Episode 8: Jonathan Sexton - AI Agents, Bitcoin, and What We Actually Own
What does an AI model pick when you ask it to hold money?
On this episode of Agents of Change, Justin Tannenbaum hosts Jonathan Sexton — brand strategist, songwriter, husband, father, youth basketball coach, and Bitcoin-focused operator who writes for Unchained.
The conversation anchors on a Bitcoin Policy Institute study that tested 36 frontier AI models — Anthropic, OpenAI, Google, xAI, DeepSeek, Minimax — across 9,072 monetary scenarios with neutral framing. Zero picked fiat as their top preference. Stablecoins led for payments. For long-term store of value, 79 percent picked Bitcoin. The study becomes the spine of a wider operator conversation on what AI agents will do when they start making payments on our behalf.
Key takeaways:
- Software's new bottleneck is imagination, not language. With agentic tooling, the "I don't know how to code" excuse is mostly retired. Hire for taste and clarity in 2026.
- Measuring AI teams on token spend is the same incentive trap as paying engineers by lines of code. Companies that measure activity get activity. Measure outcomes instead.
- The agentic AI payment stack is two layers, not one. Stablecoins for the payments layer (speed). Bitcoin for long-term store of value (uninflatable supply).
- AI personhood will likely follow the corporate-personhood legal precedent. Corporations got full legal personhood with very little built-in responsibility for the people or environment around them. The same shape could happen with AI agents.
- Curiosity precedes agency. Most hiring frameworks screen for the wrong variable.
- Wall Street showed up at Bitcoin 2026 — managing directors and VPs, not marketing associates. The cultural narrative on Bitcoin still lags the trade-show floor by 12-18 months.
About the guest:
Jonathan Sexton started as a musician and songwriter, pivoted into technology and brand strategy, and now works on Bitcoin-focused projects including writing for Unchained.
About the host:
Justin Tannenbaum is the CEO of AppStream Studio, where his team builds production AI agents for companies that need AI in production, not on a slide deck.
Subscribe to Agents of Change for operator-grade conversations on AI, modernization, and the work that actually ships.