Episode Details

Table of Contents

• Memory Management Innovations
  • Jemalloc 5.4.0
  • Speeding up gearhash on ARM64
• AI and Law
  • Astra for Law
  • Bend – A language that blocks AI mistakes via proof, on CPU and GPU
• Next-Gen AI Models
  • Bonsai 2 27B: Near-Lossless Compression in a 9x Smaller Footprint
  • Shapelearn Qwen 3.8 27B (13.1 GB VRAM)
• Security Concerns
  • A heap overflow and SSO misconfiguration to compromise OpenAI internal repos
  • Hister: A private search engine for the pages you visit and the files you keep
• New Tools and Platforms
  • OpenJev
  • Ask A Monk – A digital wilderness for thoughts with no immediate answer