18 hours ago
- The author introduces a versioning system for updates on the AI bubble instead of volumes.
- The AI bubble is described as a psyop, melodrama, financial crisis, and exploitation of ignorance for trillions in investment.
- LLMs are real but don't fulfill promises like replacing 50% of white-collar jobs, as claimed by boosters like Dario Amodei.
- AI markets are driven by fan fiction and scams, with companies like OpenAI and Anthropic discussing potential rather than actual performance.
- 89%+ of AI revenues and 90%+ of compute demand come from OpenAI and Anthropic, based on money-losing subscriptions and unsustainable token burn.
- Enterprises are capping AI spend after blowing through annual token budgets, highlighting lack of ROI.
- The entire AI industry's revenue barely reaches $100 billion by 2026, with no profitability outside data center component sellers.
- AI products are inconsistent, unreliable, and improvements are measured by increasingly useless benchmarks.
- Generative AI is limited by probabilistic hallucinations, making every output suspicious and decisions arbitrary.
- AI cannot replace most jobs; it only generates and summarizes at varying complexity, with CEOs disconnected from real work.
- Media failed to scrutinize AI hype, which is inflated by vague claims like 'AI can build software'.
- Data center construction is delayed, and AI compute demand is concentrated in Anthropic and OpenAI, which rely on debt and venture capital.
- NVIDIA's growth depends on a few customers, with demand for components high but AI compute demand low.
- The AI bubble differs from the Dot Com Bubble and is driven by 'Business Idiots' throwing cash at a dead-end technology.