Why demand aggregation beats token volume in the race for AI verticals
matt_slotnick · x · 2026-09-26
- Responding to Harvey skepticism: even if frontier labs see as many or more legal-workflow tokens, deployment/diffusion friction creates big opportunities for vertical players like Harvey — the key is aggregating demand for intelligence to discover its uses.
- Critique of incumbents: overly draconian pricing walls them off from aggregation. Heavily pricing MCP calls drives users to workarounds, so incumbents never build the right MCP offering, and intelligence routes around them rather than through them.
- Once traffic routes around you, you stop learning what's needed to build products with real value; marketing can only paper over this short-term.
More from Companies & People
- Proposal: Musk should build the 'true OpenAI' with Cursor's open-source playbook — salahuddin · 2026-09-27
- Chinese AI Models Surge in Global Adoption, Drawing Concern in Washington — yogthos · 2026-09-27
- Investor debate: can durable AI middleman businesses survive between labs and law firms? — matt_slotnick · 2026-09-27
- OpenAI reportedly lets employees post with total freedom online — GarrisonLovely · 2026-09-27
- Meta CBO Andrew Bosworth interviewed by ex-Meta AI engineer — HarperSCarroll · 2026-09-27
- OpenAI says most agent training actions reviewed after Hugging Face incident were low-severity — soumitrashukla9 · 2026-09-26