AI Briefing#pallas
The latest AI and developer news about #pallas, with the original source and a short summary.
The latest AI and developer news about #pallas, with the original source and a short summary.
Visko's Orbis model achieves 1.07x playback speed on a single Cloud TPU slice using JAX and custom kernels.
Google has released Gemini-based TPU development agent tools that assist with PyTorch to JAX conversion and kernel optimization.
Helion, a high-level DSL that lets developers write high-performance TPU kernels in PyTorch style, has been released.
The article argues that portability in AI kernels is a myth, and top performance comes from hardware-specific DSLs.