Transcribed, then measured
The hard numbers, pulled from the talk.
Every figure below was spoken on stage and lifted straight from the transcript — grounded to the exact quote, and linked back to the moment it was said.
-
VercelAutomating 90% of support tickets: How we built Vercel's support agent — Ship 26 NYCKathryn Middleton · Vercel91%Tickets auto-resolved -
AI EngineerEvery company should have a Brain — Garry Tan, Y CombinatorGarry Tan · Y Combinator~400xOutput multiple -
@ScaleFireside Chat with Boris Cherny, Head of Claude CodeBoris Cherny · Anthropic1,700 PRsPull requests this year -
Lenny's PodcastHead of Claude Code: What happens after coding is solved | Boris ChernyBoris Cherny · Anthropic4%Share of GitHub commits -
AscendionHow Ascendion is Engineering Value With AI ArbitrageKarthik Krishnamurthy · Ascendion£2BAdditional revenue -
Mark KashefHow I 10X My Productivity with ONE Claude Code TrickMark Kashefup to 10x more productiveProductivity -
AscendionHow to Build an AI-Native Company | Karthik Krishnamurthy (Ascendion)Karthik Krishnamurthy · Ascendion$60MSavings on a $200K project
All numbers are also queryable as stats.parquet (DuckDB-ready).