News
Launches, research notes and company announcements from the team building the training stack.
What we learned topping NVIDIA's SOL-ExecBench
We're #1 on SOL-ExecBench across all four tracks and learned a lot of bitter lessons about frontier models, reward hacks and why the benchmark is running out of time
Read MoreSolving the Bottleneck to Enterprise AI Sovereignty
The benefits of AI are dangerously concentrated. Changing that is the single biggest untapped lever for economic growth this decade.
Proving Kernels Correct Instead of Testing Them
We formally prove optimized ML kernels correct instead of testing them: symbolic execution shows a FlashAttention tile is equivalent to naive attention at d=64
STFS: Stop Making Every Worker Hammer S3
STFS turns repeated ML training reads into cluster-local cache hits instead of repeated S3 traffic: 97% less origin load, sub-200ms p95 reads after the first epoch.
Tensor Cloud Launch
We're launching the private beta of Tensor Cloud: a cloud-agnostic interface that runs jobs across providers without the messy details of each one.
Why We're Building SF Tensor
At the San Francisco Tensor Company, we believe that the next great wave of computing won't be defined by a single chip, a single cloud or a single framework.
1–6 of 6 Blog Results