That must have been a long time back. Having lived through the time when web pages were served through CGI and mobile phones only existed in movies, when SVMs where the new hotness in ML and people would write about how weird NNs were, I feel like I've seen a lot more concrete progress in the last few decades than this year.
This year honestly feels quite stagnant. LLMs are literally technology that can only reproduce the past. They're cool, but they were way cooler 4 years ago. We've taken big ideas like "agents" and "reinforcement learning" and basically stripped them of all meaning in order to claim progress.
I mean, do you remember Geoffrey Hinton's RBM talk at Google in 2010? [0] That was absolutely insane for anyone keeping up with that field. By the mid-twenty teens RBMs were already outdated. I remember when everyone was implementing flavors of RNNs and LSTMs. Karpathy's character 2015 RNN project was insane [1].
This comment makes me wonder if part of the hype around LLMs is just that a lot of software people simply weren't paying attention to the absolutely mind-blowing progress we've seen in this field for the last 20 years. But even ignoring ML, the world's of web development and mobile application development have gone through incredible progress over the last decade and a half. I remember a time when JavaScript books would have a section warning that you should never use JS for anything critical to the application. Then there's the work in theorem provers over the last decade... If you remember when syntactic sugar was progress, either you remember way further back than I do, or you weren't paying attention to what was happening in the larger computing world.
> LLMs are literally technology that can only reproduce the past.
That's incorrect on many levels. They are drawing upon, and reproducing, language patterns from "the past", but they are combining those patterns in ways that may have never have been seen before. They may not be truly creative, but they are still capable of generating novel outputs.
> They're cool, but they were way cooler 4 years ago.
Maybe this year has been more about incremental progress with LLMs than the shock/coolness factor of talking to an LLM for the first time, but the utility of them, especially for programming, has dramatically increased this year, really in the last 6 months.
The improvement in "AI" image and video generation has also been impressive, to the point now that fake videos on YouTube can often only be identified as such by common sense rather that the fact that they don't look real.
Incremental improvement can often be more impressive that innovation, whose future importance can be hard to judge when it first appears. How many people read "Attention is all you need" in 2017 and thought "Wow! This is going to change the world!". Not even the authors of the paper thought that.
> LLMs are literally technology that can only reproduce the past.
Funny, I've used them to create my own personalized text editor, perfectly tailored to what I actually want. I'm pretty sure that didn't exist before.
It's wild to me how many people who talk about LLM apparently haven't learned how to use them for even very basic tasks like this! No wonder you think they're not that powerful, if you don't even know basic stuff like this. You really owe it to yourself to try them out.
I've worked at multiple AI startups in lead AI Engineering roles, both working on deploying user facing LLM products and working on the research end of LLMs. I've done collaborative projects and demos with a pretty wide range of big names in this space (but don't want to doxx myself too aggressively), have had my LLM work cited on HN multiple times, have LLM based github projects with hundreds of stars, appeared on a few podcasts talking about AI etc.
This gets to the point I was making. I'm starting to realize that part of the disconnect between my opinions on the state of the field and others is that many people haven't really been paying much attention.
I can see if recent LLMs are your first intro to the state of the field, it must feel incredible.
That's all very impressive, to be sure. But are you sure you're getting the point? As of 2025, LLMs are now very good at writing new code, creating new imagery, and writing original text. They continue to improve at a remarkable rate. They are helping their users create things that didn't exist before. Additionally, they are now very good at searching and utilizing web resources that didn't exist at training time.
So it is absurdly incorrect to say "they can only reproduce the past." Only someone who hasn't been paying attention (as you put it) would say such a thing.
> So it is absurdly incorrect to say "they can only reproduce the past."
Also , a shitton of what we do economically is reproducing the past with slight tweaks and improvements. We all do very repetitive things and these tools cut the time / personnel needed by a significant factor.
> They are helping their users create things that didn't exist before.
That is a derived output. That isn't new as in: novel. It may be unique but it is derived from training data. LLMs legitimately cannot think and thus they cannot create in that way.
I will find this often-repeated argument compelling only when someone can prove to me that the human mind works in a way that isn't 'combining stuff it learned in the past'.
5 years ago a typical argument against AGI was that computers would never be able to think because "real thinking" involved mastery of language which was something clearly beyond what computers would ever be able to do. The implication was that there was some magic sauce that human brains had that couldn't be replicated in silicon (by us). That 'facility with language' argument has clearly fallen apart over the last 3 years and been replaced with what appears to be a different magic sauce comprised of the phrases 'not really thinking' and the whole 'just repeating what it's heard/parrot' argument.
I don't think LLM's think or will reach AGI through scaling and I'm skeptical we're particularly close to AGI in any form. But I feel like it's a matter of incremental steps. There isn't some magic chasm that needs to be crossed. When we get there I think we will look back and see that 'legitimately thinking' wasn't anything magic. We'll look at AGI and instead of saying "isn't it amazing computers can do this" we'll say "wow, was that all there is to thinking like a human".
That is a pedantic distinction. You can create something that didn't exist by combining two things that did exist, in a way of combining things that already existed. For example, you could use a blender to combine almond butter and sawdust. While this may not be "novel", and it may be derived from existing materials and methods, you may still lay claim to having created something that didn't exist before.
For a more practical example, creating bindings from dynamic-language-A for a library in compiled-language-B is a genuinely useful task, allowing you to create things that didn't exist before. Those things are likely to unlock great happiness and/or productivity, even if they are derived from training data.
Yeah you’ve lost me here I’m sorry. In the real world humans work with AI tools to create new things. What you’re saying is the equivalent of “when a human writes a book in English, because they use words and letters that already exist and they already know they aren’t creating anything new”.
Why is that kind of thinking required to create novel works?
Randomness can create novelty.
Mistakes can be novel.
There are many ways to create novelty.
Also I think you might not know how LLMs are trained to code. Pre-training gives them some idea of the syntax etc but that only gets you to fancy autocomplete.
Modern LLMs are heavily trained using reinforcement data which is custom task the labs pay people to do (or by distilling another LLM which has had the process performed on it).
Could you give us an idea of what you’re hoping for that is not possible to derive from training data of the entire internet and many (most?) published books?
I think the confusion is people's misunderstanding of what 'new code' and 'new imagery' mean. Yes, LLMs can generate a specific CRUD webapp that hasn't existed before but only based on interpolating between the history of existing CRUD webapps. I mean traditional Markov Chains can also produce 'new' text in the sense that "this exact text" hasn't been seen before, but nobody would argue that traditional Markov Chains aren't constrained by "only producing the past".
This is even more clear in the case of diffusion models (which I personally love using, and have spent a lot of time researching). All of the "new" images created by even the most advanced diffusion models are fundamentally remixing past information. This is really obvious to anyone who has played around with these extensively because they really can't produce truly novel concepts. New concepts can be added by things like fine-tuning or use of LoRAs, but fundamentally you're still just remixing the past.
LLMs are always doing some form of interpolation between different points in the past. Yes they can create a "new" SQL query, but it's just remixing from the SQL queries that have existed prior. This still makes them very useful because a lot of engineering work, including writing a custom text editor, involve remixing existing engineering work. If you could have stack-overflowed your way to an answer in the past, an LLM will be much superior. In fact, the phrase "CRUD" largely exists to point out that most webapps are fundamentally the same.
A great example of this limitation in practice is the work that Terry Tao is doing with LLMs. One of the largest challenges in automated theorem proving is translating human proofs into the language of a theorem prover (often Lean these days). The challenge is that there is not very much Lean code currently available to LLMs (especially with the necessary context of the accompanying NL proof), so they struggle to correctly translate. Most of the research in this area is around improving LLM's representation of the mapping from human proofs to Lean proofs (btw, I personally feel like LLMs do have a reasonably good chance of providing major improvements in the space of formal theorem proving, in conjunction with languages like Lean, because the translation process is the biggest blocker to progress).
When you say:
> So it is absurdly incorrect to say "they can only reproduce the past."
It's pretty clear you don't have a solid background in generative models, because this is fundamentally what they do: model an existing probability distribution and draw samples from that. LLMs are doing this for a massive amount of human text, which is why they do produce some impressive and useful results, but this is also a fundamental limitation.
But a world where we used LLMs for the majority of work, would be a world with no fundamental breakthroughs. If you've read The Three Body Problem, it's very much like living in the world where scientific progress is impeded by sophons. In that world there is still some progress (especially with abundant energy), but it remains fundamentally and deeply limited.
Just an innocent bystander here, so forgive me, but I think the flack you are getting is because you appear to be responding to claims that these tools will reinvent everything and introduce a new halcyon age of creation - when, at least on hacker news, and definitely in this thread, no one is really making such claims.
Put another way, and I hate to throw in the now over-used phrase, but I feel you may be responding to a strawman that doesn't much appear in the article or the discussion here: "Because these tools don't achieve a god-like level of novel perfection that no one is really promising here, I dismiss all this sorta crap."
Especially when I think you are also admitting that the technology is a fairly useful tool on its own merits - a stance which I believe represents the bulk of the feelings that supporters of the tech here on HN are describing.
I apologize if you feel I am putting unrepresentative words in your mouth, but this is the reading I am taking away from your comments.
Lot of impressive points. They are also irrelevant. The majority of people also only extrapolate from the knowledge they acquired in the past. That’s why there is the concept of inventor, someone who comes up with new ideas. Many new inventions are also based on existing ideas. Is that the reason to dismiss those achievements?
Do you only take LLM seriously if it can be another Einstein?
> But a world where we used LLMs for the majority of work, would be a world with no fundamental breakthroughs.
What do you consider recent fundamental breakthroughs?
Even if you are right, human can continue to work on hard problems while letting LLM handle the majority of derivative work
> It's pretty clear you don't have a solid background in generative models, because this is fundamentally what they do: model an existing probability distribution and draw samples from that.
After post-training, this is definitively NOT what an LLM does.
as architectures evolve, i think it can be that we learn more "side effects".. back in 2020 openai researchers said "GPT-3 is applied without any gradient updates or fine-tuning" the model emerges at a certain level of scale...
> It's pretty clear you don't have a solid background in generative models, because this is fundamentally what they do
You don’t have a solid background. No one does. We fundamentally don’t understand LLMs, this is an industry and academic opinion. Sure there are high level perspectives and analogies we can apply to LLMs and machine learning in general like probability distributions, curve fitting or interpolations… but those explanations are so high level that they can essentially be applied to humans as well. At a lower level we cannot describe what’s going on. We have no idea how to reconstruct the logic of how an LLM arrived at a specific output from a specific input.
It is impossible to have any sort of deterministic function, process or anything produce new information from old information. This limitation is fundamental to logic and math and thus it will limit human output as well.
You can combine information you can transform information you can lose information. But producing new information from old information from deterministic intelligence is fundamentally impossible in reality and therefore fundamentally impossible for LLMs and humans. But note the keyword: “deterministic”
New information can literally only arise through stochastic processes. That’s all you have in reality. We know it’s stochastic because determinism vs. stochasticism are literally your only two viable options. You have a bunch of inputs, the outputs derived from it are either purely deterministic transformations or if you want some new stuff from the input you must apply randomness. That’s it.
That’s essentially what creativity is. There is literally no other logical way to generate “new information”. Purely random is never really useful so “useful information” arrives only after it is filtered and we use past information to filter the stochastic output and “select” something that’s not wildly random. We also only use randomness to perturb the output a little bit so it’s not too crazy.
In the end it’s this selection process and stochastic process combined that forms creativity. We know this is a general aspect of how creativity works because there’s literally no other way to do it.
LLMs do have stochastic aspects to them so we know for a fact it is generating new things and not just drawing on the past. We know it can fit our definition of “creative” and we can literally see it be creative in front of your eyes.
You’re ignoring what you see with your eyes and drawing your conclusions from a model of LLMs that isn’t fully accurate. Or you’re not fully tying the mechanisms of how LLMs work with what creativity or generating new data from past data is in actuality.
The fundamental limitation with LLMs is not that it can’t create new things. It’s that the context window is too small to create new things beyond that. Whatever it can create it is limited to the possibilities within that window and that sets a limitation on creativity.
What you see happening with LEAN can also be an issue with the context window being too small. If we have an LLM with a giant context window bigger than anything before… and pass it all the necessary data to “learn” and be “trained” on lean it can likely start to produce new theorems without literally being “trained”.
Actually I wouldn’t call this a “fundamental” problem. More fundamental is the aspect of hallucinations. The fact that LLMs produce new information from past information in the WRONG way. Literally making up bullshit out of thin air. It’s the opposite problem of what you’re describing. These things are too creative and making up too much stuff.
We have hints that LLMs know the difference between hallucinations and reality but coaxing it to communicate that differentiation to us is limited.
Over half of HN still thinks it’s a stochastic parrot and that it’s just a glorified google search.
The change hit us so fast a huge number of people don’t understand how capable it is yet.
Also it certainly doesn’t help that it still hallucinates. One mistake and it’s enough to set someone against LLMs. You really need to push through that hallucinations are just the weak part of the process to see the value.
The problem I see, over and over, is that people pose poorly-formed questions to the free ChatGPT and Google models, laugh at the resulting half-baked answers that are often full of errors and hallucinations, and draw conclusions about the technology as a whole.
Either that, or they tried it "last year" or "a while back" and have no concept of how far things have gone in the meantime.
It's like they wandered into a machine shop, cut off a finger or two, and concluded that their grandpa's hammer and hacksaw were all anyone ever needed.
No, frankly it's the difference between actual engineers and hobbyists/amateurs/non-SWEs.
SWEs are trained to discard surface-level observations and be adversarial. You can't just look at the happy path, how does the system behave for edge cases? Where does it break down and how? What are the failure modes?
The actual analogy to a machine shop would be to look at whether the machines were adequate for their use case, the building had enough reliable power to run and if there were any safety issues.
It's easy to Clever Hans yourself and get snowed by what looks like sophisticated effort or flat out bullshit. I had to gently tell a junior engineer that just because the marketing claims something will work a certain way, that doesn't mean it will.
From what you've described an LLM has not invented anything. LLMs that can reason have a bit more slight of hand but they're not coming up with new ideas outside of the bounds of what a lot of words have encompassed in both fiction and non.
Good for you that you've got a fun token of code that's what you've always wanted, I guess. But this type of fantasy take on LLMs seems to be more and more prevalent as of late. A lot of people defending LLMs as if they're owed something because they've built something or maybe people are getting more and more attached to them from the conversational angle. I'm not sure, but I've run across more people in 2025 that are way too far in the deep end of personifying their relationships with LLMs.
Hang on, you're now saying that if something has ever been described in fiction it doesn't count as invention? So if somebody literally developed a working photon torpedo, that isn't new because "Star Trek Did It"?
FWIW, your "evidence" is a text editor. I'm glad you made a tool that works for you, but the parent's point stands; this is a 200-level course-curriculum homework assignment. Tens of thousands of homemade editors exist, in various states of disrepair and vain overengineering.
The difference between those is the person is actually using this text editor that they built with the help of LLMs. There's plenty of people creating novel scripts and programs that can accommodate their own unique specifications.
If a programmer creating their own software (or contracting it out to a developer) would be a bespoke suit and using software someone or some company created without your input is an off the rack suit, I'd liken these sorts of programs as semi-bespoke, or made to measure.
"LLMs are literally technology that can only reproduce the past" feels like an odd statement. I think the point they're going for is that it's not thinking and so it's not going to produce new ideas like a human would? But literally no technology does that. That is all derived from some human beings being particularly clever.
LLMs are tools. They can enable a human to create new things because they are interfacing with a human to facilitate it. It's merging the functional knowledge and vision of a person and translating it into something else.
When a computer is able to invent things, we’ve achieved AGI. Do you believe we are already in the AGI era, or is the inventor in this case actually you?
What is the point that you believe would be demonstrated by a new text editor running at the limit of hardware in a compiled editor? Would that point apply to every other text editor that exists already?
I'm being hyperbolic of course, but I'm a little dismissive of the progress that happened since the days of BBS's and car based cell phones - we just got more connectivity, more capacity, more content, bigger/faster. Likewise, my attitude toward machine learning before 2023 is a smug 'heh, these computer scientists are doing undisciplined statistics at scale, how nice for them.' Then all of a sudden the machines woke up and started arguing with me, coherently, even about niche topics I have a PhD in. I can appreciate in retrospect how much of the machine learning progress ultimately went into that, but, like fusion, the magic payoff was supposed to be decades away and always remain decades away. This wasn't supposed to happen in my lifetime. 2025 progress isn't the 2023 shock, but this was the year LLM's-as-programmers (and LLM's-as-mathematicians, and...) went from 'isn't that cute, the machine is trying' to 'an expert with enough time would make better choices than the machine did,' and that makes for a different world. More so than, going from a Commodore Vic 20 with 4k of RAM and a modem to the latest Macbook.
Most people are capable of long-term learning. Some people are capable of discovering and inventing new things. I think the two are related, and current NN architecture doesn’t allow this. An AI that can cobble together a CRUD application to spec is one thing. An AI that can come up with a new idea for a successful app on its own is a completely different ball game.
Crypto bros in hindsight were so much less dangerous than AI bros. At least they weren't trying to construct data centers in rural America or prop up artificial stocks like $NVDA.
Crazy how the two most hyped and funded technologies of the decade were: energy wasting fake money for criminals and energy wasting plagiarism machines.
It's funny how people complain about the rust belt dying and factories leaving rural communities and so on, then when someone wants to build something that can provide jobs and tax revenue, everyone complains.
How many people are employed at the average data center? A few dozen? Versus a steel mill, that’s nothing. A chicken plant in Nebraska closed down this last month. 3200 people lost their jobs. You think Meta will fill it with GPUs and the whole town will have jobs again?
Many more are employed while building it. And they will never stop building. It's modern version of rail. But instead of distances it will cover the area.
As if any taxes will be paid to the areas affected, and add to that the billions in taxes used to subsidize everything before a single cent is a net positive.
Speaking of which, we never found out the details (strike price/expiration) of Michael Burry's puts, did we? It seems he could have made bank if he'd waited one more month...
There's no reason not to use Rust for LLM-generated code in the longer term (other than lack of Rust code to learn from in the shorter term).
The stricter typing of Rust would make sematic errors in generated code come out more quickly than in e.g. Python because using static typing the chances are that some of the semantic errors are also type violations.
Have we though? I'm glad we're not shouting about it from the rooftops like it's some magical "win" button as much, but TBH the things I use routinely that HAVE been rewritten in rust are generally much better. That could also just be because they're newer and have the errors of the past to not repeat.