← All episodes

This Bittensor Subnet Could Cut Drug Discovery Costs in HALF | E2267

| 5 products mentioned
Watch on YouTube decentralized ai cryptocurrency incentives drug discovery vision models creator economy ai agents machine learning competition

This episode explores three different Bittensor subnets that demonstrate the versatility of decentralized AI incentive networks for solving real-world problems. Hosts Alex and Lon interview founders from MetaNova (drug discovery), Bitcast (creator marketing), and Score (vision AI), showing how crypto economics can reward contributors building AI models, compute, and intelligence across entirely different domains. The episode reveals how decentralized competition is accelerating progress in expensive, complex fields like pharmaceutical development while democratizing access to specialized AI tools.

Key takeaways
  • MetaNova is using Bittensor to decentralize drug discovery by having miners compete to find promising molecules from a searchable universe of ~65 billion possibilities, potentially cutting the current 10-year, $2.6 billion drug development timeline in half through improved virtual screening.
  • The Bittensor economic model rewards miners and validators based on the quality of their contributions—whether molecules, compute, or ranking decisions—while stakers vote with their Tao tokens to allocate emissions to subnets they believe in, creating a marketplace for intelligence production.
  • Bitcast removes administrative friction from creator marketing by automating brand briefs and compensating YouTubers based on watch time rather than fixed fees, allowing brands to activate hundreds of mid-tier creators simultaneously instead of negotiating with a few top-tier influencers.
  • Score (Subnet 44) distills large vision language models into tiny, specialized models (50MB instead of 3.4GB) that run on consumer CPUs, making computer vision accessible to businesses that can't afford H100 GPUs while maintaining accuracy for specific tasks like detecting gas pump collisions.
  • Miners across all three subnets benefit from adversarial dynamics—trying to exploit or "game" the system reveals weaknesses in models and incentive mechanisms, which operators can then fix, creating a feedback loop that strengthens the underlying technology.
  • The agentic economy is enabling new use cases where AI agents autonomously monitor vision streams, generate content, and flag anomalies, with Bittensor subnets providing the intelligence layer that agents depend on to operate effectively.

Recommendations (2)

OpenClaw
OpenClaw uses

"I still enjoy OpenClaw. I've bonded with my agent, but a lot of the OpenClaw projects are kind of similar."

This Week in Startups · ▶ 0:52

Claude
Claude uses

"I want them to fire up Claude code and have it build them an app and then run it because I think if you once you discover that that's possible then AI does feel like the old bicycle for your mind t..."

This Week in Startups · ▶ 54:07

Mentioned (3)

ChatGPT
ChatGPT "The average person has asked ChatGPT to write them a poem four times at this point in time. And t..." ▶ 53:47
Auto researcher "I was thinking about some combination of auto researcher from Andre Carpathy, my local OpenClaw s..." ▶ 29:32
Cursor
Cursor "If you're not coding so you're not using Claude code or you're not on Cursor or whatever." ▶ 53:31