When Crypto Becomes Intelligent
summary
Argues that crypto will shift from automation to native intelligence, citing IntoTheBlock’s own ML-driven quant strategies that attracted over a billion dollars in allocations; highlights applicable tools—transformers and graph neural networks, plus self‑supervised and reinforcement learning—and concrete builds: NFTs with vision+language agents, AMMs replacing X+Y=K with AMMs that learn market regimes, programmable stablecoins that adapt peg policy, DeFi credit scoring from on‑chain behavior, and practical barriers like costly on‑chain training and proposals to publish gradients or use federated/off‑chain training.
AI-written summary · about
chapters (14)
- 0:00Introduction
- 0:41Jesus Rodriguez Introduction
- 1:32IntoTheBlock Analytics Platform
- 2:02Can Crypto Become Intelligent
- 3:03Machine Learning is Software
- 6:06Machine Learning in Crypto
- 7:26Single Transformers
- 8:33Selfsupervised learning
- 10:16Graph Neural Network
- 11:07Future of Crypto
- 11:51Stupid Ideas
- 14:45Machine Learning
- 16:33Decentralized ML
- 17:52Summary
related talks
16mTokenization and Agents: Why It Matters
34mCrypto narratives 2025 and new cycle
27mAI to Stablecoin - Blockchain for Mainstream
20mGaming, Prediction Markets & AI: What’s Next for Blockchain
39mSuccessful AI driven startups
32mProgress In the Decentralized AI Stack
31mHacking Success in The Age of AI & Crypto
45mCrypto-AI: A Guide to Building Private and Decentralized AI Systems