On this episode of Crazy Wisdom, I, Stewart Alsop, sit down with Sweetman, the developer behind on-chain music and co-founder of Recoup. We talk about how musicians in 2025 are coining their content on Base and Zora, earning through Farcaster collectibles, Sound drops, and live shows, while AI agents are reshaping management, discovery, and creative workflows across music and art. The conversation also stretches into Spotify’s AI push, the “dead internet theory,” synthetic hierarchies, and how creators can avoid future shock by experimenting with new tools. You can follow Sweetman on Twitter, Farcaster, Instagram, and try Recoup at chat.recoupable.com.

Check out this GPT we trained on the conversation

Timestamps

00:00 Stewart Alsop introduces Sweetman to talk about on-chain music in 2025.
05:00 Coins, Base, Zora, Farcaster, collectibles, Sound, and live shows emerge as key revenue streams for musicians.
10:00 Streaming shifts into marketing while AI music quietly fills shops and feeds, sparking talk of the dead internet theory.
15:00 Sweetman ties IoT growth and shrinking human birthrates to synthetic consumption, urging builders to plug into AI agents.
20:00 Conversation turns to synthetic hierarchies, biological analogies, and defining what an AI agent truly is.
25:00 Sweetman demos Recoup: model switching with Vercel AI SDK, Spotify API integration, and building artist knowledge bases.
30:00 Tool chains, knowledge storage on Base and Arweave, and expanding into YouTube and TikTok management for labels.
35:00 AI elements streamline UI, Sam Altman’s philosophy on building with evolving models sparks a strategy discussion.
40:00 Stewart reflects on the return of Renaissance humans, orchestration of machine intelligence, and prediction markets.
45:00 Sweetman weighs orchestration trade-offs, cost of Claude vs GPT-5, and boutique services over winner-take-all markets.
50:00 Parasocial relationships with models, GPT psychosis, and the emotional shock of AI’s rapid changes.
55:00 Future shock explored through Sweetman’s reaction to Cursor, ending with resilience and leaning into experimentation.

Key Insights

1. On-chain music monetization is diversifying. Sweetman describes how musicians in 2025 use coins, collectibles, and platforms like Base, Zora, Farcaster, and Sound to directly earn from their audiences. Streaming has become more about visibility and marketing, while real revenue comes from tokenized content, auctions, and live shows.
2. AI agents are replacing traditional managers. By consuming data from APIs like Spotify, Instagram, and TikTok, agents can segment audiences, recommend collaborations, and plan tours. What once cost thousands in management fees is now automated, providing musicians with powerful tools at a fraction of the price.
3. Platforms are moving to replace artists. Spotify and other major players are experimenting with AI-generated music, effectively cutting human musicians further out of the revenue loop. This shift reinforces the importance of artists leaning into blockchain monetization and building direct relationships with fans.
4. The “dead internet theory” reframes the future. Sweetman connects IoT expansion and declining birth rates to a world where AI, not humans, will make most online purchases and content. The lesson: build products that are easy for AI agents to buy, consume, and amplify, since they may soon outnumber human users.
5. Synthetic hierarchies mirror biological ones. Stewart introduces the idea that just as cells operate autonomously within the body, billions of AI agents will increasingly act as intermediaries in human creativity and commerce. This frames AI as part of a broader continuity of hierarchical systems in nature and society.
6. Recoup showcases orchestration in practice. Sweetman explains how Recoup integrates Vercel AI SDK, Spotify APIs, and multi-model tool chains to build knowledge bases for artists. By storing profiles on Base and Arweave, Recoup not only manages social media but also automates content optimization, giving musicians leverage once reserved for labels.
7. Future shock is both risk and opportunity. Sweetman shares his initial rejection of AI coding tools as a threat to his identity, only to later embrace them as collaborators. The conversation closes with a call for resilience: experiment with new systems, adapt quickly, and avoid becoming a Luddite in an accelerating digital age.