Writing for Machines: How To Capture Your Project's "Vibe" and Survive AI Slop
Talkat Open Source Summit India 2026 · on Wednesday, 2026-06-17· at 17:15· Jasmine Hall 2 (Level 3)·
summary
Maintainers report that AI increases churn not velocity: an OSS project rises from ~120–150 PRs/month to 300–400, with merge rate collapsing from 85% to 7.5% merge rate and many PRs containing huge unnecessary diffs (one 4,000‑line PR added a 30s sleep that broke production). Analysis finds ~2.7% of AI‑generated code introduces vulnerabilities and many “100% coverage” tests are just asserts. Fixes are concrete: add a short agents.md rules file, commit small spec files and tests first, and standardize prompting to avoid AI slop.
AI-written summary · about
chapters (23)
- 0:00AI slop in open source
- 1:08AI code slows maintainers
- 3:13Data on maintainer productivity
- 5:00AI slop impact on Fossology
- 7:31AI slop breaks contributor pipelines
- 10:20Broken workflows behind AI slop
- 12:54Fixing AI slop approaches
- 15:42Anti-patterns killing AI PRs
- 17:47Context gap root cause
- 18:31Context engineering solution
- 19:12Rules files improve AI code
- 22:41Rules file example
- 23:47Spec files as feature docs
- 24:50Prompting patterns for AI
- 25:34Explore-plan-code-commit AI loop
- 26:37Validating intent beyond tests
- 27:42Test-driven design with AI
- 29:05We are all context engineers
- 30:07Implement context engineering
- 31:54Key takeaways and AI future
- 32:17Q&A: AI agents in embedded systems
- 33:15Q&A: Preventing slop in good first issues
- 38:11Q&A: Balancing maintainers and contributors
related talks
01h 01m"Looks Good to Me": A Practical Guide to Handling AI-Generated Code
19mThe AI Code Review Bottleneck: Escaping the Velocity Trap
32mFrom Vibes to Outages: Riding the AI Code Wave
27m3 data-backed ideas to deter AI code quality problems
21mHow To Make AI-Assisted OSS Contributions Review-Ready
14mAI's impact on open-source contributions and mentoring
19mIf AI Writes the Code, Who Runs Production?
44mSpeed Is Easy, Trust Is Hard: AI's Real Impact on Developer Productivity