In this episode, we sit down with Oli and Oscar from Red Team, the BitTensor subnet tackling one of GenAI’s biggest threats: synthetic fraud.

Their London-based cybersecurity firm Innerworks already protects major payment platforms and crypto exchanges. But now, with Subnet 61, they’ve built a decentralized R&D engine that uses BitTensor miners to stress-test their own defences - crowdsourcing attacks to stay ahead of adversaries.

We dig into:

• What Red Team actually does (in human language)
• Why AI-powered fraud is exploding - and how SN61 spots it in real time
• The genius way they use miners to simulate attacks before they hit the wild
• How they price and sell the product as a high-margin SaaS play
• Whether this is an “extractive” subnet or not

In this Subnet Session, we’re joined by Mike Bunting - CFO of Macrocosmos and architect of Subnet 13: Data Universe, the largest open-source social media dataset on the planet.

We explore:

• Why they’ve already scraped over 55 billion posts (and counting)
• How SN13 turns raw social data into revenue, insight, and alpha
• Who’s buying the data and why it matters for hedge funds, consumer brands, and subnets alike
• The future of metadata, visualisation, and turning big data into usable products
• How Subnet 13 could power everything from brand tracking to AI training to prediction markets

Whether you’re staking, building, or just curious about real-world utility in Bittensor - this one’s a must-listen.

In this Subnet Session, we’re joined by Wandering Weights - aka Chris - the quietly brilliant mind behind Gradients (SN56).

We talk about how Gradients is flipping the post-training market on its head:

• $250 fine-tuned LLMs that rival $10K AWS jobs
• A roadmap to 5.0 and enterprise-grade privacy
• Why replacing 5 ML engineers with one isn’t dystopian - it’s profitable
• And how 3,000 early users are already generating real-world revenue

If you’ve ever trained a model, hired an ML team, or cursed at AutoML’s pricing - this one’s for you.