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

GPT-5.5 hallucinates 3x more than MIT-licensed GLM-5.2

arrowtsx.dev

Post reply on HN
Print view View original

301–310 of 318 posts

Re: GPT-5.5 hallucinates 3x more than MIT-licensed GLM-5.2

#301

Post by gymbeaux »

O5vYtytb wrote: post #73
andybak wrote: post #55

Earlier quoted context omitted.

I can't help but feel that people continually underestimate how bad human written code becomes over time. The exception is probably single-person passion projects or open source projects that maintain quality governance over time. I strongly suspect most closed source code developed under commercial or internal pressure is pretty awful after a few years of development. All LLM code has to do is suck less than existin…

I've been sent code from vendors that didn't even compile, long before llms were a thing. Most shops that aren't primarily software have really really terrible software.

True. I used to get code that wouldn't compile all the time from Infosys "developers" in India circa 2016. Perhaps now with LLMs they still do basically no work, but at least the code compiles? That being said, I'm not sure that paying for Opus/GPT/Gemini makes sense for a company like Infosys that caters to the dumb C-Levels of large corporations who think "why pay X for U.S. devs when we can pay X/5 for foreign devs?" - such companies are fucked in the long-term anyway. Why would Infosys voluntarily pay gobs of money when such U.S. corporations seem to be content with their output as-is?

Re: GPT-5.5 hallucinates 3x more than MIT-licensed GLM-5.2

#302

Post by gymbeaux »

xvinci wrote: post #29
gymbeaux wrote: post #17

Earlier quoted context omitted.

Those numbers are abysmal. Should we really be using LLMs to write our code? I have a theory- LLMs can spit out code that gets the job done and looks ok, maybe even great, but contains small “anomalies” that compound over time. An enterprise app developed entirely with LLM-happy devs might end up virtually unmaintainable. I’m not sure how to explain it, but the more I see LLM-written code the more I feel it’s bad cod…

Not my observation. If you never look at the code and dont have basic guardrails in place (linters, architecture tests, some guidelines for best practices) - probably. But as soon as you do minimal reviews and high-level corrections, applications turn out just fine. Can there be bugs? Sure. That's the price of not reading or understanding every line. It should depend on the criticality of your software how much of th…

What I'm hearing is "thoroughly reviewing AI generated code would defeat the purpose, so we give it a cursory glance and it seems to be decent code", and that's my point - it does indeed seem to be decent code but I think we're all kicking the can down the road when we operate this way. If the alternative means there's no gain to be had by using LLMs to write code, so be it. Maybe that's the answer. Maybe we shouldn't be relying so much on AI to write our code.

I think LLMs are great for writing small snippets of code that really only have one "best answer" (something simple like writing an array to a CSV), and internal tools, where bugs and security vulnerabilities usually aren't a big deal.

Re: GPT-5.5 hallucinates 3x more than MIT-licensed GLM-5.2

#303

Post by gymbeaux »

realaleris149 wrote: post #45
gymbeaux wrote: post #17

Earlier quoted context omitted.

Those numbers are abysmal. Should we really be using LLMs to write our code? I have a theory- LLMs can spit out code that gets the job done and looks ok, maybe even great, but contains small “anomalies” that compound over time. An enterprise app developed entirely with LLM-happy devs might end up virtually unmaintainable. I’m not sure how to explain it, but the more I see LLM-written code the more I feel it’s bad cod…

Take a look at a sufficiently old random internal repo which was not written with LLMs and compare. My observation is that they are equally bad and hard to maintain or even more so than the new ones. One thing I’ve noticed is that the LLM assisted ones have a lot more comments which is nice but take more time to read.

I recognize that I seem to be an above-average developer and so my standards for LLM output are higher. I'm seeing enough comments where people are comparing LLM output to offshore dev output or even onshore mid-level or junior-level output. I guess my response would be something like "why are we celebrating this as a huge success for mankind?" If we dump obscene amounts of electricity and silicon into datacenters, we can produce mediocre or "adequate" software engineers?

I agree that LLM output tends to be absolutely flooded with comments. I think it helps to quickly see "is the output what I expected, does it do everything it should be doing in this method?" but I also think most people don't bother reading the comments.

Re: GPT-5.5 hallucinates 3x more than MIT-licensed GLM-5.2

#304

Post by unshavedyak »

coldtea wrote: post #287
unshavedyak wrote: post #203

Earlier quoted context omitted.

I often wonder how humans "know" things. I suspect (ignorant armchair) we have some ability to signal strength of those facts, via repetition. Without this layer of introspection i imagine LLMs can never avoid hallucination. It obviously breaks down with humans too, given we so easily hallucinate and confuse things we "know". However i still suspect we're more reliable at probing information we've experienced vs not.…

> I suspect (ignorant armchair) we have some ability to signal strength of those facts, via repetition. Without this layer of introspection Isn't that precisely what the LLM training does? It signals strength of those facts, via repetition.

Yea but i don't think the LLM can inspect that. Eg it's similar to a code function , the code executes but the code doesn't have the ability to inquire about the code itself.

In that silly example/thought, the LLM would effectively need the ability to query the strength of a fact/spatial region/etc.

Right now i believe the LLM is more just the output of those weights. It has no way to inspect the strength of the signal. Eg it doesn't know if blue in "The sky is " is a strong or weak signal, it just predicted that next token.

If we could somehow encode strengths along with every token, eg "The sky is " or something we could perhaps give it a sense of [un]certainty. Though i imagine it would look differently than that since we'd want to encode this information in some sort of multi-dimensional space, rather than purely by token - since tokens aren't that valuable. Eg the "knowledge" in an LLM goes beyond tokens, and so too should signal strength.

I'm of course speculating on all of this and i have no clue on anything.

Re: GPT-5.5 hallucinates 3x more than MIT-licensed GLM-5.2

#305

Post by calla »

jmalicki wrote: post #186
MattRogish wrote: post #181

Earlier quoted context omitted.

I'd have to imagine there are wildly diminishing marginal returns to additional SFT/post-training passes. There are a bounded number of (useful) derivations/combinations of Duff's device. If Frontier Labs wish to reduce hallucinations on factual things, they will have to hire people (or the data providers will need to) to do fundamental research above and beyond what is available in extant literature and the web. IE…

As a side gig, I write novel software that solves problems no existing software does, that existing LLMs have difficulty reproducing, purely for the purpose of existing as LLM training data. There are journalists being hired to write Atlantic-worthy articles that exist only as LLM training data, because they're getting paid more than the Atlantic would pay them for it. It's insane. Yes, they are hiring the experts th…

Hi i would like to hear More about this.Where is the citation?Or can you send more information ? Who is paying

Re: GPT-5.5 hallucinates 3x more than MIT-licensed GLM-5.2

#306

Post by calla »

jmalicki wrote: post #186
MattRogish wrote: post #181

Earlier quoted context omitted.

I'd have to imagine there are wildly diminishing marginal returns to additional SFT/post-training passes. There are a bounded number of (useful) derivations/combinations of Duff's device. If Frontier Labs wish to reduce hallucinations on factual things, they will have to hire people (or the data providers will need to) to do fundamental research above and beyond what is available in extant literature and the web. IE…

As a side gig, I write novel software that solves problems no existing software does, that existing LLMs have difficulty reproducing, purely for the purpose of existing as LLM training data. There are journalists being hired to write Atlantic-worthy articles that exist only as LLM training data, because they're getting paid more than the Atlantic would pay them for it. It's insane. Yes, they are hiring the experts th…

Hi so mercor is paying them? What's the prompt? Need more info!

Re: GPT-5.5 hallucinates 3x more than MIT-licensed GLM-5.2

#307

Post by jmalicki »

HarHarVeryFunny wrote: post #297
jmalicki wrote: post #251

Earlier quoted context omitted.

Yes, which is why for some things I've gotten paid as much as $1500 per training example generated. AI labs don't care about cost prohibitive.

Could you give a specific example - just the outline/shape of what it consisted of (or more detail if you care to)?

It was something that took me 8 hours of very hard intellectual work. It was, like some of the higher level contracts, something where I had to create a training example for AI out of whole cloth and prove it was roughly at the boundary of things AIs could solve, and I had to write very detailed rubrics to judge potential solutions and apportion partial credit.

I'm not comfortable saying more.

Re: GPT-5.5 hallucinates 3x more than MIT-licensed GLM-5.2

#308

Post by giancarlostoro »

x3cca wrote: post #232
giancarlostoro wrote: post #177

I wonder if this is what a “Minimally Viable LLM” looks like. I often wonder how much of an LLM do you need before you can just shove a bigger context Window and any dynamic knowledge content to it like a PDF or markdown file to give it knowledge outside of its training data. I feel like LLMs don’t need more data they just need to be refined.

You might be interested in this model. It's a densely trained on math whuch let's it punch way higher than it should https://github.com/WeiboAI/VibeThinker

Cant open the link without an account is it private or is that just GitHub being annoying?

Re: GPT-5.5 hallucinates 3x more than MIT-licensed GLM-5.2

#309

Post by flextheruler »

coldtea wrote: post #285
stevemk14ebr wrote: post #169

Earlier quoted context omitted.

An LLM is not thinking, assuming and relating it to thought and universal truths is nonsense.

How is it's process (inputs, training weights, processing those weights based on an initial given state -prompt-) different than how a brain thinks? And is it a difference of architecture and scale (like an ENIAC and a Apple Silicon M5) or a difference of fundamental operation?

While I'm sure professionals in neuroscience or philosophy can expound on this topic in a more satisfactory manner than myself, I think a relevant point is that despite all our efforts we still don't know how brains think.

Re: GPT-5.5 hallucinates 3x more than MIT-licensed GLM-5.2

#310

Post by idontpost »

claytongulick wrote: post #197
wolttam wrote: post #113

> it is clear that actual intelligence has plateaued significantly. > Moving forward, the industry cannot continue to train bigger and bigger models since their intelligence not only plateaus but often will get worse These are wild claims - why are we concluding that bigger models and more data = more hallucination? That’s actually the opposite of what’s been happening over the last couple years. Some models may stil…

My impression is that the fundamental issue is that LLMs attempt to extract reasoning (executive execution) from data (relationship between tokens). There's an open question about whether this is theoretically possible, but it doesn't seem like it to me. Human generated data is an effect of reasoning. Attempting to extract executive function from it is kind of like taking an anti-derivative of a function. This has al…

[dead]
Post reply on HN