AINewsnow

I had an AI agent port Karpathy's GPT-2 trainer from CUDA to SYCL, and made it prove every step

A lot of AI software only runs on NVIDIA GPUs, because it's written in CUDA. SYCL is an open standard (from Khronos, the group behind OpenGL and Vulkan) that lets the same C++ code run on NVIDIA, AMD, Intel and plain CPUs. Moving code from CUDA to SYCL is slow, careful work. AI coding agents are no…

Read the full story at DEV Community — AI ↗

Timeline · 1 report

  1. 2026-11-06 20:48 · DEV Community — AI
    I had an AI agent port Karpathy's GPT-2 trainer from CUDA to SYCL, and made it prove every step

More stories

  1. Chinese Firms Cleared to Buy Nvidia H200 Chips Despite Empty Offices and Sanctions Ties — New York Times AI
  2. Anthropic launches Claude Personal Finance in the US with read-only bank access — Engadget
  3. Nvidia CEO Jensen Huang faces congressional briefing request over alleged Chinese spy ties — Axios AI+
  4. Man sues OpenAI, alleging ChatGPT helped his wife plan to poison him — Bloomberg AI
  5. OpenAI launches GPT-6 Cyber, a security model that found 81 Firefox flaws and 948 Linux bugs — VentureBeat AI
  6. DeepSeek V4.1 Flash gains traction across local rigs, benchmarks, and serving platforms — Vercel Blog
  7. Microsoft winding down deal to rent AI cloud capacity from Oracle — Bloomberg AI
  8. Nasdaq Suffers Worst Day in 18 Months as Nvidia Slides on AI Chip Selloff — The Globe and Mail

Get the daily brief of stories like this at 6:30 every morning →