Kenya’s small and mid-sized chemists are moving from paper to data-first operations. Zendawa’s end-to-end platform shows how AI can streamline inventory, unlock credit, and localize delivery while reinforcing privacy and human oversight.
Multiple cofounders and senior staff are leaving Thinking Machines Lab to return to OpenAI and Meta, escalating the race for elite AI talent and raising questions about retention at well‑funded startups.
Thinking Machines Lab cofounders return to OpenAI Read Post »
MIT has expanded and renamed its central research effort on understanding natural and artificial intelligence, pairing long-horizon science with shared infrastructure and industry collaboration to accelerate real-world impact.
Hierarchical tiling—block, warp, and register—turns GPU memory hierarchy into a performance advantage for GEMM. Here’s how practitioners squeeze more throughput on modern NVIDIA hardware and when to rely on vendor libraries.
CUDA Tiled Matrix Multiply: High-Performance Guide & cuBLAS Tips Read Post »
MIT researchers are pairing generative design with materials and sensing advances to make personalized 3D-printed objects that keep their fit, function, and reliability in real-world use. The approach centers on Style2Fab for safe restyling, a dual-mode resin for complex, durable parts with dissolvable supports, and embedded sensing for interactive devices.
AI-assisted 3D printing personalization for durable, everyday products Read Post »
AI is shifting from target to tool for attackers while cloud-native lifecycles expand the attack surface. Here’s what security leaders need to know—and do—before model compromises scale across critical systems.
AI Model Security Risks: Why We’re at a Hacking Inflection Point Read Post »
