Pete and Andy talk about Claude Code falling apart and dropping from 100x to 90x productivity, plus why 95% of enterprise AI pilots are failing spectacularly.
They dig into Pete's "left curve" solutions approach - avoiding over-engineering and just building simple stuff that works, like his Craig David Bot demo that analyzes your week using NOSTR events. Plus updates on NOSTR protocol development with "bring your own database" concepts and cryptographic identity.
The conversation covers why MIT found enterprise AI pilots failing due to poor scoping and bureaucratic implementation, ROI measurement problems with compounding AI benefits, and the build vs buy question while big companies struggle with ineffective pilots.

Pete takes a walking conversation with Gigi who moved to Madeira and found 200+ merchants actually accepting Bitcoin payments.
They dig into eCash solving the micropayment problem for AI services, Nostr replacing username/password systems with cryptographic identity, and the "Netflix problem" of subscription fatigue. Plus why AI might let people build local communities instead of depending on Silicon Valley, and real examples of AI automating rental management businesses.
Gigi shares stories from Japanese Nostr communities, the chicken-and-egg problem with new payment systems, and why choosing optimism beats doom-scrolling when building permissionless tech.

Pete and Andy talk about remote AI development while Pete's traveling, using 'Wingman' for voice-driven coding and overnight AI security reviews with "Grug Brain Developer" principles.

They compare Claude Code (the overeager "dog with two dicks" that wants to build everything) versus Goose for planning and docs, why speaking ideas beats typing for ideation, and their strong feelings about chat interfaces versus structured documentation.
Plus they dig into when AI agents get stuck in debugging loops and the missing tools to turn big plans into actual work.

Pete unveils "Wingman"—his breakthrough Mac Mini setup that turns AI agents loose on their own dedicated computer while he controls them remotely from his phone. The system orchestrates multiple agents like Goose and Claude Code simultaneously, with automated handoffs, webhook chains, and a vision of humans as "air traffic controllers" managing fleets of AI workers.
They dive into the economics of building custom tools versus endless SaaS subscriptions, test AI models through Settlers of Catan gameplay, and explore the "fast fashion for software" philosophy where you build exactly what you need rather than settling for generic solutions.

Pete brings on special guest Bethan Winn—his wife and author of "The Human Edge: Critical Thinking in the Age of AI"—for a deep dive into maintaining human agency while embracing AI tools. They explore the concept of "AI veganism" (consciously choosing when to use AI versus manual processes), debate whether AI-generated creativity diminishes human expression, and discuss how AI democratizes specialist knowledge while creating new challenges around filtering and evaluation.
The conversation takes an unexpected turn into Australia's proposed social media ban for children, with Bethan arguing for education over restriction and the dangers of information control. They wrap up exploring the "ocean analogy" for confronting uncomfortable truths and why human empathy remains irreplaceable in an AI-augmented world—because as Bethan puts it, "the AI doesn't care about the outcomes that you get, but you do."

Pete and Andy prove that public accountability works—they actually shipped the projects they promised last episode. Pete launches Fat Controller, a Nostr scheduling tool that evolved from solving his own timezone problems, while Andy gets his autonomous "Hype Man" CMO operational, creating and publishing social content from podcast transcripts.
The conversation dives into their evolving development workflows, comparing Claude Code's targeted precision with Cursor's orchestrated complexity, and explores Pete's "Wingman" setup—a dedicated Mac Mini that lets him code from his phone using voice transcription. They wrestle with the meta-optimization trap of spending more time perfecting AI workflows than building actual products, and tease their vision for "vibe startups"—going from idea to launched company in a single day.

Pete and Andy tackle the modern developer's dilemma: when AI makes starting projects effortless, how do you actually finish them?

They explore their struggles with Claude Code's over-eager "young bull" mentality, the dopamine hit of endless ideation versus the grind of deployment, and why sophisticated AI workflows might be sophisticated procrastination.

Pete and Andy tackle the evolving landscape of subscription-based AI services with recent rate limit changes from Claude and Cursor while exploring their personal workflows across multiple specialised AI tools for coding, writing, and creative development.

The episode explores "Beacon," a project designed to democratize access to AI tools and Bitcoin for underserved populations through WhatsApp, addressing censorship and limited internet connectivity while highlighting how technology can bridge information and financial gaps in communities with restricted access to traditional digital services.

Pete and Andy dive into their evolving AI-assisted coding workflows, showcasing how tools like Claude and Cursor are enabling more structured and intentional development approaches. They discuss their implementation of sequential prompting methods, organised project management in Obsidian, and a shift from "vibe coding" to systematic workflows with built-in validation checks.

This episode explores uses of multi-agent panel discussions to unearth new insights and reduce bias, and how automation is fundamentally changing their relationship with work itself - shifting from execution and task completion to deeper questions of purpose and intentionality.

Pete and Andy introduce "slow code," their methodology for coding with AI that prioritizes intentionality over the quick-but-messy "vibe coding" approach, combining human planning with AI execution through a three-phase process of ideation, planning, and implementation.

The conversation explores their tool stack, how multi-agent conversations avoid single-model bias, and how this methodology parallels creative frameworks like "write drunk, edit sober."

Pete and Andy welcome Gav Fielding, digital marketing specialist and artist, exploring AI's creative potential and the art of prompt engineering as a form of "digital shamanism."

The conversation delves into how traditional creative processes translate to AI collaboration, examining creativity as a "volumes game," the importance of "set and setting" for AI interactions, and how proper context engineering resembles ancient summoning practices for birthing digital entities with the right incantations.

Pete and Andy explore common AI myths and misconceptions, challenging the "I trained the model" fallacy and examining why chat interfaces limit AI's true potential.

The conversation covers the productivity versus creativity paradigm, adaptive interface design, and how work might evolve to resemble play in an AI-enabled future.

Pete and Andy welcome their first guest, Joel Pember, brand director and co-founder of Juicebox, exploring his journey from photography to digital marketing and the philosophical implications of AI on creative industries.

The conversation delves into how AI is reshaping artistic expression, the evolving role of brand value and taste in an AI-driven world.

Pete and Andy explore how AI is reshaping business creation and product development. They discuss how drastically reduced creation costs enable rapid product-market fit.

The conversation spans from the potential death of traditional B2B SaaS to multi-agent AI systems that could replace current chat interfaces.

Pete and Andy get technical, exploring the difference between AI as tools versus building human-at-the-edge workflows. From why Cursor dominates to building personal knowledge graphs locally, they dive into the practical reality of AI development.

They tackle whether LLMs truly reason or just pattern match, discuss the "memory problem" plaguing AI systems, and explore workflow automation versus autonomous agents.

Pete and Andy dive into AI as the ultimate entrepreneurial renaissance, where the cost of experimentation has plummeted and solo founders can now compete with entire teams.

From learning to code with AI as a technical co-founder to building businesses for audiences-of-one, they explore unprecedented opportunities for individual agency.

Pete and Andy explore automating their entire podcast production pipeline and dive into how AI is reshaping knowledge work at the "subatomic" level. Rather than replacing entire jobs overnight, AI gradually automates individual tasks, creating ripple effects across the economy.

From streamlining content creation to effective AI coding strategies, they discuss why generalists with adaptable skills may be best positioned for an uncertain future.

Pete and Andy dive into how AI is reshaping education, making it more personal, project-driven, and permissionless. From kids launching businesses with AI to learning to code through hands-on projects, they explore how curiosity is outpacing credentials.

As traditional systems struggle to keep up, AI is making one-to-one, self-directed learning accessible to anyone. In this new world, you don’t wait to be taught, you can just learn things.

Pete and Andy argue that the AI we already have is more than powerful enough to transform businesses today—if leaders are willing to rethink their operations from the ground up.

Drawing parallels to the mechanization of agriculture, they suggest knowledge work is on the brink of similar upheaval, where cost structures can be radically compressed by decomposing tasks and deploying specialized agents today, and not by waiting for bigger, smarter models in the future.

Pete and Andy explore how AI will transform business processes through "The Intelligent Assembly Line" - breaking down complex knowledge work into smaller, automatable components.

This episode examines how AI is shifting business processes from human-centered to human-at-the-edge, and having a similar impact as Henry Ford's assembly line.

Pete and Andy examine whether to build new AI-native businesses or acquire and transform existing ones in the era of artificial intelligence.

This episode unpacks how industry characteristics, switching costs, trust requirements, and capital needs determine the optimal strategy for leveraging artificial intelligence in your business.

Andy and Pete dive into their 'Value Trap', a visual framework to explain how AI will transform industries and their approach to escape the value trap. 

They also discuss capital allocation strategies for the AI transition and the characteristics of businesses that fare better through this transition. 

The inaugural episode explores how AI will transform business models, where value will flow, and strategic approaches for businesses adapting to this new paradigm. Pete and Andy discuss the value shift, discussing that value from AI won't primarily accrue to companies like OpenAI but to traditional service businesses that use AI to transform their operations.