Standup Comedy
summary
Conference audio yields about six million transcribed words which get cleaned, run through peak detection and sentiment analysis and modeled with n‑grams (Grams‑13) to reveal that peaks align with Q&A openings, six high‑frequency filler tokens dominate the corpus so they remove noise, and repeated sentiment cycles track comprehension drops around Transformers talks. Models then predict next year’s format — technical keynote, seven community talks, six coffee/bathroom breaks — and even prize outcomes like books forgotten and an Oculus winner story with Roman as the likely top winner.
AI-written summary · about
related talks
34mComedy Special
34mKeynote: Lighthearted Thoughts on Conversing About Data
40mA Comedy Talk
31mPunchlines to Product Lines: How Stand-Up Comedy Boosted My Career
59mWelcome keynote
01h 15mKeynote: Creative Release: Why Improv Thinking Matters for Development
46mHow to Overcome Your Fears to Become a Conference Speaker
50mThe Future of Conferences