Microsoft’s Magnetar Turns Hash-Only Proving Into a Laptop Benchmark Win, but Recursion Is Still Unimplemented
Microsoft’s Magnetar paper supports the blog’s speedup claim on...
Browse Tag
Microsoft Research
13 Articles
Microsoft’s StudentSim Turns Tutor Testing Into a Fidelity Question, With Guidance Graded Against an Answer Key
Microsoft Research’s StudentSim trains one simulated student per...
Microsoft’s Agent Lightning v1.0 Turns Agent Training Into a Sample-Accounting Problem
Microsoft’s rebuilt Agent Lightning trains agents through an...
Microsoft’s Physical AI Toolchain Turns Robot Battery Life Into a Bandwidth Problem
A new Microsoft Research study shows that fixing a robot's dead...
Microsoft’s Skala Turns DFT Accuracy Into a Software Distribution Problem
Microsoft Research's Skala 1.1 pushes its deep-learning DFT...
Microsoft’s Orchard Turns Agentic AI Training Into an Infrastructure Problem
Microsoft Research's Orchard framework treats AI agent sandboxes...
Microsoft’s Echoverse Turns Computer-Use Agent Training Into a Depth-Over-Scale Problem
Microsoft Research built twelve synthetic worlds to train...
Microsoft’s SymCrypt Proofs Turn Verified Cryptography Into Routine Engineering
Microsoft has open sourced proofs that its Rust rewrite of SHA-3...
Microsoft’s Flint Shows Why AI-Generated Charts Need a Semantic Compiler
Microsoft Research's new Flint language and its MCP server let...
Microsoft’s SkillOpt Makes Agent Skills a Training Target
Microsoft Research’s SkillOpt trains natural-language agent...