# Wisdom of the $TAO: the future is decentralized AI Source: https://www.youtube.com/watch?v=X0AJVcKieyA Recap page: https://rapidrecap.app/video/X0AJVcKieyA Generated: 2026-03-06T23:32:48.013+00:00 --- ## Quick Overview The future of decentralized AI hinges on projects like Bit Tensor ($TAO), which programs Bitcoin's mining structure to incentivize the development of 128 specialized AI subnets, exemplified by Subnet 62's code platform competing favorably with Claude at a fraction of the cost. **Key Points:** - Bit Tensor, with coin $TAO, programs Bitcoin's mining side onto its own chain to subsidize the growth of new AI products across 128 subnets, distributing $100 million annually. - Ridg's Subnet 62, a vibe coding platform built on Bit Tensor, scores 73-88% on benchmark tests, competing with or exceeding Claude while costing five to seven times less ($29/month). - The decentralized structure allows global participation; miners compete to improve subnet products, earning subnet tokens, providing a way for individuals to earn from the AI revolution without traditional employment. - Stillcore Capital invests in specific subnets, including Ridg's (Subnet 62), Targon (Subnet 4 for private inference), and Hippius (decentralized storage), treating subnets like early-stage startups for evaluation. - Hippius storage offers significant cost reduction (400x to 4,000x cheaper than competitors) and features novel tokenomics tying the token value deterministically to usage and revenue. - Crucible Labs, co-founded by Alla Shabbana, developed the Crucible wallet to simplify staking $TAO and allocating rewards across various subnets, easing access for non-technical users. - Open Tensor Foundation development is decentralizing further, with original founders stepping back to allow the community to maintain the open-source project, mirroring Bitcoin's maintenance model. **Context:** The discussion takes place on the 'Twist' podcast, dated March 6, 2026 (designated AO41), featuring hosts Jason Calacanis and Lon Harris, alongside guests Mark Jeffrey of Stillcore Capital and Alla Shabbana, co-founder of Crucible Labs and Bit Tensor. The conversation centers on the intersection of AI, decentralized finance (crypto), and the rapidly evolving infrastructure provided by the Bit Tensor network, contrasting it with centralized AI models like Claude. ## Detailed Analysis The core value proposition discussed is Bit Tensor ($TAO), which essentially modularizes and decentralizes AI development using a structure analogous to Bitcoin mining; miners compete to improve specific AI products housed in 128 subnets and are rewarded with $TAO or subnet tokens, effectively subsidizing AI growth with $100 million annually. Mark Jeffrey highlighted the launch of Ridg's Subnet 62, a coding platform that delivers comparable or superior performance to Claude at significantly lower prices, noting that Ridg's achieved this with $10 million in emissions while Cursor raised at a $29 billion valuation. Stillcore Capital's investment thesis involves evaluating these subnets as startups, focusing on market fit and competitive advantage, exemplified by their investments in Targon for private, encrypted inference using trusted execution environments, and Hippius for decentralized storage that boasts extremely low costs and a tokenomics model directly linked to usage. Alla Shabbana emphasized that Crucible Labs created the Crucible wallet to solve the complex user experience barrier, allowing users to easily stake $TAO and automatically allocate rewards to top-performing subnets, noting that listing subnet tokens on major exchanges like Coinbase requires them to go through similar legal scrutiny as $TAO itself. The conversation concluded with the observation that this decentralized compute structure creates relentless downward pressure on the cost of various services, including storage, inference, and video processing/upscaling, pushing towards an 'age of excellence' in efficiency. ### Tool Integration for Productivity - The hosts utilize a 'Plaude pin' for recording and saving action items directly to their agents, illustrating reliance on agents for task management and delegation. ### Bit Tensor Architecture - Bit Tensor ($TAO) applies Bitcoin's mining concept to AI, incentivizing 128 subnets to build specialized AI products via programmable mining rewards. ### Subnet Performance Showcase - Ridg's Subnet 62 offers code assistance comparable to Claude at 5-7x lower cost; Targon (Subnet 4) provides industrial-grade, provably encrypted private inference; Hippius delivers decentralized storage at dramatically lower prices. ### Investment Thesis at Stillcore - Mark Jeffrey and Jason Calacanis invest in subnets by evaluating them as startups, assessing TAM, competitive advantage, and team quality, with a focus on which subnets will break out. ### Decentralized Earning Potential - Miners compete globally to improve subnet products, earning tokens that can be cashed out or held, offering a permissionless way for anyone to monetize compute resources or talent. ### Crucible Labs' Role - Alla Shabbana's Crucible Labs created a wallet that acts as an allocator, simplifying the process of staking $TAO and distributing returns across the decentralized subnet ecosystem. ### Subnet Launch and Governance - New subnets emerge permissionlessly, though a variable registration cost (based on $TAO price and demand) exists; the current cap of 128 subnets is intended to maintain quality over quantity, similar to a library rather than a party.