had just shipped a repo for training speculative decoding heads to speed up inference of llms by ~3x.
get any base model, train a few speculative heads, see the difference in throughput.
🧵on more details.
1/n
Sandboxes are all the rage (Modal, E2B, AWS, ..). Most AI teams pay a >4x markup to run sandboxes on someone else's machines.
Introducing SkyPilot Sandboxes — Run BYOC sandboxes on your own clusters.
• 50,000+ sandboxes on a single cluster
• Sub-second launches with warm pools
• Great for RL rollout (keep sandbox clusters close to GPUs)
Benchmark shows @skypilot_org Sandboxes are 4-10x cheaper than Modal at lower latency. Full results in blog.
🚀 We're excited to be a Day-0 launch partner for NVIDIA Nemotron 3 Ultra.
Deploy NVIDIA's latest open model for agentic AI on Simplismart
Our optimizations we deliver higher throughput than TensorRT + MTP + NVFP4.
Read more: https://t.co/bfi7d7jroC
#NVIDIA#Nemotron
nothing. i repeat nothing will stay the same as we know. think about this - products like salesforce, etc are used by a very minute margin of people and has a market cap of $150B, think of all the experiences that could be possible through innovations like this. endless honestly. gaming and content will converge, and world models will give us the claude-code type growth but for the rest of the distribution curve.
i dont think we know whats coming, honeslty nobody does. we are histroy in the making damn.
Excited to share that we’ve raised $300M in our Series B round, led by @radicalvcfund, bringing our total funding to more than $450M, with leading technology companies joining as both customers and investors.
We continue accelerating the path to AGI through our two core pillars: ultra-optimized infrastructure for AI workloads, and realtime world models built on top of it.
Today, we’re also launching DOS (Decart Optimization Stack) 2.0 – our next-generation inference and training platform, delivering over 1,600 tokens per second for agentic inference and over 100 FPS for world models across major hardware platforms.
Alongside DOS, we will launch new versions of our world models in the coming weeks: Lucy, for immersive realtime experiences, and Oasis, for physical AI.
Grateful to our partners, customers, and backers across media & entertainment, Physical AI, chips, hyperscalers, and the broader AI ecosystem.
@radicalvcfund, @Adobe, @alphaptrs, @amazon and @awscloud, @Atreidesmgmt, @benchmark, @eBay, @nvidia, @sequoia, @Toyota, @valor, @orenzeev; Andrej Karpathy, Michael Eisner, Yamauchi-No.10 Family Office, Moritz Baier-Lentz and more.
We made a film for this moment with Decart CEO @DLeitersdorf and Moritz Baier-Lentz - a look at the company, the technology, and what comes next.
i think it was @Suhail's post recently, where he said a good vector to have your company in is: is every new release by big token is a net scare or insane boost of your product/ service offering.
just saw claude agents view and it seems like it's over for companies like conductor etc. you just cannot exist in the path of big token. time to get back to hard engineering problems.
@andrewchen andrew, building an ai native services firm to help companies go ai native. think token factories for smbs. team is ex mastercard chief of staff, and A tier engineers from banglore.
This is rapidly becoming the greatest product demo since Steve Jobs’ “one more thing”.
Congratulations folks, you have just exited the smartphone era. Welcome to the robot era.
May you live in interesting times.