AI Coding

AI doesn’t fix your bottlenecks. It amplifies them. DORA’s report on AI-assisted software development confirmed that AI amplifies existing flow and exposes brittle processes and frictions. If your flow is good, AI makes it better. But if you have a lot of painful bottlenecks and frictions, AI makes the pain worse.

Right now, every Claude Code user is running without LSP. That means every time you ask "where is processPayment defined?", Claude Code does what you'd do with a terminal. It greps. It searches text patterns across your entire codebase, reads through dozens of files, and tries to figure out which match is the actual definition.

cross-posted from: https://lemmy.bestiver.se/post/964698

Comments

Let’s sum it up

1. Place documentation where both human developers and AI agents can expect it. For example, in README.md, not AGENTS.md.
2. Prefer exposing functionality as command-line tools and APIs, which are well accessible to developers and AI agents alike, over GUI and MCP tools.
3. Avoid parameters (command-line, environment, etc.) that segregate workflows between humans and AI agents, for example, avoid AI=true, prefer --quiet or—even better!—design your tools with limited context/mental-overload in mind.
4. In general, avoid making workflows that are available to AI but hard to access for humans, and vice-versa.

Developers need the dopamine hit of creation. That’s not a perk, it’s what keeps good engineers engaged, learning, retained, and prevents burnout. The joy of coding is probably what allowed them to become experienced devs in the first place. Replace creation with oversight and you get faster burnout, not faster shipping. You’ve turned engineering, the creative work, into the worst form of QA. The AI does all the art, the human folds the laundry.

The most useful thing you can do right now is hold two ideas in tension simultaneously. Coding has changed dramatically. Software engineering, at its core, has not.

cross-posted from: https://lemmy.bestiver.se/post/948833

Comments

TL;DR: A good mental model is to treat AGENTS.md as a living list of codebase smells you haven’t fixed yet, not a permanent configuration. Auto-generated AGENTS.md files hurt agent performance and inflate costs by 20%+ because they duplicate what agents can already discover. Human-written files help only when they contain non-discoverable information - tooling gotchas, non-obvious conventions, landmines. Every other line is noise.

AI scales the groundwork; teams that successfully adopt AI typically already have solid foundational practices in place, while those lacking them struggle to get value from their AI investments.

cross-posted from: https://lemmy.bestiver.se/post/944071

Comments

cross-posted from: https://lemmy.bestiver.se/post/939521

Comments