5 ms·
I wrote elsewhere but I’m more interpreting this distinction as “RL in real-time” vs “RL beforehand”.
by anomaloustho 11mo ago
I wrote elsewhere but I’m more interpreting this distinction as “RL in real-time” vs “RL beforehand”.
- munchler 11mo agoI agree with this description, but I'm not sure we really want our AI agents evolving in real time as they gain experience. Having a static model that is thoroughly tested before deployment seems much safer.
- mbesto 11mo ago> Having a static model that is thoroughly tested before deployment seems much safer. While that might true, it fundamentally means it's not going to ever replicate human or provide super intelligence.
- CryptoBanker 11mo ago> While that might true, it fundamentally means it's not going to ever replicate human or provide super intelligence. Many people would argue that's a good thing
- stevenpetryk 11mo agoThis is referred to as “online reinforcement learning” and is already something done by, for example Cursor for their tab prediction model. https://cursor.com/blog/tab-rl https://cursor.com/blog/tab-rl
- tinodb 11mo agoNot sure that’s the same. They just very frequently retrain and “deploy a new model”.