Blockchain-based Machine Learning Marketplaces
Fred Ehrsam

Top notch!

We’re in the age of AI for the commons: it becomes a public utility and hence drives down the marginal cost as the market opens.

  • pool-trained: bulk up from massive contributions (with attribution for the entire AI supply chain and privacy-preserving where needed — as you described)
  • forked: fine-tuned to personal benefits. Fork off massive contribution power (~ hashing power) and create your personal AI with adaptive learning

Next up:

  • MPC/FHE/PCP circuit design targeting machine learning algorithms
  • a hardware market to support this revolution (secure enclaves, hardwired personal {C,R}NNs, personal sensor spheres, …)
  • conscious incentive design for AI
  • model the state of the world resource and demand and make our future generations self-sustaining