The Data Foundry for AI with Alexandr Wang from Scale
No Priors: Artificial Intelligence | Technology | Startups
Limited Generality and Lack of Positive Transfer in Multi-Step Reasoning Models
Multi-step reasoning models like Monte Carlo tree search may not offer sufficient generality in solving multiple problems, leading to scaling challenges. Limited positive transfer exists between learning in different modalities, such as video versus text, indicating that training in one modality does not significantly help in solving problems in another. These constraints suggest a lack of broad applicability and transferability in multi-step reasoning models.
00:00
Transcript
Play full episode
Remember Everything You Learn from Podcasts
Save insights instantly, chat with episodes, and build lasting knowledge - all powered by AI.
