Nathan Lambert: all open model adoption metrics are incredibly imprecise
pstAsiatech · x · 2026-11-05
Nathan Lambert's first open model economy research note on Interconnects asks how big the open model inference market really is versus OpenAI/Anthropic. He argues every available measurement is imprecise: noisy platform metrics, marketing-driven startup numbers, imperfect public evals. He uses open-inference startups Baseten, Fireworks, and Together AI as bellwethers, plus OpenRouter's reported open model token share, while hyperscalers like Bedrock and Azure likely hold major undisclosed share. The dashboard is free; research notes are paid with discounts for policy researchers.
Related event: Nathan Lambert Launches Open Model Economics Dashboard(3 posts)→
More from Venture
- Intel CEO Lip-Bu Tan buys $9.71M of INTC at $64.73 after 40% 5-day selloff — zephyr_z9 · 2026-11-05
- SEO Teardown of Ben's Bites: 118K Subscribers but Only ~890 Google Visits a Month — iannuttall · 2026-11-05
- Perplexity agent runs $30K Meta spend at 3.5x revenue, 63% higher ROAS — krishnan · 2026-11-05
- Web design duo made $74K in October by raising prices and AI-drafting redesigns before sales calls — Murky_Explanation_73 · 2026-11-05
- Steam Next Fest hits 8,000+ demos as AI flood stretches review times to 12 days — draginol · 2026-11-05
- Two S-tier devtools startups acquired in a single day, devs say — JoshPurtell · 2026-11-05