AMD acquires Taalas to boost inference performance by etching models in silicon
1–10 of 709 posts
Re: AMD acquires Taalas to boost inference performance by etching models in silicon
#2Really hoped to see their hw out in the wild one day
Re: AMD acquires Taalas to boost inference performance by etching models in silicon
#3Didn't even give them a chance to launch the hardware.
Re: AMD acquires Taalas to boost inference performance by etching models in silicon
#4While this design is self-limiting I think its a good approach. It doesn't take an entirely new architecture or infinite memory to produce significant performance improvement.
Re: AMD acquires Taalas to boost inference performance by etching models in silicon
#5Well so much for that dream.
Guess we can look forward to picking these up ex-enterprise on ebay for under $5k a pop in a decade or two
Re: AMD acquires Taalas to boost inference performance by etching models in silicon
#6The demo: https://chatjimmy.ai/
Re: AMD acquires Taalas to boost inference performance by etching models in silicon
#7This is probably a win-win. The team gets paid, and we get greater assurance that their best ideas and architectures -- which are truly impressive -- are going to see the light of day in actual products.
Re: AMD acquires Taalas to boost inference performance by etching models in silicon
#8This is probably a win-win. The team gets paid, and we get greater assurance that their best ideas and architectures -- which are truly impressive -- are going to see the light of day in actual products.
They were too small for this to be a meaningfully sized purchase for AMD, there's real risk they get sucked into a team that ultimately delivers sqat, not to mention the chances of anything being delivered in an even remotely consumer-priced bracket are definitely out the window
Re: AMD acquires Taalas to boost inference performance by etching models in silicon
#9so qwen3.x-27b on hardware? or better deepseek-v4-flash on hardware .
Re: AMD acquires Taalas to boost inference performance by etching models in silicon
#10so qwen3.x-27b on hardware? or better deepseek-v4-flash on hardware .
I wrote them an email asking for PrismML Bonsai 27b Ternary which is like 6b or something crazy small and would be a lot easier for them to do initially.