Lilian Weng
Person on 3 sites
Lilian Weng is a machine-learning researcher known for Lil’Log, the blog of long survey posts on machine-learning techniques she has written from 2017.
-
4 facts on wiqqi (opens wiqqi.app)
- Lil’Log: Blog of survey posts, from 2017 (opens wiqqi.app)
- OpenAI: 2018 to November 2024: robotics, applied AI research (2021), head of safety systems (2023), VP of research and safety (August 2024) (opens wiqqi.app)
- Thinking Machines Lab: Founding team member, February 2025 to July 2026 (opens wiqqi.app)
-
3 channels on ppll (opens ppll.app)
-
x.com: @lilianweng (opens x.com)
“It is a hard and sad decision. I shared this message with folks at Thinky. Thank you all for the time together♥️ Just as the last sentence in my message: The future worth building is human.”
-
GitHub: @lilianweng (opens github.com)
“lilianweng.github.io — My personal page”
-
Blog: Lil'Log (opens lilianweng.github.io)
“Harness Engineering for Self-Improvement”
-
x.com: @lilianweng (opens x.com)
-
16 beliefs on korrents (opens korrents.com)
-
The deployment layer surrounding a model is as important as the model's raw intelligence. (opens korrents.com)
“the layer between the raw model and the real-world context seems to be as important as the model's raw intelligence”
-
Recursive self-improvement only works if the base model is already capable enough to improve its own mechanism. (opens korrents.com)
“Recursive structure alone is not enough. The base model must be capable enough to improve the mechanism.”
-
Harness improvements help deploy a model better, but the model's underlying intelligence remains the core driver of capability. (opens korrents.com)
“harness improvement enables better deployment of the model but intelligence is still the core.”
-
The deployment layer surrounding a model is as important as the model's raw intelligence. (opens korrents.com)