Maybe even 32GB if they use newer ICs.
More explanation (and my source of the tip): https://www.pcgamer.com/hardware/graphics-cards/shipping-document-suggests-that-a-24-gb-version-of-intels-arc-b580-graphics-card-could-be-heading-to-market-though-not-for-gaming/
Would be awesome if true, and if it's affordable. Screw Nvidia (and, inexplicably, AMD) for their VRAM gouging.
They would be incredible, as long as they're cheap. Intel would be utterly stupid for not doing this.
OpenAI does not make hardware. Also, their model progress has stagnated, already matched or surpassed by Google, Claude, and even some open source chinese models trained on far fewer GPUs... OpenAI is circling the drain, forget them.
The only "real" competitor to Nvidia, IMO, is Cerebras, which has a decent shot due to a silicon strategy Nvidia simply does not have.
The AMD MI300X is actually already "better" than Nvidia's best... but they just can't stop shooting themselves in the foot, as AMD does. Google TPUs are good, but google only, like Amazon's hardware. I am not impressed with Groq or Tenstorrent.
Yeah, I didn't mean to imply that either. I meant to write OneAPI. :D
It's just that I'm afraid Nvidia get the same point as raspberry pies where even if there's better hardware out there people still buy raspberry pies due to available software and hardware accessories. Which loops back to new software and hardware being aimed at raspberry pies due to the larger market share. And then it loops.
Now if someone gets a CUDA competitor going that runs equally well on Nvidia, AMD and Intel GPUs and becomes efficient and fast enough to break that kind of self-strengthening loop before it's too late then I don't care if it's AMDs ROCm or Intels OneAPI. I just hope it happens before it's too late.
They’re kinda already there :(. Maybe even worse than raspberry pies.
Intel has all but said they’re exiting the training/research market.
AMD has great hardware, but the MI300X is not gaining traction due to a lack of “grassroots” software support, and they were too stupid to undercut Nvidia and sell high vram 7900s to devs, or to even prioritize its support in rocm. Same with their APUs. For all the marketing, they just didn’t prioritize getting them runnable with LLM software