F2N The Internet: communication becomes a network.
1. Internet is a Communication Network, it Transformed Distribution • It connected users and computers in a many-to-many network with infinite channels • Every user had the ability to market and sell, replacing traditional centralized sales and marketing channels • This democratized distribution, internet businesses were not really about deep technology One-to-many Many-to-many means new and different sales and marketing F2N
2. Internet Distribution-Side Network Effects • More users = more value for each user → virality • E.g., a social network is more useful with 1,000 users than with 2 • Network effects accelerated growth with minimal marginal distribution cost ? More users = more value per user = more attractive for new users Less users = less value per user = less attractive for new users New user F2N
3. Internet Economics and the Growth-Monetization Tradeoff • Customer acquisition cost (CAC) decreased with scale. • Customer lifetime value (LTV) increased with network size. • Monetization early created friction → delayed monetization was optimal. • Therefore, growth was prioritized over revenue. CAC LTV Future revenue potential improves with growth via delayed monetization F2N
4. Venture Capital Was the Enabler • Delayed monetization meant a funding gap. • Only one or few players could win (winner-takes-all). • High downside risk for investors in non-winners. • Needed uncapped upside → equity, not debt. • Venture Capital (VC) rather than debt suited this model → flourished and scaled global tech giants. $ Debt financed businesses with pressure to monetize from capped upside on loan, grow slower and can’t compete with network monopolies High risk of failure subsidized by uncapped gains from winners Equity finance subsidizes gap between growth and revenue to capture market F2N
5. Internet Structural Change was Bigger Than Printing Press • Reformation triggered by printing press and removal of priesthood’s control over information • Many-to-many is a bigger change, is creating similar disruption to Reformation • Over-connected networks favor misinformation, ideas spread too easily. • Finite truths, infinite lies = compelling lies win and networks vulnerable to conspiracy and groupthink • Future stable equilibrium = small world networks - tribal clusters, few trusted connectors between Hand written information, few channels Printed information, many channels Networked information, infinite channels More democratic but new vulnerabilities F2N
F2N AI: software itself becomes a network
6. AI is a Software Producing Network, Production-Side Network Effects • Unlike internet, AI networks generate production value, not distribution • AI writes software, robots builds robots → AI creates agents of agency • Software improves through larger models and more usage/data (feedback loop) • Bigger models replace smaller, niche applications, big AI kills little AI AI = software network that produces software = network effects in production of software, from apps to eventually building a better version of itself. F2N
7. AI Era Capital Flows Shift Upstream • Entire software ecosystem dependent on upstream hardware infrastructure and energy costs which create running costs • Value accumulates in LLM providers, chipmakers, energy suppliers, and hyperscalers • Software becomes commoditized, infrastructure becomes critical • AI creates upstream concentration of power and capital AI becomes use-metered infrastructure for all apps, not just like electricity, but ultimately dependent on it F2N
8. AI Capital Flow Changes Disrupt VC Funding Model Most niche AI startups are vulnerable due to: • Lack of defensibility vs. hyperscalers. • High infrastructure cost. • Network effects at the AI model level, not the app level. Funding model shifts: • Toward corporate VC (tech platforms), infrastructure funds, or true deep tech. • Niche AI apps only survive by monetizing early, agency style, and avoiding equity dependency. Agencies Deep Tech VC Backed SaaS, AI Death Zone F2N
F2N AI embedded hardware: manufacturing eats the world
9. Manufacturing is Eating the World • US tech platforms dominated in Europe and elsewhere (China built its own), coining the phrase “software is eating the world” • Now, the commanding heights of new economy are based on hardware with embedded smarts (e.g. self driving cars, drones, phones) and China dominates. Manufacturing is now eating the world. • Europe is caught in a vice as it has pre-digital era services and manufacturing. F2N digital era manufacturing software China US Europe
10. AI Era Economic Commanding Heights: AI-Embedded Hardware • Critical economic infrastructure: AI + hardware (e.g., drones, phones) • Physical goods retain marginal cost → cannot be zero • Therefore, manufacturers with best cost structure win • Software may be given away to boost hardware sales (e.g., BYD vs. Tesla) Chinese Model: unbeatable manufacturing margins, embedded software bundled for free Loss Leader Margin US Model: Subscription for software embedded in hardware F2N
11. China is the New Hegemon • China has the majority of the global share of many of the manufactured items that constutute the new economic commanding heights. • This includes supply chains and raw materials. F2N
12. AI Manufacturing and Automation • AI replaces services industry jobs but not all as many services are premium if they have human touch • AI also makes manufacturing more automated, reducing labor needs (robotic, lights-out factories), less manufacturing jobs benefit from human touch (i.e artisanal), but not all. • China’s manufacturing workforce has already halved due to automation • Jobs shift to services, but only human-centric, scarce ones endure Premium, high touch, mostly services labor replaces faceless manufacturing, clerical and driving/delivery No high touch human demand = AI economy stalls, equilibrium reached via consumer demand from sufficient redistributed AI driven wealth. F2N
F2N A New social Pact: sustainable free-market consumerism as demand side socialism
13. Societal & Economic Impact of AI • AI agents (virtual + physical) replace human labor • Ownership of AI = ownership of the labor force • Capital dominates labor more than in the Industrial Age • A new social contract is needed to distribute labor productivity gains Means of production Capital Labor AI robotics = Labor and means of production F2N
14. Consumerism as a Social Model • Redistribution may come via demand-side leverage, not supply-side labor organizing • Historical precedent: Ford paying workers enough to buy the cars they made • Popular culture succeeded by serving large, low-income audiences. Hollywood bigger market than Opera • Competitive consumption drives reinvestment and distribution • AI dramatically lowers production costs and so the gains only work with large, mass markets and popular culture More consumers, higher return on capital = consumer leverage through aligned interest rather than collective bargaining Products that only service the elites, circle capital among them with less return and less redistribution Opera Movies F2N
15. Free markets vs Capitalism • China has distinguished between free markets and democracy and capitalism to allow for consumer choice and competitiveness • There can be choice of leadership (democracy), choice of products (competition) and choice of components within products (supply chain non integration) • China has highly vertically integrated manufacture (BYD makes its own motors, batteries and chips, SpaceX in the US is also vertically integrated) but with extreme competition amongst end product providers at early stages. F2N Fierce competition (choice) Supply chain vertical integration (efficiency)