Haider.(@haider1):pretty crazy that AI agents are now helping researchers optimize inference down to the GPU kernel level... combining MXFP6 quantization, custom matmul kernels and optimized inter-GPU communication helped cut Krea 3 image editing latency from 22s to 3.6s on the same hardware in under a week

2026.11.10 17:56

pretty crazy that AI agents are now helping researchers optimize inference down to the GPU kernel level... combining MXFP6 quantization, custom matmul kernels and optimized inter-GPU communication helped cut Krea 3 image editing latency from 22s to 3.6s on the same hardware in under a week

Hedra@hedra_labs

2026.11.10 17:30

Introducing Hedra Model Factory: an autonomous inference and training platform for building models for the real world. The recursive self-improvement loop behind frontier research has been locked inside a few labs. We're opening it to teams in robotics, manufacturing, biology and defense. Our researchers come from Stanford, NVIDIA and Meta Superintelligence Labs. Now we work alongside your team to post-train and optimize your models.