Guest
Fernando Nikolić
Bitcoin, AI and the Way Out of the Permanent Underclass | Fernando Nikolić #230
- Fernando left Blockstream as VP of Marketing to build Perception solo, achieving profitability within seven months while still in beta with zero employees or external funding. - The original dashboard product failed to gain traction; he pivoted to offering data via Model Context Protocol (MCP) for integration into Claude, ChatGPT, and internal AI workflows after discovering users wanted to reduce tool sprawl rather than add another layer. - Coding has become commoditized through AI; the hard competitive edges now lie in sales, marketing, positioning, and understanding customer psychology—not technical execution. - Running a business entirely on a fleet of AI agents instead of human staff eliminates politics, meetings, and overhead, though it requires extensive prompt engineering and orchestration to prevent mistakes and ensure security. - AI-generated content ("AI slop") is ubiquitous and indistinguishable; differentiation now demands storytelling, founder philosophy, and authentic human insight that AI cannot generate independently. - In a world of AI abundance and potential monetary debasement, non-builders should hold Bitcoin as a scarce, predictably distributed asset; builders should use AI to create products people want.
Has Bitcoin Lost Its Narrative? | Fernando Nikolic
- Saylor narrative shift: Michael Saylor's messaging evolved from Bitcoin maximalism (2020–2024) toward credit, derivatives, and Strategy as a financial product (2024–2026). Data-driven analysis shows this was a structured, three-phase narrative transition, not spontaneous. - Narrative-driven market reactions: Strategy's sale of 32 Bitcoin triggered emotional and price reaction, whereas a sale of 700+ Bitcoin in 2022 went unnoticed. The difference lay in narrative engineering—when the story breaks, actions carry weight; when it's solid, they're rationalized away. - Death of monoculture: Bitcoin and internet culture have fragmented into insulated niches via algorithmic feeds, personalization, and social media. The unified Bitcoin movement that existed pre-ETF approval no longer exists; adoption now happens across disconnected communities with contradictory understandings. - Crisis-era adoption patterns: Bitcoin adoption is shifting from counterculture movement to boring, slow, fragmented growth across multiple interpretations. Without shared narratives or memes, outsiders perceive Bitcoin as stagnant, yet adoption quietly accelerates at the edges. - Homogenization via AI and algorithm: Netflix, Spotify, and now AI homogenize culture by optimizing for safe, bland consensus. Self-sovereignty and high-agency content-seeking remain possible but only for a small minority; most accept the "wrapped" version and lose taste autonomy. - Print press analogy: The chaos following Gutenberg's invention lasted 300 years; we're in a similar "chaos gap" with Bitcoin, AI, and the internet. Current disruption may take generations to resolve—expect persistent fragmentation, not quick clarity.
#185 - Fernando Nikolić - AI, One-Person Companies & The New Economic Elite
- AI has dramatically lowered the cost of building software; one developer can now accomplish what previously required teams of 15+ people and €1 million budgets in weeks using Claude, Gemini, and agentic workflows. - Running a solo company with AI agents as virtual team members is viable; Fernando operates perception.to as a one-person organization with 94% profit margins while spending only €200/month on Google AI Ultra subscriptions. - AI-generated content quality varies sharply by domain: coding and data processing are reliable, but creative work—jokes, marketing strategy, tone—remains mediocre and homogenized, lacking human flair and cultural nuance. - The "SaaS apocalypse" is underway; tools like Squarespace, Photoshop, and Grammarly face obsolescence as custom AI-built software becomes cheaper and more flexible than subscription products. - Democracy and accountability tools powered by AI are emerging; Peter built a system tracking 628 UK MPs across 10,915 tweets with 81% accuracy in fact-checking claims, potentially surfacing corruption and political dishonesty at scale. - The productivity surge is unevenly distributed; exponential AI gains risk concentrating wealth among early adopters while creating a permanent underclass without access to inference costs, tokens, and computational infrastructure.