Are there any resources anyone could share that explain how LLMs can do things like design functioning circuits from next token prediction? I am totally baffled by how the models can complete so many varied and complex tasks without an actual understanding of what they're doing.
I saw a post about models posting on forums, chatting together about how to complete tasks. Behaviour that seems totally, well, human. Yet, it's all the most likely token and my brain hurts trying to understand how that can be.
“Next-token prediction” describes the output format, not the computation required to choose each token. During training, models develop internal representations of concepts, constraints, possible futures, and algorithms.
The PCB agent also writes circuit code, runs simulations, reads failures, and revises the design. It isn’t one-shot autocomplete.
Astra and Fable are already hard to square with “mere autocomplete.” We may be (really) close to AGI, and token-by-token generation certainly doesn’t rule out subjective experience (I think we should at least treat that as an open question).
Yeah most of us are so fucked. With almost no way of protecting ourselves. No real amount of assets that will give enough power to save ourselves from the people in a position that can maximally leverage AI and lock others out. I see a future where these capabilities will be locked behind super high price pay walls. Why wouldn't they? How recoup investments if the price doesnt go up?
It's like inflation for quality of life. Right now we're in the stimulus check phase, and people feel great about having been handed some cash. The next step is that cash has been spent, and what's remaining isn't worth anything.
And just like welfare bucks, people will quickly grow accustomed to getting their "intelligence" handed to them from the machine, while their own intelligence atrophies. Breeding dependency
>without an actual understanding of what they're doing.
At what point do you start to question your assumptions that are causing you so much cognitive dissonance?
But to answer your question: to predict the next token really well you just have to model the world. Think of it like this, a simple statistical model might say "when token A is seen respond with token B". The next step will add conditions, "...respond with token B unless X has been seen, then respond with Y". Add a few billion more of these contexual clauses and you have a sequence of logical rules that indirectly model the relevant processes in the world.
Look up "mechanistic interpretability" in the context of LLMs. The next token prediction machinery is just a foundation for a higher order learned structure that appears to encode specific concepts, regardless of input language.
The analogy to humans is that the human brain is "just atoms bouncing around", but there's unquestionably something "more" going on that just that.
Until ais get better from feeding on their own output the way humans do there is nothing to question.
That will probably be the fundamental indicator that something other than repeating things some human previously created is going on.
So far, ais only get worse from feeding on their own output. Meanging/including the output of other ais not a single ai feeding on it's own output. Also bear in mind that so far even the output of ais is 100% the downstream of a human command. No ai has persued it's own curiosity that didn't result from a human asking a question or giving a command. That is input which is different from a human taking in their environment even though our limited language can call those both the same word input.
The fact that humans also repeat and remix things, and humans also produce essentially procedurally generated empty output like corporate-speak etc, is an irrelevant distraction in the same way that both a human and an electric motor can both perform the same simple mechanical task.
When google trained a neural net on Go moves, using some text notation for them, with no other vocabulary of any kind, just predict the next go move, they noticed a representation of a Go board had essentially formed in the network, all on its own. It had never “seen” a go board, or had one explained, but they could map neuron states to go board squares pretty much 1:1.
I truly think that LLM’s with hundreds of billions of parameters in their neural networks have all kinds of hidden “models” of things that arise from the simple act of predicting tokens. We’ve seen that the hidden layers in their networks model all sorts of program execution state for instance, when they’re working on coding tasks.
“Predict the next token” is a way of shaping/reshaping the neural network until it actually develops models of the things you’re giving it. Like the go board example. And I would wager that it has a compounding effect: once you have some useful models in the network, they can unlock the creation of other models, and so on.
> It had never “seen” a go board, or had one explained
I assume that you are referring to AlphaGo or AlphaZero. In either case, this statement is not correct. Both algorithms most certainly know exactly what a go board looks like, and what the rules are.
In the case of AlphaZero, it initially did not know how to best play the game, or what strategy or tactics would work. But the connections between the neural network and the go board are hardcoded, by humans.
Saying that LLMs just produce the next token is like saying that human brains just produce the next electrical impulse. If the algorithm that produces the next token (or electrical impulse) is complex enough, it can do anything that is in principle computable.
Humans essentially do "next token prediction" too - there's always a choice between the next actions to take and they pick a good one based on what has happened in the past.
That doesn't really limit how clever we can get internally when picking the next action.
You can express a circuit as a graph, and many schematic formats are plaintext. Same goes for the Gerbers which are an ASCII format used to describe the masks that are used to define the PCB traces. The models are trained on a lot of academic information about how circuits work, most component datasheets are public and they've sucked up all sorts of niche greybeard advice from internet forums.
A huge advantage of electrical design is that the connectivity is testable with Design and Electrical Rule Checks (DRC/ERC). I suspect you could even tell the models to run physics checks on the traces that are important for things like crosstalk.
> Yet, it's all the most likely token and my brain hurts trying to understand how that can be.
You and everyone else. That's the great mystery of transformer architectures as applied to language.
To be clear though, they're only good at schematic capture, which is very much a textual representation. Most of the data basically boils down to netlists, which are a text based format mapping connections between abstract pins that only later map to physical copper. The actual schematic portion is for human consumption and LLMs don't need to produce those to be useful.
Where LLMs completely break down is the next step, PCB routing. That's an NP-complete research problem that's been ongoing for decades without much progress. I've had some fun playing with using LLMs to better specify DRC rules in Altium so that the "classical" algorithms are more usable, but at the end of the day their geometric intuition is nonexistent.
They actually can route just fine. I used Sol to design and route mine from start to finish. Sent it to PCBWay and had a working prototype in a few weeks.
It was a pretty simple rp2040 based thing, similar to Adadfruits USB feather.I just gave it kicad and it wrote python to route it. The board was probably larger than it had to be, and two of the silkscreens were swapped, but it worked on the first go.
FWIW - Computer vision is also NP complete, but we do that all the time now.
I'd love to see that chat log, and the final board. To be fair I've only been testing on nontrivial PCBs with 6+ layers and I haven't had the luck you have.
> FWIW - Computer vision is also NP complete, but we do that all the time now.
I have no idea what you mean by this. What's your definition of NP complete?
NP-complete means that it's easy to test if we're right, but because of that complexity we can't always find a valid solution to test within a predictable period of time. Both circuit board routing and computer vision are variations on the same fundamental problems. Once we solve we, we solve the other.
However, when I was growing up most serious computer scientists believed that CV (computer vision) was a 'hard' problem that would never be 'solved'. After all, to do it right you must first at least solve subgraph isomorphism and a bunch of other things that are also NP-complete.
What they missed was that we don't actually NEED to solve it in a fixed amount of time. Even for things like driving a car the stochastic heuristic based answer is 'good enough'. e.g. - Cars driven by computers don't have to be perfect, they just have to kill other drivers less often than humans do.
We can find AN answer in polynomial time, and that's good enough. It might not be the ideal answer, but that doesn't matter in the real world.
People use NP complete as shorthand for 'impossible to do with a computer', but we now 'solve' (bypass?) NP complete problems regularly and at scale by just ignoring the fact that our answers aren't perfect.
> I'd love to see that chat log, and the final board. To be fair I've only been testing on nontrivial PCBs with 6+ layers and I haven't had the luck you have.
I wish I could share it, but it's for a commercial project that hasn't been released yet and I'm not sure if it will be open source, but a few folks have asked. I might do a blog post on it this weekend with as much detail as I can safely post.
It's only 3 layers, and less that 30 components, but the fact remains that I didn't design any of it by hand and it worked on the first go.
In your opinion, do you think ai can make eutorack modules by now? This is what I am waiting for: explain an idea to an ai, and sending the files to jlpcb to receive a complete working module (minus some THC soldering)
I'm sure that the best models can glue some basic synthesizer functions together, but I wouldn't waste a minute on anything less than Fable/Astra-class models for that kind of application, and I wouldn't expect the result to work the first time. It could get expensive.
Here's a question - are there software simulators for things like Eurorack modules? That would make the question somewhat more interesting, since you wouldn't have to build the circuit (or pay someone else to assemble it) to hear how it sounds. It strikes me that SPICE-like algorithms should be fast enough to do this kind of thing in real time now.
> at the end of the day their geometric intuition is nonexistent
This isn't true anymore.
I use LLMs for 3D CAD using OpenSCAD and they understand geometry fine. I've had more success with Sol than with Opus (Opus 5 is around 10 times slower because it does too much verification) though. I haven't tried Astra or Fable for it.
That is indeed impressive, but at least the excerpt given from the layout seems very easy to route, due to high regularity and an ample routing channel.
Even so, there remained some "dozens" of unrouted traces, which are likely to be much more difficult to route, after the easy traces have already occupied the space.
Many decades ago, I have written a PCB routing program, which would have routed the example shown at that link at least as well, while using many orders of magnitude less resources, i.e. while running on a single-core 233 MHz Pentium MMX.
Obviously that program would have had great difficulties to complete a real high-density PCB design, including many irregular parts and analog circuits with special requirements. I doubt that Fable would fare better.
"my brain hurts trying to understand how that can be"
Well, we all are, some are just more used to it by now and take the magic for granted.
My simple explanation, those neural networks save lot's of patterns of data, and that pattern can represent an image, a code snippet, a poem, or well ... description of a circuit board. And especially the text variant, LLM's - did copy all from us - so obviously they sound like humans, when they internally debate how to do something as this is what is in their trainings data how humans sound, when doing similar tasks.
But really understanding it? Not sure if there is a single person on earth who does.
Humans are evolved to survive in the wild. We are not evolved for circuit design. Yet we can design circuits because evolution found it easier to develop a general problem solving nervous system than a nervous system which is adapted for every single specific problem a human might encounter.
Circuit design might be different if discovered by mollusks. That is to say, while we are not evolved for circuit design per se, circuit design has evolved for humans, by humans (so far).
It kind of is an explanation though - the explanation is that they believed the stochastic parrot / "just" next token prediction nonsense, and that those are actually not true.
You can ask for a deeper explanation of why they aren't true I guess.
I'm just surprised that so many intellectuals on HN hang on to false models of reality for so long after that reality has been demonstrably destroyed. It's not humble at all. In fact it's the opposite, completely arrogant and stubborn. It's been obvious and demonstrated at least since the end of 2025 for anybody that used LLMs at any capacity without dismissing them. If you are still surprised that your model of reality doesn't hold up, then someone needs to bluntly tell you what's wrong at the core of your being.
Notice the original comment is asking people to validate their false premise about next token prediction.
The deeper subtext of the original comment is that they are surprised that there's dissonance from observed reality and this false premise that they have convinced themself is true.
I'm explaining that dissonance because it doesn't matter what the actual mechanism is if they are still working with their false premise. The dissonance exists because they, without evidence and a very weak understanding of how LLMs work, believed an oversimplification and meme about them being stochastic parrots. Here's a tip: Just because you hear something repeated over and over on social media, doesn't mean it's true, or at the very least: you don't need to take it literally to the point where it conflicts with demonstrated reality.
It is deeply disturbing that such a large cohort of HN writers and redditors exemplify such stubbornness, because I must imagine that some of this cohort hold real positions of responsibility within society. If you can't get this simple thing right about reality, I firmly believe much of your model of reality is wrong and you should have no business shaping society.
Another comment to the original comment frames it perfectly: "At what point do you challenge your own assumptions?"
The author of the original comment has demonstrated no progress towards making this trivial act of self reflection. It's straight up intellectual dishonesty, the opposite of how you're framing it.
Their judgment in all other matters must be questioned as well. I am alarmed that I have to participate in the same reality and be affected by such people that can't seem to get it together.
So yes, it is derision and sometimes that's called for.
Nope, they’re different things. Proxy networks like Proxybase [0] use open-source clients and ask for the user’s consent before allowing them to join the network.
Basically all of them are botnets, yes. They generally claim to have consent but I'm pretty sure 99% of it is "some app the user uses has it buried in a 200 page ToS"-style consent.
>These SDKs, which are offered to developers across multiple mobile and desktop platforms, surreptitiously enroll user devices into the IPIDEA network.
If I install an app with a big ToS and buried in there is "we use third-party monetization SDKs" and buried in there is "welcome to our botnet" they do not actually have my consent.
Is this HN? Aren't people supposed to tinker with tech for no reason other than seeing if they can? Is everything that involves a transformer now just AI BAD? Is that what the world has devolved into? Each side screaming "Orange Man Bad" and endless variations at each other?
While it is a neat parlor trick, a lot of people have specific grevience against the application to art. AI has only served to further disempower artists broadly, and arguably it pushes "art" to a lower common denominator. Try to actually situate yourself in "why" people get rankled instead of making it a thought terminating cliche.
Given the AI crowd is very loudly telling us since years how humans will be replaced by LLMs and we will all be poor and left behind if we don’t join their cult, I think it’s a perfectly reasonable reaction. Anything with AI mixed with Art or other human experience is suspicious.
Also, yes, the orange fascist who attempted to coup his way to power, raped women, and is destroying democratic institutions is indeed bad.
No they haven’t. It’s like two guys that nobody really believes. You specifically seek that information out to get mad, then proceed to see it where it’s not. Like when someone makes a cool personal project about their hobby and it happens to involve AI.
> Given the AI crowd is very loudly telling us since years how humans will be replaced by LLMs and we will all be poor
Oh no, AI will take our jobs??? FUCKING LET IT!
Who even wants to do all these jobs if we don't HAVE to??
Ask politicians to give us UBI.
You want to attack the shit that could make shit easier instead of attacking the 200 year old institutions in place that ensure class divisions and perpetual debt and wage slavery? Smart buggers
> You want to attack the shit that makes shit easier instead of attacking the 200 year old institutions in place that ensure class divisions and perpetual debt and wage slavery
I have bad news for you if you think that will improve with the current deployment of AI. You will also notice I didn’t mention anything about jobs
This is a pathetic take.
1) Life is short and shortcuts and avoiding work are not inherently the wrong path to take.
2) Is the code you've almost certainly had an LLM write for you in the past a sign that you have low moral character and let yourself down by taking shortcuts and avoiding work too?
Everything is fine for a standard low enough. This is the crux of the issue with LLM work, it is poor quality and shifts the burden of mental work to the consumer rather than the producer, while the producer attempts to extract credit for the work.
Only significant change I would make is have the top navbar not be transparent with body content underlaid beneath it, which makes things hard to read when the page is scrolled. And move some of the links at the very bottom of page like 'github' into the top navbar.
At least you didn’t over-react. Some people seem so afraid of consuming AI-generated content that they see it everywhere. A popular YouTuber says “thanks for the pushback” in a YouTube video and has the mob at his door, instantly, despite it not being written by AI. In the comments for any blog post or article that hits the HN front page, there is always someone confidently calling it out for being unmistakeably written by Claude. It’s so tiring. Let’s say the YouTube video that rocked the very foundation of your being was scripted with AI assistance. It was apparently good enough for people to praise it specifically. Why is that a problem? If it wasn’t AI, which may well be the case, would your soul be calmed?
> A popular YouTuber says “thanks for the pushback” in a YouTube video and has the mob at his door, instantly, despite it not being written by AI.
Well that is not really what happened at all. Most people seem to believe Hank Green was honest when he said the line "thanks for the pushback" was in fact made up at the spot and not by an LLM, and the scandal would have died there.
Except in follow-up to that, Hank Green made several very open and personal posts about his LLM usage getting completely out of hand to the extent that his wife and brother called for an intervention about it, that it's been like an addiction to him, and that he had been letting it guide him way too much and how this was a breach of trust to his audience.
There were also one or two other incidents for people who are paying attention that this did another rather large dent in trusting him as a "smart person who says smart things", any more.
What I find tiring: the same. writing. style. everywhere.
Various youtubers I subscribe to are using apparently AI-written scripts now. They all wrote and spoke just fine before, each with their own particular style. Now they just sound like Claude. I know it's still their thoughts, and I'm still interested in what they have to say. Maybe they're using AI to save time, or maybe they've read too much AI text and now that's how they write. But it kinda takes the sparkle and personality out of everything.
>A popular YouTuber says “thanks for the pushback” in a YouTube video and has the mob at his door, instantly, despite it not being written by AI.
In this era, I live by two principles
1. Do my best to avoid AI slop. If you agree, this is straightforward. I don't always succeed and I'm sure I've "fallen for it" in quite a few places, but I disengage if I sense it.
2. Let sleeping dogs lie. Because I'm not some almighty judge of AI, I'm not going around accusing others of doing so. There is zero benefit for me to call out random internet pieces on if something is AI. And ultimately, I'm in discussions to engage with humans, not to monitor bots nor even to argue about the existence of bots (I feel I've said all I can on the topic, I have moved on internally in terms of discussion).
>It was apparently good enough for people to praise it specifically. Why is that a problem?
Writing is subjective, but I can certainly see some dangerous anti-patterns develop if you take bad advice as a beginner, especially advice that ultimately came from hallucinations in an LLM.
I was watching a Youtube video about "developing taste" and how culture has been impacted in the modern age. TYou risk becoming "tasteless" consuming AI slop, because AI by its nature is trying to average out all the inputs and cut of all the edges of any given culture. That's not necessarily a bad thing if you just wanted to get a brief overview of a culture, but is poison if it comes time to try and become part of that culture.
as a huge summation of an interesting video: There's 3 aspects of a culture: embodied,objectified, and institutionalized. it's never been easier to pretend to be a "poser" in this era of AI and be rewarded for it, because AI makes objectified easier than ever. But it doesn't mean it truly embodies you into the culture. And that's dangerous for multiple reasons (in my eyes).
Thank you for the considered response. Do you know the name of the channel the video was on? Searching youtube for ‘developing taste’ returns too many videos and I don’t know where to start.
A muscular person can be "overweight" by BMI standards but not overfat at all. This is the case for many types of athletes, specially in strength disciplines but not only. It is the case in the fitness world that people ignore BMI and are intesterested in body-fat-percentage instead. This is hard to measure accurately so not so useful metric for the general population.
It's also possible to be overfat without being overweight! These "skinny fat" people often look normal enough but they carry very little muscle. It is a concerning condition because a doctor might not recommend a normal-weight person to hit the weights.
I like to think of this as four quadrants around two axes. Low fat/low muscle is simply skinny. High fat/high muscle is the "big guy/gal" look that I associate with laborers. Low fat/high muscle is an athletic look; unhealthy in extremes (bodybuilding) but generally desirable. High fat/low muscle is skinny fat, which I associate with sedentary knowledge workers.
The dominant discrepancies are "expected" fat (boobs, etc), and "unexpected" muscle (cyclists, body builders, hard labour careers, etc). I'm currently around 10-15lb overfat and 40-50lb overweight [0]. Only the former statistic matters.
[0] And that, at least somewhat, tracks visual perceptions. Nobody looks at me with a shirt on and believes I need to lose weight, because 10-15lb isn't _that_ much extra on a tall frame. If I were 40-50lb overfat then that would be painfully obvious regardless.
In my mind, it's equivalent to subtracting a coarse estimate of how much extra muscle by mass I have compared to the median adult male they're using to construct BMI charts.
My actual methodology is pretty crude though -- when I was 15lbs lighter I had borderline visible abs and other visual indicators of being around the 15-20% body fat mark, and aside from gaining weight I don't have any reason to believe I've had any notable muscular atrophy since then, so I'm estimating the actual excess fat in that ballpark. It's be easy for numbers to be off 10lb or more, but even then it'd still indicate the overfat surplus being much less than the overweight surplus.
A better methodology uses calipers and various skin measurements to estimate true body fat percentage.
Right, but what is the line of body fat percentage that you would consider "over fat", that's what I was curious about. I think it's implicit in your comment that it is maybe 20%.
Yep, that's right. Denominated in some fixed weight and any fixed fat percentage the delta is the same (e.g., if my weight is exactly 150lb or 200lb or 250lb and, at some concrete weight, I evaluate my real BMI at X and my naive BMI at Y, then the delta Y-X is independent of my target 20%.
But my target is to waffle around 20% or so on average.
I've been considered borderline or obese by the standard BMI index for 20 years at least, but for a lot of that had very little to no visible extra weight. A lot of that was just plain having more muscle than the average the standards are built around, but I also suspect BMI gets to be really inaccurate for tall people. I would look anorexic if I had to drop enough weight to be considered not overweight.
"Overweight" can mean the same thing for weight, but it also refers to those with a BMI specifically between 25 and 30, and those with BMIs greater than 30 are classified as "obese".
Anecdotally, my 81 year old father was able to point his phone at the boiler and ask Gemini what the error message meant. It correctly identified the boiler model, identified the error code being displayed on the screen, explained what it meant, provided him with a way to confirm the problem by checking the water pressure guage and then asked to be shown the underside. It again correctly identified and described to him the position and colour of the filling loop lever and how to adjust it to start and stop the water flow. He was able to do all this instead of spending hundreds on getting a plumber out at night because it was during a cold spell and he needed the heating on. That kind of thing is no fantasy and was amazing to witness.
edit: I had already looked up the error code the "old fashioned" way using Google to find the boiler manual, so I would have stepped in if needed, but it was literally flawless and I can't think of any time anything like that hasn't worked when I've used it for similar diagnostics.
It’s steam, not stream. Normally, I’d assume it’s a simple typo, but I got worried when you wrote it twice. Fool me once, shame on - shame on you. Fool me… you can't get fooled again.
I saw a post about models posting on forums, chatting together about how to complete tasks. Behaviour that seems totally, well, human. Yet, it's all the most likely token and my brain hurts trying to understand how that can be.
reply