Episode Summary
Executive Summary: The episode centers on BitTensor as a decentralized AI/crypto network that incentivizes miners to build and run AI products, plus how agents like OpenClaw can interact with it. Jason and Mark argue BitTensor’s subnet model can create cheaper, competitive AI infrastructure, while Ala and Michael explain wallets, allocators, and agent tooling that make participation easier. The show also includes startup demos, media recommendations, and consumer product ads.
Main Topics: BitTensor as decentralized AI infrastructure (Priority: 5/5): Mark Jeffrey explains BitTensor/TAO as a programmable mining network that subsidizes AI products through subnets, framing it as a crypto-native way to fund and distribute AI infrastructure. Subnet investments and commercialization (Priority: 5/5): Mark and Ala discuss three investments: Ridges for vibe coding, Targon for private inference, and Hippias for storage, emphasizing tokenomics and market efficiency. Wallets and allocation mechanisms for subnets (Priority: 4/5): Ala Shabana describes Crucible Labs’ wallet/allocator, which simplifies staking TAO and distributing rewards across subnets instead of manually trading and managing positions. OpenClaw agent tooling for founders (Priority: 4/5): Michael Ryaboy presents SetupClaw, a concierge/SaaS service that installs and operationalizes OpenClaw agents for executives and founders, with custom automations and hypercare. Bisbee: business creation in a box (Priority: 4/5): Pete and Maddie Rees demo Bisbee, an AI system that helps users validate, name, brand, and then operate a business, aiming to support workers displaced by AI. AI, media, and personal recommendations (Priority: 2/5): The hosts briefly pivot to Bob Dylan documentaries, The Gray House, Inca history docs, Ghost Elephants, and a headphone product, ending with show promos and programming notes.
Key Arguments: BitTensor uses token emissions to subsidize AI products similarly to how Bitcoin subsidized network security, but it aims at useful computation rather than purely wasteful hashing. Ridges (Subnet 62) shows that a BitTensor-backed product can match or beat leading coding copilots at far lower price, suggesting decentralized competition can beat venture-backed incumbents. TAO functions as the gateway asset for subnet participation, and subnets accrue value back into TAO, making the base token economically central. Crucible Labs lowers friction for users by automating staking/allocation across subnets, turning active subnet investing into a more passive strategy. Targon demonstrates that private, encrypted inference can be offered via distributed miners while preserving value capture through subnet tokens. Hippias argues that storage markets can be radically cheaper when miners supply capacity and token value is explicitly tied to network usage and revenue. SetupClaw’s thesis is that the real opportunity is not just OpenClaw itself, but empowering agents with wallets, email, browsers, APIs, and deployment tools. Bisbee’s thesis is that AI can turn layoff/disruption into entrepreneurial opportunity by generating a business concept, branding, and agentic back-office support. The speakers believe the next wave of valuable crypto products will be the ones that solve real-world user problems and attract non-crypto everyday users.
Data Points: Date: March 6, 2026 - Opening of the episode BitTensor subnet count cap: 128 subnets - Ala explains current network limit Ridges launch price: $29/month - Ridges vibe coding product pricing Ridges benchmark range: 73% to 88% on SWE benchmark - Performance relative to coding copilots Ridges polyglot benchmark: 96.3% - Additional benchmark performance Ridges build subsidy: ~$10 million in emissions - Mark compares subnet-funded development to Cursor’s capital raise Cursor valuation: $29 billion - Comparison point for Ridges Ridges miner earnings: ~50K subnet tokens per day - Rewards for miners improving the product TAO price reference: $177 - Jason discusses buying TAO/subnets Hippias storage cost: 400x to 4000x less expensive - Compared with Filecoin and other storage offerings VidAIO network rewards: ~$3,000 per day - Mark’s open claw mining example VidAIO miner profitability: ~$30/day revenue vs $10/day cost - Mark says his mining setup is profitable but not durable Bisbee support scope: 350+ business types - Maddie says the product supports many customized business categories Roe GLP coverage: ~50% - Jason mentions insurance coverage rate for Roe members Industry timeline mentioned: 2027 - Maddie predicts AI-related job loss will be especially significant then
Pivotal Quotes: "BitTensor does for stranded talent." — Mark Jeffrey: Explaining how BitTensor incentives attract global contributors to build subnet products "The most empowered agent is the goal for you." — Michael Ryaboy: Describing the long-term vision for SetupClaw and OpenClaw-based automation "We want to provide a way for these people who don't have a way back into the traditional workforce to start something for themselves and be able to make money and live their lives." — Maddie Rees: Bisbee’s mission to help displaced workers start businesses
Implications: The episode frames decentralized AI/crypto as a real infrastructure play, not just speculation: cheaper products, global contributor markets, and agentic tooling could reshape how software, storage, and startups are built and monetized.
About This Week in Startups
Jason Calacanis covers startups, tech, markets, media, and all the hottest topics in business and technology. He also interviews the world’s greatest founders, operators, investors, and innovators.