Hypothesis

Impact of AI

  1. Coding Gets Commoditized — Engineering Is No Longer the Bottleneck
  2. SaaS Decays — Software Becomes Radically Customizable
  3. Context engineering and data may be the short-term moat. Over time, network effects, physical assets and regulation become the durable moat.
  4. Long-Running, Parallel, Autonomous Agents Become the Norm [already consensus]
  5. UI/UX Collapses Into Terminal + Agent Interface (potentially CLI)
  6. The Workforce Transforms — Rise of Self-Employment & Death of Junior Roles
  7. AI-Native Companies Win Every Incumbent Category
  8. The Real Opportunities Are Second-Order Effects
  9. Massive Societal & Demographic Shifts
  10. Financial Infrastructure Gets Rewired
  11. AI Slop, Malicious Agents & Human Verification
  12. Agents Become Economic Actors — Collaborating, Delegating & Hiring

AI x Financial Services

  1. Banks will still exist in 100 years in some shape and form. Banks do three things: Store value, move value, and transform value across time. The first two are commoditizing toward zero. The third (who to lend to, at what price, under what terms) is where all the profit lives, and society will always need institutions that intermediate risk. The form will change radically; the function won't disappear.
  2. It is easier to build a bank than it is to sell a bank AI tools. Banks are still built on COBOL and mainframe. Their org structure, tech debt, and risk culture prevent transformative AI adoption. The market for being the AI-native company is 100x larger than selling tools to the incumbent. Don't sell AI to banks. Be the bank.
  3. The global retail banking market is expected to grow from ~$2.5T today to ~$4T+ in the next decade (~5-7% CAGR). Neobanks are growing 20-35% CAGR. Nubank (100M+ users), Revolut (70M), Chime (35M), and Cash App (57M MAU) prove the model works at scale.
  4. Neobanks were built on a mobile-first paradigm. They improved the interface, not the bank: better UX and distribution, but the same products and cost structure. AI-native banks can dramatically reduce costs, unlock new features, and reinvent banking.
    1. An AI-native bank's cost structure is a 5-10x structural advantage — but it's an enabler, not the product. Traditional banks run 1,000-2,000 employees per 1M customers; AI-native banks can run 150-350. That doesn't win you customers, but it lets you survive long enough to build the thing that does.
    2. Neobanks (Chime, Current, Dave) are stuck in no-man's land. Too big to pivot to AI-native architecture, too small to compete with JPMorgan on products and trust. They all have the same ICP (young, lower-income, underbanked) and the same features (no fees, early DD, cash advances). Incredibly commoditized.