Bittensor Guru S2E9 - Subnet 44 Score w/ Max
May 19, 2025
Max discusses cutting-edge advancements in computer vision and AI for sports analytics, emphasizing self-supervised learning and data challenges. He highlights the role of community engagement and innovative mining incentives within the Bittensor ecosystem. The conversation also touches on improving quality through human feedback and the importance of collaboration among miners and subnet owners. Lastly, they explore the intricacies of financial strategies and the need for ethical practices in maintaining system integrity.
AI Snips
Chapters
Transcript
Episode notes
Vision Is Key for Human-Level AI
- Vision is critical for AI to reach human-level understanding, as babies ingest more visual data than text models acquire.
- Sports offers a complex yet practical use case to develop resilient self-supervised computer vision models on BitTensor.
Incentivizing Fast, Accurate Vision
- Score incentivizes miners to autonomously reconstruct sports footage frame by frame, emphasizing speed and accuracy.
- The subnet achieves annotation speeds vastly faster than humans with near expert-level accuracy, critical for live data streams.
Complete Sports Analytics Pipeline
- Score’s workflow includes data annotation, simulation on the extracted data, and attributing frame-by-frame impact scores.
- This comprehensive approach gives clients superior insights, like player contributions and injury prevention data.