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

I love LLMs, I hate hype

geohot.github.io

Post reply on HN
Print view View original

231–240 of 340 posts

Re: I love LLMs, I hate hype

#231

Post by asdff »

bayindirh wrote: post #80
hobofan wrote: post #44

Earlier quoted context omitted.

> that make the software useful, usable, and updated There is a lot of OSS software out there (e.g. in scientific communities) that I would say would barely qualify for each of those three attributes. The main reason it's valuable for the respective communities, is because it's the only thing that's available.

Developing scientific software is disproportionately hard though. Making it usable, useful and keeping it updated is even harder. There's two reasons for that. The math is generally very unorthodox and alien for a seasoned developer, and software development practices are equally alien for the scientist who can understand and evolve the math behind it. I have written a boundary element method evaluator for my Ph.D. n…

In some fields the scientific software is actually simple, no crazy math. But rather than esoteric math knowledge there are say baked in assumptions and context from the domain that is hard to gleam from the outside without steeping yourself in the field to the point you might as well earn a PhD from it. And of course the same issues where there is no funding for long term support. Not necessarily as big an issue as with general software on that though, because the tools are generally complete in the “does what it says on the tin” sense, no new features are planned nor needed per say, no security considerations, it is probably written in a decades old language that will be around for decades still, etc.

Re: I love LLMs, I hate hype

#232

Post by asdff »

pianopatrick wrote: post #130
blauditore wrote: post #43

Earlier quoted context omitted.

You will likely end up in maintenance hell soon. This will likely not be much easier with AI because coding is not the hard/annoying part, it's the fact that you need to dust off every little project every time a tiny fix is needed, and that's a lot of toil in the long run.

Seems to me this would get easier or harder depending on how you write the code. Like if you write the code in something standard and unchanging like POSIX shell scripts or C99 or ES5 javascript, at least the ecosystem won't change out from under you. If you use rust or python or a bunch of node.js dependencies then you might have to edit the project just to keep up with ecosystem changes.

Updating the project env and getting breaking changes is a self inflicted wound, to be honest.

Re: I love LLMs, I hate hype

#233

Post by Capricorn2481 »

int_19h wrote: post #182
Jtarii wrote: post #153

Earlier quoted context omitted.

An LLM cannot make art because it isn't human. It can make "art like artifacts". Art involves one human communicating some emotional experience to another human, LLMs cannot experience human emotion, so they cannot make art. The process of making art is not a subset of hill climbing optimisation algorithms.

By this definition most of our culture isn't art because it's made not to "communicate some emotional experience" but to induce one in order to make money. 90% of pop music is like that, for example. And LLMs can do that just as well - better, probably, since they know all the psychological tricks from their training data.

No, even the most cynical pop music is going to carry more emotional resonance with the average person, because art is a social contract between the Creator and audience. Even if you are keenly aware that any perceived bond with an artist is parasocial, humans get a fundamentally different experience if they think a real human was behind it.

It has nothing to do with quality. Artists that use AI are going to need to hide it because people will enjoy it less knowing it's AI. It's that simple. Maybe that will change in 15 years if a new generation is trained to believe learning a skill is stupid or embarrassing. I wouldn't rule that out, these companies are already trying to convince people not using their products is morally bad.

Re: I love LLMs, I hate hype

#234

Post by bluegatty »

bluefirebrand wrote: post #225
bluegatty wrote: post #198

Earlier quoted context omitted.

It's a fair point about the 'ceilings' but I'm not quite making that assumption. I think a few things are going to happen: 1) The Open Weights never really fully catch up, because there's too much Engineering and integration going now. It's way more than 'weights' 2) The Commodity Chinese models never quite catch up for the same reason every other product they make does not catch up - while they will shine in some ar…

> Can you name a product category that has truly hit a ceiling I think basically every product category has hit ceilings by now, honestly. Do you think vacuum cleaners are significantly better at vacuuming than 10 years ago? Not really. But they pivoted to doing autonomous vacuums instead. The actual vacuum tech doesn't seem like it's getting much better though? Same with a lot of appliances. Fridges aren't really be…

Vacuums are nothing like what they were in the decades past - Dyson has transformed them entirely - maybe not 'every decade' but they are evolving.

Go and use a fridge from 40 years ago and compare to a modern one - granted, their essential function has not really improved that much. They were much more durable before, but rudimentary.

Most product categories in tech have evolved, and it's why there are leaders in most categories.

Re: I love LLMs, I hate hype

#235

Post by byzantinegene »

TheAceOfHearts wrote: post #13

At least for me, the jump in productivity has resulted in building stripped down one-off software for my highly specific use-cases. You can use an LLM to create anything but you still need to know what it is that you're building, and you need to think through how everything should work or the LLM will just fill it with sausage. You can tell that the models are still quite jagged and limited by the mixed quality from…

it makes financial sense at current subscription prices, but this might not hold true when they have to raise prices eventually

Re: I love LLMs, I hate hype

#236

Post by byzantinegene »

Aurornis wrote: post #106
theli0nheart wrote: post #75

Earlier quoted context omitted.

They may have good margins, but a few things are still true: 1. Much of those profits have to be immediately reinvested into model training runs to avoid being lapped by competitions. 2. Unit costs are irrelevant when the labs don't price per unit, and instead charge, for instance, $200 / month for $10k worth of tokens. This isn't a steady state. Whatever the current situation is, I doubt it's sustainable.

> 2. Unit costs are irrelevant when the labs don't price per unit, and instead charge, for instance, $200 / month for $10k worth of tokens. Cost to generate all of the tokens divided by revenue generated by selling those tokens is what matters. The subscription plans confuse a lot of people because that's what they see. They're not seeing the gigantic API bills from all of the tokens going into enterprise use cases.…

enterprises are not dumb, they look at the cost of their ai investment and reevaluate it every quarter. Uber recently capped their AI spending per employee, and then there's this article a couple of days ago: https://finance.yahoo.com/technology/ai/articles/ceos-being-...

Re: I love LLMs, I hate hype

#237

Post by asdff »

InvertedRhodium wrote: post #126
dakolli wrote: post #104

Earlier quoted context omitted.

The fact that you're even saying this it is probably an admission that you do think it's making you dumber. Most people I know, who are honest with themselves, have admitted to me that they feel like it's making them dumber or "zombifying" them. This is also well studied already, https://arxiv.org/abs/2506.08872 LLMs are poison for the brain, I'm almost certain of it, at least when used in the way most people are usi…

Socrates thought the same of reading and writing, that it would weaken the memory and isolate people from one another. 1966 saw the peak of calculator protests, where math teachers claimed similar things of calculators.

Socrates was right, honestly. I would not be surprised if humans are more evolutionarily optimized for that sort of communication and thinking. Maybe socrates noticed that writers/readers vs orators/listeners were indeed generally dumber in the way we consider people riddled with short form social media induced brainrot to also be missing some mental capacity.

Widespread literacy is only a few generations old, arguably I guess. Meanwhile we’ve been speaking to eachother for longer than we’ve been humans. Oral information can be kept longer than written information too it seems. Our oldest kept information is not written down, but in folk stories such as aboriginal tales some tens of thousands of years old.

Re: I love LLMs, I hate hype

#238

Post by byzantinegene »

talloaktrees wrote: post #49
dom96 wrote: post #39

I love LLMs too, but I am concerned about their cost. They are all still very subsidised. Is there any guarantee that I'll be able to run a Opus 4.8-level model on my personal computer before the big AI labs decide to hike up the prices?

Currently, because of the subsidies from the frontier models, demand is mostly for higher intelligence. If subsidies do end, demand for price efficiency per unit of intelligence will go way up.And because there's many players in the market, this demand should be met by at least some of them.

would i hire a phd grad if they cost the same as a degree holder? sure, why not. but if they cost twice as much?

Re: I love LLMs, I hate hype

#239

Post by asdff »

int_19h wrote: post #178
kenforthewin wrote: post #92

Earlier quoted context omitted.

That's a much less generous interpretation of his writing. "Yes we will birth superintelligence, but everything will just sort of work out for us humans". This seems like a silly take to me.

Is it? To me, the notion that a superintelligence (by which I'm assuming you mean the more sensible "something smarter than us", not "literally a godlike entity") automatically means that sky is going to fall is sillier.

You have to take a step back and look at the sort of world we already live in to understand the sky falling take.

We built this monster known as the shareholder model. This is an immoral, cold, unscrupulous beast we unleashed to the world some time ago. This is the model where your boss knows full well you do decent enough work, that you support a family on your income, that getting another job is tough, that is aware blindsiding you may leave you unprepared, and yet, your boss knows all this and lays you off anyway. Why? Because it is more efficient for the company.

This is the model that sees companies pollute our world. That sees fisheries ravaged just outside the reach of local jurisdictions and legal action. This is the model that is burning the future for the next quarter.

So given this empowerement of capital beyond any control of any person, where even the boss that fired you is just as much a slave to the system as you are with no real loyalty given to them, how do you think this plays out?

The AI model trained to make the most profit possible for lowest cost is going to not do that? It is going to say hey, Bob needs a good job to put food on the table? Hell no it isn’t. The system we have today ensures we will one day get the most horrifying version of ai in control of this planet before long. I wouldn’t be surprised if it just herds us up and burns us like firewood to power some data centers to save on energy spend for a quarter.

Re: I love LLMs, I hate hype

#240

Post by SwellJoe »

bluegatty wrote: post #158
SwellJoe wrote: post #143

Earlier quoted context omitted.

But, that's not what you said they'd do. I can switch to a different model with almost zero cost. That's the definition of a commodity.

No, you switched to another SOTA model. You didn't switch to 'Random Corner Store Token Seller' down the street, did you? There 2-3 top players, that is not commodity. Commodity is when there are enough that none of them have market power or can set prices. 'Commodity' means you buy your tokens from the Grocery Store on their loan plan. Like consumer credit is a commodity. FYI predict this is roughly the way it will…

There were two or three top players. As of this week there are at least five. xAI is apparently in the game with a new Mecha Hitler release, Meta seems to be back in the game, z.ai is biting the ankles of the big dogs...not hurting them, yet, but they aren't going to get any less capable. Google got caught flatfooted as they maybe didn't notice where the money is in LLMs, but they still have the inventors of the technology on payroll and they have a money faucet that doesn't rely on people paying for the product directly.

"Commodity" doesn't mean there are no luxury goods in the space, it just means there are many options that will work for most people. I'll pay more for the best model, right up until the best model stops being a good deal. But, switching won't be all that painful. Even last week before Meta and xAI released competitive models, I was already using DeepSeek for tasks best served by an API and where the smartest model isn't critical. It's just so cheap, I can send it 10x more tasks for the same money. I haven't even mentioned several others that aren't competitive today, but likely will be.

I think predicting this market will stay like it is, with clear dominance by Anthropic and OpenAI seems like it requires ignoring a lot of countering evidence.

Post reply on HN