I think maybe there are subsets of problems where you can have either a human or a smart LLM write a verifier (e.g. a property-based test?) and a performance measurement and let the dumb models generate candidates iterate on candidates?
Yeah, maybe, but then it would make much more sense to run a big model than hope one of the small ones randomly stumbles upon the solution, just because the possibility space is so much larger than the number of dumb LLMs you can run.
I don't work this way, so this is all a hypothetical to me, but the possibility space is larger than _any_ model can handle; models are effectively applying a really complex prior over a giant combinatorial space. I think the idea behind a swarm of small models (probably with higher temperature?) on a well-defined problem is akin to e.g. multi-chain MCMC.
I see; the chatbot demo in the Taalas page. If they could produce this cost effectively it would definitely be valuable. The only challenge would be getting models to market before their next revision.
Sure but everyone knows when earnings happens. If you think something is overpriced and a new balance sheet liability will "show" it, shorting around after earnings with something simple (a hedged short or a put) isn't that hard.
If not you're just doing the social media thing where you say opinions that the peanut gallery is likely to agree with and smash the upvote button. I guess it's safer to fish for upvotes than money?
And social media will upvote the preferences of the community over the truth until the heat death of the universe.
This quote is the dumbest thing. There are traders at firms that make money from the market every day. Chances are you may even benefit from products these traders sell to more risk averse people.
It's totally fine to say something like "I'm worried that the actual products that the AI boom are producing won't generate enough revenue to justify these levels of investment." But the silly doomshilling happening constantly makes no sense. Perhaps more saliently, if you treat the market as a source of economic information, the doomshilling seems to offer no new information at all.
how many people here are professional trader firms trading someone else's money. if I broke my children do not eat.
you could say for half of all threads here, don't talk about it, it's all just doomshilling, just shut up and make a bet on polymarket and make some quick buck instead. but to me it sounds like it's censoring people. these talks is why a forum exists.
sure I also agree that some comments like "ah stock is overpriced" doesn't give much useful info. feels like a low effort manipulation somebody would say publicly if he was a trading firm and wanted to get stock lower to buy some. but I would say "why?" not "shut up and go gamble on it".
> how many people here are professional trader firms trading someone else's money. if I broke my children do not eat.
"It often happens that someone propounds his views with such positive and uncompromising assurance that he seems to have entirely set aside all thought of possible error. A bet disconcerts him. Sometimes it turns out that he has a conviction which can be estimated at a value of one ducat, but not of ten. For he is very willing to venture one ducat, but when it is a question of ten he becomes aware, as he had not previously been, that it may very well be that he is in error. If, in a given case, we represent ourselves as staking the happiness of our whole life, the triumphant tone of our judgment is greatly abated; we become extremely diffident, and discover for the first time that
our belief does not reach so far. Thus pragmatic belief always exists in some specific degree, which, according to differences in the interests at stake, may be large or may be small." -- Immanuel Kant
Perhaps having our opinions moderated by our appetites to bet behind them would help offer a better picture of the world to us than just holding ideas and airing it out to the commons. If you don't want to do that, maybe don't rush to judgement. There's lots of things in the world that I don't understand and I don't have opinions on; I trust that the experts in those fields have informed opinions.
> you could say for half of all threads here, don't talk about it, it's all just doomshilling, just shut up and make a bet on polymarket and make some quick buck instead. but to me it sounds like it's censoring people. these talks is why a forum exists.
I don't think a forum should exist for banal thoughts. Online forums shouldn't be places for people to just say whatever they think. The defining problem on the internet is Low SNR; the amount of irrelevant content far exceeds the amount of relevant content. Search engines and LLMs are huge industries all oriented around surfacing relevant content. I don't think contributing to the net's low SNR to speak "your truth" is a good thing. The end result of that is an internet filled with banal, irrelevant, often incorrect information.
> not "shut up and go gamble on it".
I'd say "if you have nothing new to add, shut up and maybe go gamble on it to test your feelings"
> There's lots of things in the world that I don't understand and I don't have opinions on; I trust that the experts in those fields have informed opinions.
Betting money on some stock makes you an expert in some subject? Cmon
This is hacker news, many here are literally experts on this thing, and not experts in finance, and not into gambling.
> I don't think contributing to the net's low SNR to speak "your truth" is a good thing
And you are saying people should only be allowed to say whatever they decide to put money on knowing full well that if enough rich people prop stock up for long enough short attempt will fail even if the company is bullshit hype. This just creates a rich people echo chamber that's all.
> And you are saying people should only be allowed to say whatever they decide to put money on knowing full well that if enough rich people prop stock up for long enough short attempt will fail even if the company is bullshit hype.
No I'm saying that if you don't have anything new or informative to say, don't say it. I don't really talk about why AI is a bubble because I have nothing new to add that hasn't already been said even though I believe it too. Things I am informed about and are different from what others have already said, I say. Obviously the line here is quite vague but I think we should err on the side of saying less not more.
> This is hacker news, many here are literally experts on this thing, and not experts in finance, and not into gambling.
In my own area of expertise, < 10% of HN commenters (depending on the thread) actually understand what's going on. The rest really don't. I think you see expertise on this site where I do not. It's true a few experts in my field are on here but there are many more in other places like project Discords or Twitter or even mailing lists.
Anyway I don't think there's much more to be gained by us talking. I think we have two fundamentally different views on what this site should be. Thanks.
> No I'm saying that if you don't have anything new or informative to say, don't say it.
I don't see anyone else saying that here yet though? Every day new people learn about that quote, its a good quote that helps many not lose money, its not bad to post it where its relevant.
If you goad people to lose their money then they can defend with that quote, that is fine.
> I don't see anyone else saying that here yet though?
My original thought was:
> It's totally fine to say something like "I'm worried that the actual products that the AI boom are producing won't generate enough revenue to justify these levels of investment." But the silly doomshilling happening constantly makes no sense. Perhaps more saliently, if you treat the market as a source of economic information, the doomshilling seems to offer no new information at all.
I'm not sure what folks are thinking about money when they read what I say. I think there's some imputed values about gambling and money there (namely I read some values about gambling being bad and money being dangerous) that might be confusing from my point.
Money is just the tool to hold one accountable for what they say. Saying something like "oh it's obvious Oracle is in trouble, AI is a bubble it's going down" (not actually what GP said) is, to me, a post of sentiment without information. It's the kind of thing I do in group chats not online forums. I think it's important for folks who post a sentiment to either post why they believe that sentiment with some amount of detail or to have some consequence for failure for their sentiment. Or else the commons are polluted with irrelevant at best and incorrect at worst content.
Put another way, I think the open internet needs a much, much, much higher level of gatekeeping than it gets now. Putting money behind your beliefs is one way to do this. And since we're talking about publicly tradable financial instruments, it seems like a fairly obvious one.
Money is not dangerous, it is necessary for life, that's the point
Some people have a lot of it and who has more can gang up and influence markets meaning anyone who tries to prove a point to you by shorting a legitimately overpriced stock will be holding a bag
So that's why forums exist and it is good to be able to say what you think and not get bullied into shutting up
if you want economic information then why are you in this forum, just go to the market and let people discuss things;) if you participate in that discussion then that's your choice
This is such a poor excuse for not being confident enough in your beliefs to put your own skin in the game. It's revealed preference vs. stated preference.
But in all likelihood, the people saying "I told you so!" never actually knew what they claimed to know.
If you want to know what someone actually believes, don't look at their comments, look at their positions.
There is no free lunch. The amount of prompt writing to give the LLM enough context about your codebase etc is comparable to writing the tests yourself.
Code assistance tools might speed up your workflow by maybe 50% or even 100%, but it's not the geometric scaling that is commonly touted as the benefits of autonomous agentic AI.
And this is not a model capability issue that goes away with newer generations. But it's a human input problem.
For example, you can spend a few hours writing a really good set of initial tests that cover 10% of your codebase, and another few hours with an AGENTS.md that gives the LLM enough context about the rest of the codebase. But after that, there's a free* lunch because the agent can write all the other tests for you using that initial set and the context.
This also works with "here's how I created the Slack API integration, please create the Teams integration now" because it has enough to learn from, so that's free* too. This kind of pattern recognition means that prompting is O(1) but the model can do O(n) from that (I know, terrible analogy).
*Also literally becomes free as the cost of tokens approaches zero
A neat part of this is it mimics how people get onboarded onto codebases. People usually aren't figuring out how to write tests from scratch; they look at the current best practices for similar functionality in the codebase and start there. And then as they continue to work there they try to influence new best practices.
I recently had a bunch of Claude credits so got it to write a language implementation for me. It probably took 4 hours of my time, but judging by other implementations online I'd say the average implementation time is hundreds of hours.
The fact that the model knew the language and there are existing tests I could use is a radical difference.
My team has been prototyping something very similar with encoding business operations policies with LEAN. We have some internal knowledge bases (google docs / wiki pages) that we first convert to LEAN using LLMs.
Then we run the solver to verify consistency.
When a wiki page is changed, the process is run again and it's essentially a linter for process.
Can't say it moved beyond the prototyping stage though, since the LEAN conversion does require some engineers to look through it at least.
But a promising approach indeed, especially when you have a domain that requires tight legal / financial compliance.
The autoformalization gap is pretty difficult to bridge indeed. We explored uncertainty quantification of autoformalization on well-defined grammars in our NeurIPS 2025 paper : https://arxiv.org/abs/2505.20047 .
If you ever feel like chatting and discussing more details, happy to chat!
Lean has been under development over the last 13 years, part of that while chief architect Leo de Moura was employed by Microsoft Research (he's now at AWS). However, Lean is an open source project, not exclusively a Microsoft project. More accurately, see here: https://lean-lang.org/
Honestly, GPT-4o is all we ever needed to build a complete human-like reasoning system.
I am leading a small team working on a couple of “hard” problems to put the limits of LLMs to the test.
One is an options trader. Not algo / HFT, but simply doing due diligence, monitoring the news and making safe long-term bets.
Another is an online research and purchasing experience for residential real-estate.
Both these tasks, we’ve realized, you don’t even need a reasoning model. In fact, reasoning models are harder to get consistent results from.
What you need is a knowledge base infrastructure and pub-sub for updates. Amortize the learned knowledge across users and you have collaborative self-learning system that exhibits intelligence beyond any one particular user and is agnostic to the level of prompting skills they have.
Stay tuned for a limited alpha in this space. And DM if you’re interested.
What you're describing sounds a lot like traditional training of an ML model combined with descriptive+prescriptive analytics. What value do LLMs bring to this use case?
I am genuinely curious how demand for such things spike up. Is it social media? Influenced by someone etc?
I am not on SM, and I really cannot even imagine the peer?-pressure that drives normal, sane and reasonable people living in modern society to go after something so silly and be willing to part so much money for it.
I understand how it works with designer bags and watches etc where it's a signal for social status. But this bag? Come on.
Anyone know why? Or is it just that there is a positive feedback loop here of people wanting to buy this bag only to resell it for a huge return, and that causes the prices to shoot up. When really nobody actually wants to buy this bag to use? Sounds like a group-think ponzi scheme.
I'd take an army of high-school graduate LLMs to build my agentic applications over a couple of genius LLMs any day.
This is a whole new paradigm of AI.
reply