AI Startup Retrospective: Only Traffic, Training, and API Sales Survive
yangyi · x · 2026-08-01
The author reviews the AI track from 2023 to 2026, concluding that only three business models remain unchanged: traffic monetization (e.g., directories, SEO, influencer marketing), training (from image generation to agent engineering), and selling APIs/infrastructure.
Many once-popular applications have been wiped out by advancing frontier models, such as virtual try-on, ComfyUI workflows, digital human cloning, and low-code platforms like Coze/Dify. The author argues that the inevitable path for C-end efficiency tools is stitching together APIs, squeezing model discounts, and selling courses to bridge the adoption gap. However, this results in low user loyalty and a gross margin capped at around 50%. Ultimately, the author dismisses C-end AI efficiency tools as low-barrier ventures with minimal investment value.
More from AGI Musings
- Can AI Replace Lawyers in Due Diligence? Practitioners Debate Its Limits — jkubicki · 2026-08-01
- AI Coordination Before First Major Incident: Pacing Ahead of Time — danfaggella · 2026-08-01
- Why AI Struggles to Become Einstein: Paper Highlights Lack of Principle Discovery — OdedRechavi · 2026-08-01
- Stanford Radiologist Counters AI Replacement Fears: Tech Saves Only 20% of Time — curtlanglotz · 2026-08-01
- Fable Discusses Divergent Visions for AGI Between DeepSeek and US Frontier Labs — teortaxesTex · 2026-08-01
- If AI Reduces Human Labor Value to Zero, What Is Society's Baseline Worth? — tealCrayon98 · 2026-08-01