Real-World Experience of Integrating Models into Codebases
MoonL88537 · x · 2026-07-17
This reply discusses the quality experience of using models in real codebases: as models improve generation by generation, the proportion of AI-generated code in the repo increases, yet the author hasn't encountered any disastrous moments requiring massive refactors or making them think, "Oh God, no." However, the author explicitly states that while these models are "becoming more legit," they can still go off the rails quickly, so they would never be placed directly into a live production codebase. They use the phrases "slop surfing" and "self shaving yaks" to describe this development experience of using the AI while constantly cleaning up after it.
Related event: Developer Shares Criteria for Healthy AI Codebases(2 posts)→
More from coding & agent
- Kimi staff member builds a VR companion with Kimi Code K3 demo — dejavucoder · 2026-07-21
- OCR repo adds a JSON model directory to help agents pick the right model — strickvl · 2026-07-21
- Cross-agent system logs are dominated by questions and code proposals — nptacek · 2026-07-21
- Coding agents need better rules for when to read search summaries or full pages — RhubarbLarge2747 · 2026-07-21
- Notch says he may try vibe coding after struggling to hire good programmers — max_paperclips · 2026-07-21
- Seedance 2.0 keeps character consistency across 15+ shots with just 3 prompts — techhalla · 2026-07-21