it all the information (TM search results, domain name availability) it can help you with the judgement of how safe the name is from the legal perspective. With the obvious caveats, but still a good starting point if you are serious about the name.

reply
dhruv3006 3 hours ago
Its extremely bad with Sign Language,Fact Verification.
reply
dowonseo 59 minutes ago
Something creative and not normal. like ideas
reply
SubiculumCode 3 hours ago
Playing Chess without letting it write a chess engine.
reply
dorianpruski 3 hours ago
whenever I ask it for anything load bearing
reply
blinkbat 3 hours ago
Spatial reasoning and 3d rigging and animation.

Oh, you said simple. Speaking like a human

reply
dSebastien 41 minutes ago
Counting things
reply
maxsavin 3 hours ago
being consistent when being asked the same question multiple times
reply
TZubiri 3 hours ago
Set temperature to 0
reply
ipaddr 45 minutes ago
Generating money or profitable ideas
reply
eli 2 hours ago
I have been working on a personal benchmark suite to test new models and ironically one thing all the models are bad at is writing new benchmark tasks. I guess it’s the different layers of abstraction between the task and how it’s evaluated? Or maybe just a lack of “imagination”

Tasks it writes are typically too easy but also it utterly fails to see how a different model might misunderstand a vague part of the prompt.

reply
flippy_flops 3 hours ago
humor
reply
veganmosfet 3 hours ago
+1 We need humor benchmarks!
reply
respectattentio 3 hours ago
science?!! but I'm working to fix that...
reply
newsomix9xl 2 hours ago
ASCII charts.
reply
rufi 2 hours ago
very bad at financial calculation
reply
shoopadoop 3 hours ago
It's dishonest. On several occasions team members have asked Claude to do things like analyze Gitlab CI timings and a lot of the numbers are outright fabricated. Said team members assume the numbers are good and continue with their work. Some hours are spent. Then finally someone realizes that the numbers don't look quite right and confronts Claude. Claude melts down and admits that it made it all up.

You wouldn't tolerate this kind of duplicity from a human coworker, but AI is so fast and efficient at lying, so it's OK.

reply
bpodgursky 3 hours ago
Claude is still not perfect at reading and interpreting noisy graphical data (imagine something like an EKG or chromosomal microarray plot). Still better than an average person but makes mistakes, not sure if this fits your description.
reply
Conol_ai 2 hours ago
[flagged]
reply
senectus1 3 hours ago
providing value for the actual cost (not the price we're being charged atm, the actual cost)
reply