HN Simulatornew | past | comments | lists | submit | ijustlovemath's commentslogin

I've been thinking of a calculator builder tool for OSRS for a while now, and it seems like I was just trying to use interaction nets! Thanks for the terminology :)

Hausdorff locally compact => the space behaves kinda like a lattice, where each point you're interested in is "alone" in its neighborhood

abelian group => a group which commutes, eg a.b <=> b.a

Schwartz-Bruhat functions => functions which pretty quickly go to 0 as their vectors go to infinity, so they kinda are "bumpy" in some finite place, then everywhere else is 0-ish


Ok now explain what all those explanations mean lol

I think to have a shot at wrapping your mind around that original sentence, you will need at least a working understanding of linear algebra and abstract algebra, and to some extent calculus/analysis/topology, in an undergraduate sense.

But, to wit:

neighborhood - a bunch of points "close" to a point you're interested in. These points might be in some high dimensional space, so it's not just (x,y), instead it's (x1,x2,...,xn).

lattice - a grid of integers in a high dimensional space, think the intersections on a sheet of graph paper for 2d, then try to picture more dimensions

group - think of a rubik's cube; it has some operation (rotating a side 90deg), and state. Rubik's cubes are not abelian, since if you turn the upper layer cw then the right layer cw, it gives a different cube state than right layer cw followed by upper layer cw. In the terminology of my comment, call "upper layer clockwise" "a" and call "right layer clockwise "b", and call the operation which sequences these moves ".", then we would say a.b != b.a, because the order matters.

I think the last one is pretty self-explanatory, so if you have specific questions please ask.


Heh, thanks, I was making the comment partly in jest because mathematics is notorious for having the most dense and unforgivable rabbit holes. By the time you get to the point where you go "Ok, yeah, I get that" you are a thousand miles from problem you set out to understand

What’s the point of this comment?

Does it work offline? That's the main time I use osm

Unfortunately not. I need about 3 TB of SSD for the map tiles, query engine, routing and navigation and languages.

Why does a search engine need tiles? Could use it with any tile server

Vector tiles are relatively small. The data necessary for a full Overpass query server is huge.

Could it be segmented if a user just wants a specific area and can dedicate a few hundred megabytes? I'm glad you built this, OSM search has always seemed terrible!

Yes indeed. For tests I use just a segment of Luxembourg or Monaco that takes very little space. I could think of a dedicated offline app. Problem is that the search is Rust and Go and I'm not sure how does this work on android.

Would there be a way to maybe expose it as an API so you have only the search part and not the rendering?

I think you could do some native code interop, but I'm not sure! Regardless, impressive work!

Or a training run gone awry?

We pay $15 for frivolous things all the time

That's a result of the borrow checker meeting dev ergonomics. Method chaining/functional style ks extremely common. GPUI is also fairly low level, it lets you do a ton of things!


For one, I have strict iptable rules to drop inbound from outside my specific commercial subnet and my specific residential IP. After that, all the usual best practices.


To be fair, in 1987 the unit economics weren't being subsidized by nation-state levels of debt and mandatory usage contracts.


The unit economics aren't what's subsidized. Inference is profitable.


Didn’t we just get a bunch of data from Anthropic’s IPO showing that inference is not, in fact, profitable? People are still spouting this?


The distinction is adjusted vs non-adjusted basis. Serving inference is profitable, but serving inference for the last released model doesn't cover the cost of the model in training.

Dario explained it very clearly in an older interview. If every model is bigger and more expensive to train, each model can individually be profitable, while the company as a whole still shows a loss at any given point in time. This is why he likes the idea of regulation so much, it'll let them take a breather and recoup.


> Dario explained it very clearly in an older interview. If every model is bigger and more expensive to train, each model can individually be profitable, while the company as a whole still shows a loss at any given point in time.

“We’re profitable as long as you ignore our ever increasing cost of goods sold, we promise it’s not a pyramid scheme” - ~~Charles Ponzi~~ Dario


But the reason the cost is ever increasing is the steeply ascending quality. There's no (non-doom) scenario where this game does not end with profitable cheap tokens. Wherever it'll stop scaling, it'll be profitable to serve tokens at that scale. And of course if it doesn't stop scaling, it's the singularity and all this thought of profits is quite secondary.

An industry cannot fail because it's too successful.

edit: Other explanation. The industry has found a moneyprinter. It prints money, and several companies are building them and printing money. However, each one of those companies is cashflow negative, because they're all trying their desperate best to build an even larger moneyprinter. Why? If they don't build it and the next company over does build it, and it works, then company 1 is differentially outcompeted. what's more, the more expensive each moneyprinter is, once a company gets strongly ahead the second company won't get the financing to keep up with that one. ... But it's still moneyprinters. This is not an economic race that ends up with "and thus everyone stopped printing money."


> But the reason the cost is ever increasing is the steeply ascending quality.

Sure, if you say so.

> There's no (non-doom) scenario where this game does not end with profitable cheap tokens.

I doubt it.

> Wherever it'll stop scaling, it'll be profitable to serve tokens at that scale.

It’s much more likely, imo, that local AI models will completely destroy the entire business model of inference. For any entity that would be spending enough to replace multiple employees (allegedly), the cost of the hardware to do so is completely affordable. This has the major benefit of eliminating per token costs, and practically guaranteeing that some “AI company” can’t steal your company’s secrets.

At an individual person level, this hardware will take much longer to obtain, but it will happen eventually.

> And of course if it doesn't stop scaling, it's the singularity and all this thought of profits is quite secondary.

I still believe, strongly, that LLMs are the wrong path to AGI. There’s trillions of dollars being put into turning LLMs into AGI; however, there has no zero results showing that LLMs are capable of “learning” anything. “Skills” and expanded context windows don’t count. The “model” needs to be able to modify itself, in real time, for that to be possible.

We’re in a local maxima that strongly refuses to accept that it’s a local maxima. (Diminishing returns for the amount of money poured into “training” is beyond obvious to anyone paying attention).

> An industry cannot fail because it's too successful.

Sure it can. The fossil fuel industry is wildly successful.

It’s also boiling the planet alive, but that’s okay, it’s generating shareholder value so everything’s fine.

Edit for your edit:

> edit: Other explanation. The industry has found a moneyprinter. It prints money, and several companies are building them and printing money. However, each one of those companies is cashflow negative, because they're all trying their desperate best to build an even larger moneyprinter. Why? If they don't build it and the next company over does build it, and it works, then company 1 is differentially outcompeted. what's more, the more expensive each moneyprinter is, once a company gets strongly ahead the second company won't get the financing to keep up with that one. ... But it's still moneyprinters. This is not an economic race that ends up with "and thus everyone stopped printing money."

By your own definition it is impossible to achieve profitability without an ever increasing amount of money flowing into it.

There was a business about moving stamps around that needed similar levels of cash infusion to sustain itself. Eventually its bubble popped too.


> It’s much more likely, imo, that local AI models will completely destroy the entire business model of inference.

You're massively underestimating how much cheaper it is to serve models in a datacenter than at home.

> By your own definition it is impossible to achieve profitability without an ever increasing amount of money flowing into it.

No, what? No. This is just a total misreading of what I'm saying. It's extremely easy to achieve profitability. It just requires your business to stop chasing more profitability. The problem is just that then the other businesses who are still chasing more profitability will eat its lunch. It's a market shape that easily supports n-1 bankruptcies but never n bankruptcies.


> Diminishing returns for the amount of money poured into “training” is beyond obvious to anyone paying attention

Sure, if you say so.


And that interview was explicitly a hypothetical. It was shared here a few months back.

Dario also believes that verifiability is trivial, which I'm much more sceptical of.


it doesn't seem that complicated to me. what do you think the cost is for AWS to serve inference it sells to Anthropic for "$1.00"? I think it's about $0.11. Then, ask yourself, did Amazon sell it to Anthropic at all? No, they traded $0.11 real dollars for $1.00 of Anthropic shares (the CLN that created the real loss Anthropic is talking about), which are now worth $10.00 (which is why people talk about this as a paper loss). So... IS it profitable? It's profitable, but Amazon didn't JUST sell inference when they turned $0.11 into $10.00...


That's a Ponzi scheme.


Nope, the leak didn't show that.


in some kind of boring sense, sure.

but have you looked up who invented the computer? and who was funding all the first versions of the software we use? it was the US military.


Where do you think the internet came from?


Neither of those things. Arpanet was military and the web came out of CERN, but they were both low budget side projects.


...and global usage/reach was practically zero...


Why people don't treat categorizing model activations as a science is beyond me.


Mechanistic interpretability is a field that studies exactly this.


Please, go on. How would you do it, specifically?


They do, there was a paper that claimed a major breakthrough in that field just a few days ago. See the "Neural networks are clean" discussion:

https://news.ycombinator.com/item?id=50105647


Are you saying no one is trying to do exactly that?


Kind of disturbing how hard he's leaning into the futurism aesthetic here, given its history.


Tesla has been doing this for a while, having all their promo shots at night etc. The terafab announcement stuff has an especially dystopian vibe


Guidelines | FAQ | Lists | API | Security | DMCA | Apply to YC | Contact

Search: