Video wird geladen...
Video konnte nicht geladen werden
Our heterogeneous cloud just got better. We're announcing a cluster scale deployment of Positron AI as our next partner. We believe demand for fast inference will outstrip supply for the next decade. Our bet on Positron is another step towards getting fast inference for opensource models into builders' hands.... show more
467,236 Aufrufe • vor 28 Tagen •via X (Twitter)
44 Kommentare

@positron_ai The building will continue unabated

@positron_ai

@positron_ai LFG ! Proud to be partnering with @mitesh711 and @trsohmers to get more fast inference into developers' + builders' hands.

@positron_ai Its all about the speed! Excited for this one

@positron_ai Fitting! ⚡️

@positron_ai AI agents are going to drive so much more inference demand.

@positron_ai Well That sounds very serious Congrats

@positron_ai This is how we push SI/AI forward 🔥🔥🔥

@positron_ai

@positron_ai Big step toward making fast inference more accessible. Exciting to see this partnership focused on open-source models!

@positron_ai wow, so many innovations cominng the inference space. prefill on GPUs and decode on memory-first chips is exactly what agent era needs

@positron_ai Open-source AI needs serious compute behind it. This partnership could unlock something huge..

@positron_ai This signals a major shift in AI infrastructure and compute needs. Very excited to see this partnership meet those demands. Also, @mitesh711 is killing it. :)

@positron_ai @mitesh711 @FPuklowski and @GoodisonJason are huge fans of @mitesh711 and the team! 🫶

@positron_ai Since saas is overloaded, probably inference is one of not many good options

@positron_ai This is an incredibly exciting collaboration. Congratulations to both @general_compute and @positron_ai! I'd love to experience this new cluster when it comes online. 🦾

@positron_ai LETS GOOOO!

@positron_ai Q1 2027 will be an interesting milestone for this partnership.

@positron_ai specialized chips for inference are definitely worth watching

@positron_ai Gotta go fast

@positron_ai Alright, LFG 🔥

@positron_ai The agentic economy is going to require so much more. LETS GO!!!

@positron_ai Great to see more partnerships focused on scaling open-source AI.

@positron_ai Quick question: do you have any benchmarks on how long it takes to move the KV cache to different hardware?

@positron_ai AI demand is not the question anymore The opportunity is in delivering more useful compute from every dollar of infrastructure already deployed

@positron_ai which open-source models are you guys putting on the cluster first

@positron_ai

@positron_ai We're stoked to bring @positron_ai on as a partner!

@positron_ai

@positron_ai Great to see another step toward faster AI model replies.

@positron_ai The one contract part matters too. Builders shouldn't have to manage two hardware setups just to split prefill and decode.

@positron_ai Very cool!

@positron_ai gemini’s long-context mode already separates prefill cost from decode in the API billing, so you can see the same split today without waiting for new chips. the difference is you still pay for the memory bandwidth either way

@positron_ai MORE INFERENCE!

@positron_ai can u pick what runs on positron or its all one endpoint

@positron_ai So the slow part was memory the whole time That changes how I think about inference speed

@positron_ai Fast inference is about to be the new GPU shortage. Smart move adding Positron

@positron_ai Totally agree. Silicon is going to be worth more than gold soon enough!

@positron_ai congrats on the partnership, this one makes a lot of sense

@positron_ai Thanks, we're stoked with this one too! @positron_ai 🙏

@positron_ai Love seeing more options for fast decode. Agents need this.

@positron_ai We're building an awesome selection of options!

@positron_ai This makes total sense to me as a developer

@positron_ai LFG! 🚀 More compute for open-source models means faster inference and more innovation for builders!
