Baidu says RSI auto-optimized ERNIE inference, 6.24x decode speedup on Kunxin P800
xiaohu · x · 2026-11-11
Baidu disclosed at Baidu World that it used Recursive Self-Improvement (RSI) for the first time to optimize ERNIE inference: engineers only define the modifiable scope, and the system autonomously locates bottlenecks, patches code, and runs tests, keeping changes only if they're faster with identical outputs.
Results: decode speed reached 6.24x the pre-optimization baseline on a Kunxin P800 low-latency test, and another GPU test processed 33.7% more tokens in the same time with consistent outputs.
More from Infra
- Three years building CPUs and spiking nets: one engineer's path to AI accelerators — k7agar · 2026-11-11
- CoreWeave: 2027 GPU renewals could add ~$4B annualized revenue, analyst skeptical — zephyr_z9 · 2026-11-11
- Mark Says George Hotz Was Prescient About Nvidia's Situation — mark_k · 2026-11-11
- Amazon Prime 360 uses AI to rebuild NBA plays in 3D for live, any-angle replays — import_jmr · 2026-11-11
- Altman says OpenAI aims to replace Nvidia in the medium to long term — zephyr_z9 · 2026-11-11
- Chinese University Bought Nvidia H200 Chips Despite US Blacklist — pstAsiatech · 2026-11-11