It's a shifting goalpost, but one of the things that struck me was how some questions could still be trivial for a fairly qualified human (a doctor in this case) but difficult for an AI model. Reasoning, visual or logic, is built on a set of assumptions that are better gained through IRL experience than crawling datasets and matching answers.
This leads me to believe that much of the future for training AI models will lie in exposing them to "meatspace" and annotating their inferences, much like how we train a child. This is a long, long process, and one that is already underway at scale. But it's what might give us emergent intelligences rather than just a basket of competing yet somehow-magic thesaurus.
Benchmarks are like SAT scores. Can they guarantee you'll be great at your future job? No, but we are still roughly okay with what they signify. Clearly LLMs are getting better in meaningful ways, and benchmarks correlate with that to some extend.
Humans are much better at out of sample prediction than LLMs. And inherently benchmarks cannot be out of sample. So I believe that leads to the disconnect between LLMs getting better and better at in sample prediction (benchmarks) while not improving nearly as much at out of sample (actual work).
We should make a collective git repo full of every kind of annoying bug we (expert developers) can think of. Then use that to benchmark LLMs.
Someone want to start? I've got a Yjs/CRDT collaborative editing bug that took like a week and a half of attempts with Claude Code (Sonnet 4.5), GPT5-codex (medium), and GLM-4.6 many, many attempts to figure out. Even then they didn't really get it... Just came up with a successful workaround (which is good enough for me but still...).
Aside: You know what really moved the progress bar on finding and fixing the bug? When I had a moment of inspiration and made the frontend send all it's logs to the backend so the AIs could see what was actually happening on the frontend (near real-time). Really, I was just getting sick of manual testing and pasting the console output into the chat (LOL). Laziness FTW!
I have the Google Chrome Dev Tools MCP but for some reason it doesn't work as well :shrug:
> took like a week and a half of attempts with Claude Code ...
What kind of expert developer wastes that much time prompting a bunch of different LLMs to end up with a workaround, instead of actually debugging and fixing the bug themselves?
there is a lot of disdain for vibe coding/coders, as Im sure you already know. I was going to post something similar as soon as I read a week and a half of prompts. I pray that any gainfully employed expert coders don't spend 10 days prompting, rather than coding lol
This may be intentional, but I'd like to point out that your basically suggesting that others aggregate high-quality training data for AI companies to use free of charge to replace software engineers.
> We should make a collective git repo full of every kind of annoying bug we (expert developers) can think of. Then use that to benchmark LLMs.
I think any LLM-user worth their salt have been doing this pretty much since we got API access to LLMs, as otherwise there is no way to actually see if they can solve the things you care about.
The only difference is that you must keep the actual benchmarks to yourself, don't share them with anyone and even less put them publicly. The second you do, you probably should stop using it as an actual benchmark, as newly trained LLMs will either intentionally or unintentionally slurp up your benchmark and suddenly it's no longer a good indicator.
I think I personally started keeping my own test cases for benchmarking around the GPT3 launch, when it became clear the web will be effectively "poisoned" from that part on, and anything on the public internet can be slurped up by the people feeding the LLMs training data.
Once you have this up and running, you'll get a much more measured view of how well new LLMs work, and you'll quickly see that a lot of the fanfare doesn't actually hold up when testing it against your own private benchmarks. On a happier note, you'll also be surprised when a model suddenly does a lot better in a specific area that wasn't even mentioned at release, and then you could switch to it for specifically that task :)
I actually started a collection of annoying bugs I’ve seen in the wild. I give the llm the buggy implementation and ask it to write a test that catches it. So far not even a frontier model (Claude Sonnet) can do it, even though they can find and fix the bug itself.
Have you tried the Playwright libraries? Not the MCP, instead telling Claude Code to use the Node.js or Python Playwright libraries directly. I have had some really good results for this for gnarly frontend challenges.
I don't really like MCPs, at least when I'm working with coding agents like Claude Code or Codex CLI. I'd rather let the agents write code that can do anything the underlying library is capable of, rather than restricting them to just the functionality that the MCP exposes.
It's more token efficient too since I don't need to load the full MCP description into my context.
When I have a bug I’m iterating on it’s much easier and faster to have it write out the playwright script. That way it does not have to waste time or tokens performing the same actions over and over again.
> "For example, if a benchmark reuses questions from a calculator-free exam such as AIME," the study says, "numbers in each problem will have been chosen to facilitate basic arithmetic. Testing only on these problems would not predict performance on larger numbers, where LLMs struggle."
When models figure out how to exploit an effect that every clever college student does, that should count as a win. That’s a much more human-like reasoning ability, than the ability to multiply large numbers or whatever (computers were already good at that, to the point that it has become a useless skill for humans to have). The point of these LLMs is to do things that computers were bad at.
College exam takers use those tricks because they are on a time limit and are gaming the system. It's clever and wink wink nudge nudge ok everyone does it. But it's one tiny signal in a huge spectrum of things we use to evaluate people.
Instead, these metrics are gamed and presented as the entire multi special signal of competence for LLMs because it is literally impossible to say that success in one domain would translate the way it might with a good hire.
What I want is something I don't have to guard against gaming. Something conscientious and capable like my co workers. Until then it's google version 2 married to intellisense and I'm not letting do anything by itself.
I don’t think the fact that LLMs can handle small numbers more reliably has anything to do with their reasoning ability. To the contrary, reasoning ability should enable them to handle numbers of arbitrary size, just as it enables humans to do so, given some pencil and paper.
However:
> Testing only on these problems would not predict performance on larger numbers, where LLMs struggle.
Since performance on large numbers is not what these exams are intended to test for, I don’t see this as a counterargument, unless the benchmarks are misrepresenting what is being tested for.
> reasoning ability should enable them to handle numbers of arbitrary size, just as it enables humans to do so, given some pencil and paper.
Or given a calculator. Which it's running on. Which it in some sense is. There's something deeply ironic about the fact that we have an "AI" running on the most technologically advanced calculator in the history of mankind and...it can't do basic math.
This is like saying it's ironic that an alternator in a car cannot combust gasoline when the gasoline engine is right beside it, even though the alternator 'runs' on the gasoline engine.
Or similarly having a gasoline engine without an alternator and making the observation that there's an absurdity there in that you're generating large amounts of energy, yet aren't able to charge a relatively small 12V battery with any of it. It's a very practical and natural limitation, yet in some sense you have exactly what you want - energy - you just can't use it because of the form. If you step back there's an amusing irony buried in that. At least in my humble opinion :-)
Thing is, a LLM is nothing but a prediction algorithm based upon what it trained. So it missing basic calculator functionality is a given. This is why tool usage is more and more a thing for LLMs. So that the LLM can from itself use a calculator for the actual math parts it needs. Thus increasing accuracy ...
If they were selling LLMs as “LLMs” instead of magic code-writing, answer-giving PhD replacements, the lack of basic arithmetic capability would be a given… but they aren’t. Judging a paid service using their own implied claims is perfectly reasonable.
Why is it a given? The universal approximation theorem should apply since addition is a continuous function. Now whether the network is sufficiently trained for that is another question but I don’t think it's a given that a trillion parameter model can’t approximate the most basic math operations.
I think the tokenization is a bigger problem than the model itself.
Easy to answer that one ... predictions are based upon accuracy. So if you have a int4 vs a float16, the chance that the prediction goes off is higher with a int4. But even with a float16, your still going to run into issues where your prediction model goes off. Its going to be a lot less, your still going to get rounding issue, what may result in a 5 being a 8 (just a example).
So while it can look like a LLM calculates correctly, its still restricted by this accuracy issue. What happens when you get a single number wrong in a calculation, everything is wrong.
While a calculator does not deal with predictions but basic adding/multiplying/subtracting etc .. Things that are 100% accurate (if we not not count issues like cosmic rays hitting, failures in silica etc).
A trillion parameter model is just that, a trillion parameters, but what matter is not the tokens but the accuracy as in, the do they use int, float16, float32, float64 ... The issue is, the higher we go, the memory usage explodes.
There is no point in spending terabytes of memory, to just get a somewhat accurate predictive calculator, when we can just have the LLM call a actual calculator, to ensure its results are accurate.
Think of a LLM more like somebody with Dyslexia / Dyscalculia... It does not matter how good you are, all it takes is to switch one number in a algebraic calculation to get a 0/10 ... The reason why i mention this, is because i often think of a LLM like a person with Dyslexia / Dyscalculia. It can have insane knowledge, be smart, but be considered dumb by society because of that less then accurate prediction (or number swiping issue).
Take it from somebody that wasted a few years in school thanks to that issue, it really does not matter if your a good programmer later in life, when you flunk a few years thanks to undiagnosed issues. And yet, just like a LLM, i simply rely on tool usage to fix my inaccuracy issues. No point in wasting good shoulder space trying to graft a dozen more heads/brains onto me, when i can simply delegate the issue away. ;)
The fact that we can get computer models, that can almost program, write texts, ... and do so much more like a slightly malfunctioning human, amazes me. And at the same time, i curse at it like my teachers did, and also call it dumb at times hehehe ... I now understand how my teachers felt loool
That's confusing basic arithmetic as a user feature and as an implementation requirement.
I guarantee that computer vision and email clients both use basic arithmetic in implementation. And it would be trivially easy to bolt a calculator into an email app, because the languages used to write email apps include math features.
That's not true of LLMs. There's math at the bottom of the stack. But LLMs run as a separate closed and opaque application of a unique and self-contained type, which isn't easily extensible.
They don't include hooks into math features on the GPUs, and there's no easy way to add hooks.
If you want math, you need a separate tool call to conventional code.
IMO testing LLMs as if they "should" be able to do arithmetic is bizarre. They can't. They're not designed to. And even if they did, they'd be ridiculously inefficient at it.
> Pretty sure the only thing computer vision does is math.
That is only marginally less pedantic than saying that the only thing computer vision does is run discrete electrical signals through billions of transistors.
Yes, everything that a computer does, it does using math. This does not imply that things running on the computer can do basic arithmetic tasks for the user.
On some level this makes sense, but on the other hand LLMs already have perfect recall of thousands of symbols built into them, which is what pencil and paper gives to a human test taker.
If you're not doing clever hacks for very long windows, I thought a basic design fed in the entire window and it's up to the weights to use it properly.
Agreed. I don't like when the prompt sets up a good portion of how to go about finding the answer by saying which tools to use and how. The LLM needs to decide when and how to use them, not the prompt.
I don't think it should be completely open ended. I mean, you could have an "ask_hooman" tool that solves a ton of problems with current LLMs. But that doesn't mean the LLM is capable with respect to the benchmark.
Why not? One of the most intelligent things to do when stuck on a problem is to get outside help.
If allowing this behaviour raises a problem, you can always add constraints to the benchmark such as "final answer must come out under 15s" or something. The LLM can then make the decision to ask around in accordance to the time risk.
Because AI are good at devolving to the highest score, regardless of test intent. For most problems "ask_hooman", or especially the plural, would be much more effective. So, the degenerate case would dominate and tell you precisely zero about the intelligence of the AI. If a specific "tool" is more adept than the "AI" then "choose tool" will always be the correct answer. But I agree, a tight time constraint would help.
I'm not addressing an argument, just stating that's already a form of LLM testing done today for people wanting to look at the difference in results the same as the human analogy.
> To the contrary, reasoning ability should enable them to handle numbers of arbitrary size, just as it enables humans to do so, given some pencil and paper.
People interested can see the results of giving LLMs pen and paper today by looking at benchmarks with tools enabled. It's an addition to what you said, not an attack on a portion of your comment :).
I see now. My focus was on the effect of LLMs’ (and by analogy, humans’) reasoning abilities argued by bee_rider. The fact that tool use can enable more reliable handling of large numbers has no bearing on that, hence I found the reply confusing.
Hmm, maybe it depends on the specific test and reasoning in it? I certainly think reasoning how and when to use allowed tools and when not to is a big part of the reasoning and verification process E.g. most human math scores allow for a pen and paper calculation, or even a calculator, and that can be a great way to say spot check a symbolic derivative and see it needs to be revisited without relying on the calculator/paper to do the actual reasoning for the testee. Or to see the equation for motion of a system can't possibly have been right with some test values (without which I'm not sure I'd have passed my mid level physics course haha).
At the very least, the scores for benchmarking a human on such a test with and without tools would be different to comparing an LLM without the analogous constraints. Which is (IMO) a useful note in comparing reasoning abilities and why I thought it was interesting to note this kind of testing is just called testing with tools on the LLM side (not sure there is an equally as standard term on the human testing side? Guess the same could be used for both though).
At the same time I'm sure other reasoning tests don't gain much from/expect use of tools at all. So it wouldn't be relevant for those reasoning tests.
> Since performance on large numbers is not what these exams are intended to test for,
How so? Isn't the point of these exams to test arithmetic skills? I would hope we'd like arithmetic skills to be at a constant level regardless of the size of the number?
No. AIME is a test for advanced high schoolers that mostly tests higher level math concepts like algebra and combinatorics. The arithmetic required is basic. All the answers are 3-digit numbers so that judging is objective and automated while making guessing infeasible. You have 12 minutes on average for each question, so even if you are terribly slow at arithmetic, you should still be able to calculate the correct answer if you can perform all the other math.
That's probably a great test for high schoolers but it doesn't really test what we want from AI, no? I would expect AI to be limited by the far greater constraints of its computing ability, and not the working memory of a human high schooler.
IMO I think the calculator problem goes away with tool use or NN architectures that basically add a calculator equivalent as one of the potential 'experts' or similar. It won't be much of a trope for longer.
LLMs can probably be taught or configured to use external tools like Excel or Mathematica when such calculations are needed. Just like humans. There are plenty of untapped optimization opportunities.
>the point of these LLMs is to do things that computers were bad at.
The way they’re being deployed it feels like the point of LLMs is largely to replace basic online search or to run your online customer support cheaply.
I’m a bit out on a limb here because this is not really my technical expertise by any stretch of the imagination, but it seems to me these benchmark tests don’t really tell us much about how LLM’s perform in the ways most people actually use them. Maybe I’m off base here though
Nobody really knows "the point" of LLMs yet. They weren't even "invented" as much as they emerged as a trick to get computers to better understand human language.
They're still brand spanking new and everyone's trying to figure out how to best use them. We don't even really know if they're ever going to be "really good at" any given task!
Are they "really good at" these things or are they merely "OK-ish"?
* Answering factual questions.
* Programming.
* Understanding what the user wants from natural language.
* Searching/recommending stuff.
Real world testing suggests that with billions and billions of dollars spent, you really can get an LLM to be "OK-ish" at all those things :D
I don't claim to know anything but I thought tool usage was a major sign of intelligence. For example floats are a wonderful technology but people use them as if chainsaws are great for cutting bread and butter. We now have entire languages that cant do basic arithmetic. I thought it was alarming: People it cant compute like this! Now we have language models, those are still computers, why cant we just give them.. you know... calculators? Arguably the best thing their universe has to offer.
edit: I forgot my point: calculating big numbers is not a real world problem anyone has.
This is solvable at the level of an individual developer. Write your own benchmark for code problems that you've solved. Verify tests pass and that it satisfies your metrics like tok/s and TTFT. Create a harness that works with API keys or local models (if you're going that route).
At the developer level all my LLM use is in the context of agentic wrappers, so my benchmark is fairly trivial:
Configure aider or claude code to use the new model, try to do some work. The benchmark is pass/fail, if after a little while I feel the performance is better than the last model I was using it's a pass, otherwise it's a fail and I go back.
Building your own evaluations makes sense if you're serving an LLM up to customers and want to know how it performs, but if you are the user... use it and see how it goes. It's all subjective anyway.
> Building your own evaluations makes sense if you're serving an LLM up to customers and want to know how it performs, but if you are the user... use it and see how it goes. It's all subjective anyway.
I'd really caution against this approach, mainly because humans suck at removing emotions and other "human" factors when judging how well something works, but also because comparing across models gets a lot easier when you can see 77/100 vs 91/100 as a percentage score, over your own tasks that you actually use the LLMs for. Just don't share this benchmark publicly once you're using it for measurements.
So what? I'm the one that's using it, I happen to be a human, my human factor is the only one that matters.
At this point anyone using these LLMs every day have seen those benchmark numbers go up without an appreciable improvement in the day to day experience.
> So what? I'm the one that's using it, I happen to be a human, my human factor is the only one that matters.
Yeah no you're right, if consistency isn't important to you as a human, then it doesn't matter. Personally, I don't trust my "humanness" and correctness is the most important thing for me when working with LLMs, so that's why my benchmarks focus on.
> At this point anyone using these LLMs every day have seen those benchmark numbers go up without an appreciable improvement in the day to day experience.
Yes, this is exactly my point. The benchmarks the makers of these LLMs seems to always provide a better and better score, yet the top scores in my own benchmarks have been more or less the same for the last 1.5 years, and I'm trying every LLM I can come across. These "the best LLM to date!" hardly ever actually is the "best available LLM", and while you could make that judgement by just playing around with LLMs, actually be able to point to specifically why that is, is something at least I find useful, YMMV.
Benchmarks are nothing more than highly contextual specs (in traditional code). They demonstrate your code works in a certain way in certain use cases, but they do not prove your code works as expected in all use cases.
yes it does - it has to be meaningful or rigorous for the comparative ranking to be meaningful or rigorous, or else wtf are you doing? Say I have all the information on my side but only these questions that you are showing the user? Who cares about that comparison?
For statistical AI models, we can use out of sample prediction error as an objective measure to compare models. What makes evaluating LLMs difficult is that comparisons are inextricable from utility (whereas statistical AI models do have a pre-utility step wherein it can be shown out of sample prediction epsilon is minimized).
I'd like to see some video generation benchmarks. For example, one that tested a model's ability to generate POV footage of a humanoid form carrying out typical household tasks
Even if it requires human evaluators at first, and even if the models completely suck at this task right now: it seems like the kind of task you'd want them to be good at, if you want these models to eventually carry out these tasks in embodied forms in the real world.
Just having the benchmark in the first place is what gives model makers something to optimize for.
Generating footage wouldn't help with the opposite but navigating a simulation would which is a pretty standard type of evaluation for multimodal AIs designed to act in the real world.
Do you mean that it wouldn't help with ingesting footage and then determining how to act?
I can imagine a robotics architecture where you have one model generating footage (next frames for what it is currently seeing) and another dumber model which takes in the generated footage and only knows how to generate the motor/servo control outputs needed to control whatever robot platform it is integrated with.
I think that kind of architecture decoupling would be nice. It allows the model with all the world and task-specific knowledge to be agnostic from its underlying robot platform.
I wish the big providers would offer some sort of trial period where you can evaluate models in a _realistic_ setting yourself (i.e cli tools or IDE integrations). I wouldn't even mind strict limits -- just give me two hours or so of usage and I'd already be happy. Seriously.
My use-case is probably pretty far from the usual tasks: I'm currently implementing a full observability platform based on VictoriaMetrics / Victorialogs + Grafana. It's quite elaborate and has practically no overlap with the usual/cloud solutions you find out there. For example, it uses an authenticated query stack: I use the Grafana oauth token to authenticate queries by injecting matchers via prom-label-proxy and forward that to promxy for fan-out to different datasources (using the label filter to only query some datasources). The IaC stuff is also not mainstream as I'm not using any of the big cloud providers, but the provider I use nonetheless has a terraform provider.
As you can imagine, there's probably not much training data for most of this, so quality of the responses varies widely. From my experience so far Claude (Sonnet 4.5 ) does a _much_ better job than GTP-5 (Codex or normal) with the day-to-day task. Stuff like keeping documentation up to date, spotting inconsistencies, helping me find blind spots in the Alerting rules, etc. It also seems to do better working with provided documentation / links.
I've been using Claude for a couple of weeks now but recently switched to codex after my subscription to Claude ran out. I was really curious after reading a lot of good things about it but I gotta say, so far, I'm not impressed. Compared to Claude it gives wrong answers much more frequently (at least in this domain). The results it produces take much more effort to clean up than Claude's. Probably on a level where I could just invest the time myself. Might be that I do not yet know how to correctly prompt GPT but giving both tools the same prompt, Claude does a better job 90% of the time.
Anyway, I guess this is my long-winded way of saying that the quality of responses "off the beaten track" varies widely and is worth testing several models with. Especially if your work is not 70+% of coding. Even then I guess that many benchmarks have seized being useful by now?
There's the github copilot 30 day trial? "Access to Anthropic Claude Sonnet 4, GPT-5, Gemini 2.5 Pro, and more
300 premium requests to use the latest models and code review"
Benchmarks optimize for fundraising, not users. The gap between "state of the art" and "previous gen" keeps shrinking in real-world use, but investors still write checks based on decimal points in test scores.
we try to make benchmarks for users, but it's like that 20% article - different people want different 20% and you just end up adding "features" and whackamoling the different kinds of 20%
if a single benchmark could be a universal truth, and it was easy to figure out how to do it, everyone would love that.. but that's why we're in the state we're in right now
The problem isn’t with the benchmarks (or the models, for that matter) it’s their being used to prop up the indefensible product marketing claims made by people frantically justifying asking for more dump trucks of thousand-dollar bills to replace the ones they just burned through in a few months.
Definitely one of the weaker areas in the current LLM boom. Comparing models, or even different versions of the same model, is a pseudo-scientific mess.
I'm still using https://lmarena.ai/leaderboard. Perhaps there is something better and someone will pipe up to tell me about it. But we use LLMs at work and have unexplainable variations between them.
And when we get a prompt working reliably on one model, we often have trouble porting it to another LLM - even straight "version upgrades" such as from GPT-4 to -5. Your prompt and your model become highly coupled quite easily.
I dunno what to do about it and am tending to just pick Gemini as a result.
I work on LLM benchmarks and human evals for a living in a research lab (as opposed to product). I can say: it’s pretty much the Wild West and a total disaster. No one really has a good solution, and researchers are also in a huge rush and don’t want to end up making their whole job benchmarking. Even if you could, and even if you have the right background you can do benchmarks full time and they still would be a mess.
Product testing (with traditional A/B tests) are kind of the best bet since you can measure what you care about _directly_ and at scale.
I would say there is of course “benchmarketing” but generally people do sincerely want to make good benchmarks it’s just hard or impossible. For many of these problems we’re hitting capabilities where we don’t even have a decent paradigm to use,
For what it's worth, I work on platforms infra at a hyperscaler and benchmarks are a complete fucking joke in my field too lol.
Ultimately we are measuring extremely measurable things that have an objective ground truth. And yet:
- we completely fail at statistics (the MAJORITY of analysis is literally just "here's the delta in the mean of these two samples". If I ever do see people gesturing at actual proper analysis, if prompted they'll always admit "yeah, well, we do come up with a p-value or a confidence interval, but we're pretty sure the way we calculate it is bullshit")
- the benchmarks are almost never predictive of the performance of real world workloads anyway
- we can obviously always just experiment in prod but then the noise levels are so high that you can entirely miss million-dollar losses. And by the time you get prod data you've already invested at best several engineer-weeks of effort.
AND this is a field where the economic incentives for accurate predictions are enormous.
In AI, you are measuring weird and fuzzy stuff, and you kinda have an incentive to just measure some noise that looks good for your stock price anyway. AND then there's contamination.
Looking at it this way, it would be very surprising if the world of LLM benchmarks was anything but a complete and utter shitshow!
I have actually been thinking of hiring some training contractors to come in and teach people the basics of applied statistical inference. I think with a bit of internal selling, engineers would generally be interested enough to show up and pay attention. And I don't think we need very deep expertise, just a moderate bump in the ambient level of statistical awareness would probably go a long way.
It's not like there's a shortage of skills in this area, it seems like our one specific industry just has a weird blindspot.
A/B testing is radioactive too. It's indirectly optimizing for user feedback - less stupid than directly optimizing for user feedback, but still quite dangerous.
Human raters are exploitable, and you never know whether the B has a genuine performance advantage over A, or just found a meat exploit by an accident.
It's what fucked OpenAI over with 4o, and fucked over many other labs in more subtle ways.
Are you talking about just preferences or A/B tests on like retention and engagement? The latter I think is pretty reliable and powerful though I have never personally done them. Preferences are just as big a mess: WHO the annotators are matters, and if you are using preferences as a proxy for like correctness, you’re not really measuring correctness you’re measuring e.g. persuasion. A lot of construct validity challenges (which themselves are hard to even measure in domain).
Yes. All of them are poisoned metrics, just in different ways.
GPT-4o's endless sycophancy was great for retention, GPT-5's style of ending every response in a question is great for engagement.
Are those desirable traits though? Doubt it. They look like simple tricks and reek of reward hacking - and A/B testing rewards them indeed. Direct optimization is even worse. Combining the two is ruinous.
Mind, I'm not saying that those metrics are useless. Radioactive materials aren't useless. You just got to keep their unpleasant properties in mind at all times - or suffer the consequences.
Has your lab tried using any of the newer causal inference–style evaluation methods? Things like interventional or counterfactual benchmarking, or causal graphs to tease apart real reasoning gains from data or scale effects. Wondering if that’s something you’ve looked into yet, or if it’s still too experimental for practical benchmarking work.
Even professional human evaluators are quite vulnerable to sycophancy and overconfident-and-wrong answers. And LMArena evaluators aren't professionals.
A lot of the sycophancy mess that seeps from this generation of LLM stems from reckless tuning based on human feedback. Tuning for good LMArena performance has similar effects - and not at all by a coincidence.
It's biased to small context performance, which is why I don't pay much attention to it as a developer aside from a quick glance. I need performance at 40-100k tokens which models like Deepseek can't deliver but Gemini 2.5 Pro and ChatGPT 5.0 Thinking can.
And even "long term performance" splits itself into "performance on multi-turn instruction following" and "performance on agentic tasks" down the line. And "performance on agentic tasks" is a hydra in itself.
Capturing LLM performance with a single metric is a hopeless task. But even a single flawed metric beats no metrics at all.
This is something I've stuggled with for my site, I made https://aimodelreview.com/ to compare the outputs of LLMs over a variety of prompts and categories, allowing a side by side comparison between them. I ran each prompt 4 times for each model with different temperature values available as a toggles.
My thinking was to just make the responses available to users and let them see how models perform. But from some feedback, turns out users don't want to have to evaluate the answers and would rather see a leaderboard and rankings.
The scalable solution to that would be LLM as judge that some benchmarks already use, but that just feels wrong to me.
LM Arena tries to solve this with the crowd sourced solution, but I think the right method would have to be domain expert human reviewers, so like Wirecutter VS IMDb, but that is expensive to pull off.
Tech companies/bloggers/press/etc are perpetually bad at benchmarks. For browsers they kept pushing simplistic javascript-centric benchmarks even when it was clear for at least 15 years that layout/paint/network/etc were the dominant bottlenecks in real-world usage.
It's primarily marketing-driven. I think the technical parts of companies need to attempt to own this more.
It gets really weird when engineering priorities shift because of these mostly irrelevant benchmarks.
The problem with the LLM benchmarks is that if you see one that shows high performance by something that isn’t from Anthropic, Google or OpenAI, you don’t believe it, even if it were “true.” In that sense, benchmarks are a holistic social experience in this domain, less a scientific endeavour.
When people claim that there is such a thing as "X% accuracy in reasoning", it's really hard to take anything else seriously, no matter how impressive.
AI (and humans!) aside, claiming that there was an oracle that could "answer all questions" is a solved problem. Such a thing cannot exist.
But this is going already too deep IMO.
When people start talking about percentages or benchmark scores, there has to be some denominator.
And there can be no bias-free such denominator for
- trivia questions
- mathematical questions (oh, maybe I'm wrong here, intuitively I'd say it's impossible for various reasons: varying "hardness", undecidable problems etc)
- historical or policital questions
I wanted to include "software development tasks", but it would be a distraction. Maybe there will be a good benchmark for this, I'm aware there are plenty already. Maybe AI will be capable to be a better software developer than me in some capacity, so I don't want to include this part here. That also maps pretty well to "the better the problem description, the better the output", which doesn't seem to work so neatly with the other categories of tasks and questions.
Even if the whole body of questions/tasks/prompts would be very constrained and cover only a single domain, it seems impossible to guarantee that such benchmark is "bias-free" (I know AGI folks love this word).
Maybe in some interesting special cases? For example, very constrained and clearly defined classes of questions, at which point, the "language" part of LLMs seems to become less important and more of a distraction. Sure, AI is not just LLMs, and LLMs are not just assistants, and Neural Networks are not just LLMs...
There the problem begins to be honest: I don't even know how to align the "benchmark" claims with the kind of AI they are examinin and the ones I know exist.
Sure it's possible to benchmark how well an AI decides whether, for example, a picture shows a rabbit.
Even then: for some pictures, it's gotta be undecidable, no matter how good the training data is?
I'm just a complete layman and commenting about this; I'm not even fluent in the absolute basics of artificial neural networks like perceptrons, gradient descent, backpropagation and typical non-LLM CNNs that are used today, GANs etc.
I am and was impressed by AI and deep learning, but to this day I am thorougly disappointed by the hubris of snakeoil salespeople who think it's valuable and meaningful to "benchmark" machines on "general reasoning".
I mean, it's already a thing in humans. There are IQ tests for the non-trivia parts. And even these have plenty of discussion revolving around them, for good reason.
Is there some "AI benchmark" that exclusively focuses on doing recent IQ tests on models, preferably editions that were published after the particular knowledge cutoff of the respective models? I found (for example) this study [1], but to be honest, I'm not the kind of person who is able to get the core insights presented in such a paper by skimming through it.
Because I think there are impressive results, it's just becomimg very hard to see through the bullshit at as an average person.
I would also love to understand mroe about the current state of the research on the "LLMs as compression" topic [2][3].
I'm already quite put off by the title (it's science -- if you have a better benchmark, publish it!), but the contents aren't great either. It keeps citing numbers about "445 LLM benchmarks" without confirming whether any of the ones they deem insufficiently statistical are used by any of the major players. I've seen a lot of benchmarks, but maybe 20 are used regularly by large labs, max.
"For example, if a benchmark reuses questions from a calculator-free exam such as AIME," the study says, "numbers in each problem will have been chosen to facilitate basic arithmetic. Testing only on these problems would not predict performance on larger numbers, where LLMs struggle."
For a math-based critique, this seems to ignore a glaring problem: is it even possible to randomly sample all natural numbers? As another comment pointed out we wouldn't even want to ("LLMs can't accurately multiply 6-digit numbers" isn't something anyone cares about/expected them to do in the first place), but regardless: this seems like a vacuous critique dressed up in a costume of mathematical rigor.
At least some of those who design benchmark tests are aware of these concerns.
In related news, at least some scientists studying climate change are aware that their methods are imperfect. More at 11!
Not the person you replied to, but I'm a programmer and up until ~3 weeks ago I really only used AI for auto-complete, looking up API information, and constructing arcane CLI commands.
I decided to take a leap and use AI as much as possible to complete a ticket at work. Now, 3 weeks later AI is writing 90% of my code.
Granted, I'm not sitting back sipping on a latte while AI does my job. It's a very interactive process, I spend more time reviewing code and going back and forth with the AI to get the result I want. But it's become surprisingly good.
I wouldn't say I'm anywhere close to 10x more productive, but perhaps 50%.
It definitely has its amazing moments, but sometimes I get caught in a loop of expecting it to do the thing, it not working, and spinning my wheels a lot instead of just solving it myself. I think I’m still learning how to use the tools effectively, but the random nature of it makes it difficult.
I've been getting flagged by high-on-their-own-supply AI boosters for identifying that LLM benchmarks have been obvious bullshit for at least the last year and a half.
What changed to make "the inevitable AI bubble" the dominant narrative in last week or so?
It is possible to be right on the main theme but only by accident (with arguments and claims being wrong), communicating in a highly faulty way, with pointless insults, doing it in offtopic derails, being correct on minor point while being mostly wrong etc.
Can you link some of these comments you consider useful but got flagged?
Link those comments please because I checked your history and the flagged ones were pure nonsense with zero insights. Also, calling out LLM benchmarks has never been a radical take and basically the default on this site.
The market was down for AI related stocks especially, while down only over 3% it’s the worst week since April, and there’s no single event that is to blame it just looks like market sentiment has shifted away from the previous unchecked exuberance.
https://www.happiesthealth.com/articles/future-of-health/hum...
It's a shifting goalpost, but one of the things that struck me was how some questions could still be trivial for a fairly qualified human (a doctor in this case) but difficult for an AI model. Reasoning, visual or logic, is built on a set of assumptions that are better gained through IRL experience than crawling datasets and matching answers.
This leads me to believe that much of the future for training AI models will lie in exposing them to "meatspace" and annotating their inferences, much like how we train a child. This is a long, long process, and one that is already underway at scale. But it's what might give us emergent intelligences rather than just a basket of competing yet somehow-magic thesaurus.
For example a test of “multiply 1765x9392” would have some correlation with human intelligence but it wouldn’t make sense to apply it to computers.
Someone want to start? I've got a Yjs/CRDT collaborative editing bug that took like a week and a half of attempts with Claude Code (Sonnet 4.5), GPT5-codex (medium), and GLM-4.6 many, many attempts to figure out. Even then they didn't really get it... Just came up with a successful workaround (which is good enough for me but still...).
Aside: You know what really moved the progress bar on finding and fixing the bug? When I had a moment of inspiration and made the frontend send all it's logs to the backend so the AIs could see what was actually happening on the frontend (near real-time). Really, I was just getting sick of manual testing and pasting the console output into the chat (LOL). Laziness FTW!
I have the Google Chrome Dev Tools MCP but for some reason it doesn't work as well :shrug:
> we (expert developers) ...
> took like a week and a half of attempts with Claude Code ...
What kind of expert developer wastes that much time prompting a bunch of different LLMs to end up with a workaround, instead of actually debugging and fixing the bug themselves?
I think any LLM-user worth their salt have been doing this pretty much since we got API access to LLMs, as otherwise there is no way to actually see if they can solve the things you care about.
The only difference is that you must keep the actual benchmarks to yourself, don't share them with anyone and even less put them publicly. The second you do, you probably should stop using it as an actual benchmark, as newly trained LLMs will either intentionally or unintentionally slurp up your benchmark and suddenly it's no longer a good indicator.
I think I personally started keeping my own test cases for benchmarking around the GPT3 launch, when it became clear the web will be effectively "poisoned" from that part on, and anything on the public internet can be slurped up by the people feeding the LLMs training data.
Once you have this up and running, you'll get a much more measured view of how well new LLMs work, and you'll quickly see that a lot of the fanfare doesn't actually hold up when testing it against your own private benchmarks. On a happier note, you'll also be surprised when a model suddenly does a lot better in a specific area that wasn't even mentioned at release, and then you could switch to it for specifically that task :)
Probably because Sonnet is no longer a frontier model, it isn't even the best model Anthropic offers, according to themselves.
It's more token efficient too since I don't need to load the full MCP description into my context.
Think of it as TDD.
When models figure out how to exploit an effect that every clever college student does, that should count as a win. That’s a much more human-like reasoning ability, than the ability to multiply large numbers or whatever (computers were already good at that, to the point that it has become a useless skill for humans to have). The point of these LLMs is to do things that computers were bad at.
College exam takers use those tricks because they are on a time limit and are gaming the system. It's clever and wink wink nudge nudge ok everyone does it. But it's one tiny signal in a huge spectrum of things we use to evaluate people.
Instead, these metrics are gamed and presented as the entire multi special signal of competence for LLMs because it is literally impossible to say that success in one domain would translate the way it might with a good hire.
What I want is something I don't have to guard against gaming. Something conscientious and capable like my co workers. Until then it's google version 2 married to intellisense and I'm not letting do anything by itself.
However:
> Testing only on these problems would not predict performance on larger numbers, where LLMs struggle.
Since performance on large numbers is not what these exams are intended to test for, I don’t see this as a counterargument, unless the benchmarks are misrepresenting what is being tested for.
Or given a calculator. Which it's running on. Which it in some sense is. There's something deeply ironic about the fact that we have an "AI" running on the most technologically advanced calculator in the history of mankind and...it can't do basic math.
I think the tokenization is a bigger problem than the model itself.
So while it can look like a LLM calculates correctly, its still restricted by this accuracy issue. What happens when you get a single number wrong in a calculation, everything is wrong.
While a calculator does not deal with predictions but basic adding/multiplying/subtracting etc .. Things that are 100% accurate (if we not not count issues like cosmic rays hitting, failures in silica etc).
A trillion parameter model is just that, a trillion parameters, but what matter is not the tokens but the accuracy as in, the do they use int, float16, float32, float64 ... The issue is, the higher we go, the memory usage explodes.
There is no point in spending terabytes of memory, to just get a somewhat accurate predictive calculator, when we can just have the LLM call a actual calculator, to ensure its results are accurate.
Think of a LLM more like somebody with Dyslexia / Dyscalculia... It does not matter how good you are, all it takes is to switch one number in a algebraic calculation to get a 0/10 ... The reason why i mention this, is because i often think of a LLM like a person with Dyslexia / Dyscalculia. It can have insane knowledge, be smart, but be considered dumb by society because of that less then accurate prediction (or number swiping issue).
Take it from somebody that wasted a few years in school thanks to that issue, it really does not matter if your a good programmer later in life, when you flunk a few years thanks to undiagnosed issues. And yet, just like a LLM, i simply rely on tool usage to fix my inaccuracy issues. No point in wasting good shoulder space trying to graft a dozen more heads/brains onto me, when i can simply delegate the issue away. ;)
The fact that we can get computer models, that can almost program, write texts, ... and do so much more like a slightly malfunctioning human, amazes me. And at the same time, i curse at it like my teachers did, and also call it dumb at times hehehe ... I now understand how my teachers felt loool
My email client runs on my computer and it doesn’t do basic arithmetic either.
Something running on a computer does not imply that it can or should do basic arithmetic
I guarantee that computer vision and email clients both use basic arithmetic in implementation. And it would be trivially easy to bolt a calculator into an email app, because the languages used to write email apps include math features.
That's not true of LLMs. There's math at the bottom of the stack. But LLMs run as a separate closed and opaque application of a unique and self-contained type, which isn't easily extensible.
They don't include hooks into math features on the GPUs, and there's no easy way to add hooks.
If you want math, you need a separate tool call to conventional code.
IMO testing LLMs as if they "should" be able to do arithmetic is bizarre. They can't. They're not designed to. And even if they did, they'd be ridiculously inefficient at it.
I’ve also observed email clients tallying the number of unread emails I have. It’s quite obnoxious actually, but I qualify adding as math.
That is only marginally less pedantic than saying that the only thing computer vision does is run discrete electrical signals through billions of transistors.
If allowing this behaviour raises a problem, you can always add constraints to the benchmark such as "final answer must come out under 15s" or something. The LLM can then make the decision to ask around in accordance to the time risk.
Personally, I’d say that such tool use is more akin to a human using a calculator.
People interested can see the results of giving LLMs pen and paper today by looking at benchmarks with tools enabled. It's an addition to what you said, not an attack on a portion of your comment :).
At the very least, the scores for benchmarking a human on such a test with and without tools would be different to comparing an LLM without the analogous constraints. Which is (IMO) a useful note in comparing reasoning abilities and why I thought it was interesting to note this kind of testing is just called testing with tools on the LLM side (not sure there is an equally as standard term on the human testing side? Guess the same could be used for both though).
At the same time I'm sure other reasoning tests don't gain much from/expect use of tools at all. So it wouldn't be relevant for those reasoning tests.
How so? Isn't the point of these exams to test arithmetic skills? I would hope we'd like arithmetic skills to be at a constant level regardless of the size of the number?
The way they’re being deployed it feels like the point of LLMs is largely to replace basic online search or to run your online customer support cheaply.
I’m a bit out on a limb here because this is not really my technical expertise by any stretch of the imagination, but it seems to me these benchmark tests don’t really tell us much about how LLM’s perform in the ways most people actually use them. Maybe I’m off base here though
They're still brand spanking new and everyone's trying to figure out how to best use them. We don't even really know if they're ever going to be "really good at" any given task!
Are they "really good at" these things or are they merely "OK-ish"?
Real world testing suggests that with billions and billions of dollars spent, you really can get an LLM to be "OK-ish" at all those things :Dedit: I forgot my point: calculating big numbers is not a real world problem anyone has.
Configure aider or claude code to use the new model, try to do some work. The benchmark is pass/fail, if after a little while I feel the performance is better than the last model I was using it's a pass, otherwise it's a fail and I go back.
Building your own evaluations makes sense if you're serving an LLM up to customers and want to know how it performs, but if you are the user... use it and see how it goes. It's all subjective anyway.
I'd really caution against this approach, mainly because humans suck at removing emotions and other "human" factors when judging how well something works, but also because comparing across models gets a lot easier when you can see 77/100 vs 91/100 as a percentage score, over your own tasks that you actually use the LLMs for. Just don't share this benchmark publicly once you're using it for measurements.
At this point anyone using these LLMs every day have seen those benchmark numbers go up without an appreciable improvement in the day to day experience.
Yeah no you're right, if consistency isn't important to you as a human, then it doesn't matter. Personally, I don't trust my "humanness" and correctness is the most important thing for me when working with LLMs, so that's why my benchmarks focus on.
> At this point anyone using these LLMs every day have seen those benchmark numbers go up without an appreciable improvement in the day to day experience.
Yes, this is exactly my point. The benchmarks the makers of these LLMs seems to always provide a better and better score, yet the top scores in my own benchmarks have been more or less the same for the last 1.5 years, and I'm trying every LLM I can come across. These "the best LLM to date!" hardly ever actually is the "best available LLM", and while you could make that judgement by just playing around with LLMs, actually be able to point to specifically why that is, is something at least I find useful, YMMV.
Maybe we need something similar for benchmarks, and updated for today's LLMs, like:
> LLM benchmarks can be used to show what tasks they can do, but never to show what tasks they cannot.
Even if it requires human evaluators at first, and even if the models completely suck at this task right now: it seems like the kind of task you'd want them to be good at, if you want these models to eventually carry out these tasks in embodied forms in the real world.
Just having the benchmark in the first place is what gives model makers something to optimize for.
I can imagine a robotics architecture where you have one model generating footage (next frames for what it is currently seeing) and another dumber model which takes in the generated footage and only knows how to generate the motor/servo control outputs needed to control whatever robot platform it is integrated with.
I think that kind of architecture decoupling would be nice. It allows the model with all the world and task-specific knowledge to be agnostic from its underlying robot platform.
My use-case is probably pretty far from the usual tasks: I'm currently implementing a full observability platform based on VictoriaMetrics / Victorialogs + Grafana. It's quite elaborate and has practically no overlap with the usual/cloud solutions you find out there. For example, it uses an authenticated query stack: I use the Grafana oauth token to authenticate queries by injecting matchers via prom-label-proxy and forward that to promxy for fan-out to different datasources (using the label filter to only query some datasources). The IaC stuff is also not mainstream as I'm not using any of the big cloud providers, but the provider I use nonetheless has a terraform provider.
As you can imagine, there's probably not much training data for most of this, so quality of the responses varies widely. From my experience so far Claude (Sonnet 4.5 ) does a _much_ better job than GTP-5 (Codex or normal) with the day-to-day task. Stuff like keeping documentation up to date, spotting inconsistencies, helping me find blind spots in the Alerting rules, etc. It also seems to do better working with provided documentation / links.
I've been using Claude for a couple of weeks now but recently switched to codex after my subscription to Claude ran out. I was really curious after reading a lot of good things about it but I gotta say, so far, I'm not impressed. Compared to Claude it gives wrong answers much more frequently (at least in this domain). The results it produces take much more effort to clean up than Claude's. Probably on a level where I could just invest the time myself. Might be that I do not yet know how to correctly prompt GPT but giving both tools the same prompt, Claude does a better job 90% of the time.
Anyway, I guess this is my long-winded way of saying that the quality of responses "off the beaten track" varies widely and is worth testing several models with. Especially if your work is not 70+% of coding. Even then I guess that many benchmarks have seized being useful by now?
You can get a lot of free usage out of the models.
if a single benchmark could be a universal truth, and it was easy to figure out how to do it, everyone would love that.. but that's why we're in the state we're in right now
I'm still using https://lmarena.ai/leaderboard. Perhaps there is something better and someone will pipe up to tell me about it. But we use LLMs at work and have unexplainable variations between them.
And when we get a prompt working reliably on one model, we often have trouble porting it to another LLM - even straight "version upgrades" such as from GPT-4 to -5. Your prompt and your model become highly coupled quite easily.
I dunno what to do about it and am tending to just pick Gemini as a result.
Product testing (with traditional A/B tests) are kind of the best bet since you can measure what you care about _directly_ and at scale.
I would say there is of course “benchmarketing” but generally people do sincerely want to make good benchmarks it’s just hard or impossible. For many of these problems we’re hitting capabilities where we don’t even have a decent paradigm to use,
Ultimately we are measuring extremely measurable things that have an objective ground truth. And yet:
- we completely fail at statistics (the MAJORITY of analysis is literally just "here's the delta in the mean of these two samples". If I ever do see people gesturing at actual proper analysis, if prompted they'll always admit "yeah, well, we do come up with a p-value or a confidence interval, but we're pretty sure the way we calculate it is bullshit")
- the benchmarks are almost never predictive of the performance of real world workloads anyway
- we can obviously always just experiment in prod but then the noise levels are so high that you can entirely miss million-dollar losses. And by the time you get prod data you've already invested at best several engineer-weeks of effort.
AND this is a field where the economic incentives for accurate predictions are enormous.
In AI, you are measuring weird and fuzzy stuff, and you kinda have an incentive to just measure some noise that looks good for your stock price anyway. AND then there's contamination.
Looking at it this way, it would be very surprising if the world of LLM benchmarks was anything but a complete and utter shitshow!
It's not like there's a shortage of skills in this area, it seems like our one specific industry just has a weird blindspot.
Human raters are exploitable, and you never know whether the B has a genuine performance advantage over A, or just found a meat exploit by an accident.
It's what fucked OpenAI over with 4o, and fucked over many other labs in more subtle ways.
GPT-4o's endless sycophancy was great for retention, GPT-5's style of ending every response in a question is great for engagement.
Are those desirable traits though? Doubt it. They look like simple tricks and reek of reward hacking - and A/B testing rewards them indeed. Direct optimization is even worse. Combining the two is ruinous.
Mind, I'm not saying that those metrics are useless. Radioactive materials aren't useless. You just got to keep their unpleasant properties in mind at all times - or suffer the consequences.
Even professional human evaluators are quite vulnerable to sycophancy and overconfident-and-wrong answers. And LMArena evaluators aren't professionals.
A lot of the sycophancy mess that seeps from this generation of LLM stems from reckless tuning based on human feedback. Tuning for good LMArena performance has similar effects - and not at all by a coincidence.
Capturing LLM performance with a single metric is a hopeless task. But even a single flawed metric beats no metrics at all.
My thinking was to just make the responses available to users and let them see how models perform. But from some feedback, turns out users don't want to have to evaluate the answers and would rather see a leaderboard and rankings.
The scalable solution to that would be LLM as judge that some benchmarks already use, but that just feels wrong to me.
LM Arena tries to solve this with the crowd sourced solution, but I think the right method would have to be domain expert human reviewers, so like Wirecutter VS IMDb, but that is expensive to pull off.
It's primarily marketing-driven. I think the technical parts of companies need to attempt to own this more.
It gets really weird when engineering priorities shift because of these mostly irrelevant benchmarks.
AI (and humans!) aside, claiming that there was an oracle that could "answer all questions" is a solved problem. Such a thing cannot exist.
But this is going already too deep IMO.
When people start talking about percentages or benchmark scores, there has to be some denominator.
And there can be no bias-free such denominator for
- trivia questions
- mathematical questions (oh, maybe I'm wrong here, intuitively I'd say it's impossible for various reasons: varying "hardness", undecidable problems etc)
- historical or policital questions
I wanted to include "software development tasks", but it would be a distraction. Maybe there will be a good benchmark for this, I'm aware there are plenty already. Maybe AI will be capable to be a better software developer than me in some capacity, so I don't want to include this part here. That also maps pretty well to "the better the problem description, the better the output", which doesn't seem to work so neatly with the other categories of tasks and questions.
Even if the whole body of questions/tasks/prompts would be very constrained and cover only a single domain, it seems impossible to guarantee that such benchmark is "bias-free" (I know AGI folks love this word).
Maybe in some interesting special cases? For example, very constrained and clearly defined classes of questions, at which point, the "language" part of LLMs seems to become less important and more of a distraction. Sure, AI is not just LLMs, and LLMs are not just assistants, and Neural Networks are not just LLMs...
There the problem begins to be honest: I don't even know how to align the "benchmark" claims with the kind of AI they are examinin and the ones I know exist.
Sure it's possible to benchmark how well an AI decides whether, for example, a picture shows a rabbit. Even then: for some pictures, it's gotta be undecidable, no matter how good the training data is?
I'm just a complete layman and commenting about this; I'm not even fluent in the absolute basics of artificial neural networks like perceptrons, gradient descent, backpropagation and typical non-LLM CNNs that are used today, GANs etc.
I am and was impressed by AI and deep learning, but to this day I am thorougly disappointed by the hubris of snakeoil salespeople who think it's valuable and meaningful to "benchmark" machines on "general reasoning".
I mean, it's already a thing in humans. There are IQ tests for the non-trivia parts. And even these have plenty of discussion revolving around them, for good reason.
Is there some "AI benchmark" that exclusively focuses on doing recent IQ tests on models, preferably editions that were published after the particular knowledge cutoff of the respective models? I found (for example) this study [1], but to be honest, I'm not the kind of person who is able to get the core insights presented in such a paper by skimming through it.
Because I think there are impressive results, it's just becomimg very hard to see through the bullshit at as an average person.
I would also love to understand mroe about the current state of the research on the "LLMs as compression" topic [2][3].
[1] https://arxiv.org/pdf/2507.20208
[2] https://www.mattmahoney.net/dc/text.html
[3] https://arxiv.org/abs/2410.21352
If anyone doubts my concerns and thinks this article is in good faith, just check out this site's "AI+ML" section: https://www.theregister.com/software/ai_ml/
https://openreview.net/pdf?id=mdA5lVvNcU
And the review is pretty damning regarding statistical validity of LLM benchmarks.
I decided to take a leap and use AI as much as possible to complete a ticket at work. Now, 3 weeks later AI is writing 90% of my code.
Granted, I'm not sitting back sipping on a latte while AI does my job. It's a very interactive process, I spend more time reviewing code and going back and forth with the AI to get the result I want. But it's become surprisingly good.
I wouldn't say I'm anywhere close to 10x more productive, but perhaps 50%.
What changed to make "the inevitable AI bubble" the dominant narrative in last week or so?
Can you link some of these comments you consider useful but got flagged?
Nothing says confidence that AGI is imminent like needing the US government to prevent your investments from losing you money.