Hacker Timesnew | past | comments | ask | show | jobs | submit | XCSme's commentslogin

Good point, everyone was expecting GPU prices to go down and all crypto bros to drop their mining rigs, but they can just repurpose them now...

I thought anyone serious about mining switched to ASICs and custom hardware a while ago. At least for BTC and maybe eth, a custom asic setup is ~2-3x more profitable

Yep. You've got it right.

BTC has been an ASIC game for a very long time. GPUs haven't been profitable there since the ASICs showed up a dozen or so years ago (with odd exceptions where power is "free").

Eth kept hitting the GPU market hard until ~4 years ago, when the network switched from proof-of-work to proof-of-stake. That cut the home-gamer hobbyist miners out rather completely.

(That last one kind of sucked for me. When my office-room had resistive heating, I rather liked getting paid for my otherwise-idle GPU to make heat for me in the winter time. It wasn't much, but >0 is more than <0.)


Ethereum switched to proof of stake in 2022. The cryptocurrency inflicted GPU shortage began in 2017. You're talking about an event that is basically a decade ago at this point.

It also does A LOT better, for my hamster test: https://aibenchy.com/showcase/?q=claude#showcase=6efb87c28e3...


> We recently submitted a confidential S-1. We expect it to leak so we’re just announcing it.

What?


They expect someone to leak that they had submitted it, so they’re just saying it themselves. I don’t think they mean that the actual contents (like financial projections and all that) will be leaked.

Narcissist marketing, Sam loves it.

I like how the Gemini 3.5 Flash (medium) one[0] added a ladder, so the hamster can get on the table.

[0]: https://aibenchy.com/showcase/#showcase=eb4878dbff331c67


Not sure how I feel about Grok... [0]

[0]: https://aibenchy.com/showcase/?q=grok


I use V4 Flash as my main model, it really is exceptionally capable for the price.

It can follow instructions quite well, if you have a plan, it usually executes it to completion and you get code that works.

My only concern with it, is that it's mostly served by Chinese companies on OpenRouter.


Svgs are just shapes, they (models) mostly only use sqaures, circles, etc. Rarely paths.

Plus most models are multi-modal, so they have a visual understanding too.


Right, but rendering the visual understanding is based on neighboring pixels normally, which is much more directly related to the visual than plaintext like SVG tags.

I kind of disagree, because describing an image is more like writing an svg than rendering pixels.

If I ask you to describe how would you draw a cat head, you could do it in text like: "a big circle as the head, 2 small circles as the eyes, 2 triangles for ears, 3 lines on each side of the mouth as moustaches, etc..."


Hmm, sure. I'm still surprised - it also has to say where they are in coordinate space. It feels like the way genAI works ("what comes next") is not amenable to this use case (demonstrably I'm wrong, of course).

I remember when I was learning WebGL/OpenGL and had to draw some test shapes to test my shader, I would manually think what the vertex positions should be to draw a triangle, a pyramid, etc. I think for AIs it's quite easy, because most are quite decent at math, and have been trained on many geometry problems, and probably also a lot of OpenGL code too, and 3D assets.

Sorry, accidentally posted the Romanian version, lol, here is en: https://aibenchy.com/showcase/hamster-playing-table-tennis-s...

If mods could update the link please.


If I were English, I would not love the US flag beside "English".

Well, it is US English I think, I could add UK English too.

If you were a former colony, it would be kinda funny ;)

Not if the original code is secure...

Any reason why they are not on openrouter yet, but the speech models are? https://openrouter.ai/models

Curious to test them and see how they perform.


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

Search: