Rendered at 18:02:11 GMT+0000 (Coordinated Universal Time) with Cloudflare Workers.
0-_-0 3 hours ago [-]
If AI models were trained for optimising kernels, they could easily replace ML compilers. With no branches a numerical check gives you the correctness guarantees you need. However we already see anthropic deliberately nerfing kernel optimisation capabilities of their latest models, Chinese AI companies on the other hand have the opposite incentive.
augment_me 2 hours ago [-]
Really? Can you give source for kernel generation being nerfed?
> However, during the model’s training, we took steps to limit advances in the specific research capabilities that could most directly accelerate AI development. Thus, in relevant areas, including accelerator kernel optimization and improving large language model architectures, Fable 5.2 and Claude Mythos 5.2 perform at about the level of Opus 5.5.
> Fable 5.2 and Claude Mythos 5.2 also include transparent fallbacks to Opus 5 for a small set of topics in this domain, including kernel development on certain nonstandard accelerators.
etskinner 2 hours ago [-]
As someone who doesn't know much about kernels and compilation, this reads like a modern-day Retroencabulator
gavmor 1 hours ago [-]
Yeah, I'm surprised I hadn't heard of kernel in this context until today.
> However, during the model’s training, we took steps to limit advances in the specific research capabilities that could most directly accelerate AI development. Thus, in relevant areas, including accelerator kernel optimization and improving large language model architectures, Fable 5.2 and Claude Mythos 5.2 perform at about the level of Opus 5.5.
> Fable 5.2 and Claude Mythos 5.2 also include transparent fallbacks to Opus 5 for a small set of topics in this domain, including kernel development on certain nonstandard accelerators.