Spill the tea
summary
No-slides, no-filter session of hot takes and concrete anecdotes: code generation using Claude code restarts a hands-on hobby and produces surprisingly reliable output, but LLM calls incur unreliable latencies and ballooning costs and require verification; database marketing often touts claims of ACID without specifying isolation levels (defaults like read‑committed cause subtle correctness bugs); CockroachDB name stays defended as a net positive with real enterprise recognition and customer billing workarounds.
AI-written summary · about
chapters (11)
- 0:00Introduction
- 1:11Coding to AI-assisted development
- 3:34AI’s future database demands
- 6:03Spill the tea game begins
- 8:29Debunking vendor claims
- 10:11Tech buzzwords to eliminate
- 11:15AI projects unlikely to survive
- 14:24LLM infrastructure mistakes
- 15:47Oracle’s legacy response
- 16:31Why CockroachDB’s name
- 17:54Social media comments
related talks
54mSeeing Like a Programmer - Resiliency, Limits, And Moral Hazards In SE
01h 02mOpening Keynote: FIXME
57mCoffee-break with EngX Experts
45mDynamic vs Static Having a Discussion without Sounding Like a Lunatic
31mFuture-proofing operational data at scale
50mStop Punching Yourself in the Face
55mKeynote: Experimenting in a Sociotechnical System in Software Engineering
20mFireside Chat