Bring your task, your latency budget, and your hardware ceiling. We'll tell you honestly whether a Seed shape fits — and if it does, what accuracy you can expect.
All three land in the same inbox. Tell us which one you're sending so we can route it to the right human and reply with something useful.
You're building a humanoid platform, a robotics product, or an embedded ML pipeline and need a specific Seed shape that fits your latency, memory and accuracy budget.
Tell us your budgetQuestions about the Banking77 result, the protocol, or how the frontier moves on your task. We'll answer anything we can share without disclosing the methodology.
Ask a questionCoverage, interviews, or benchmark follow-up. We prefer specific, technical questions to general briefings — send them over and we'll reply fast.
Start a threadIf you think your task might fit the Seed shape envelope — even if you're not sure — tell us. The worst answer we can give is “not yet.”