The New Rules of Software Engineering (Google DeepMind Lead)
TL;DR
Code is no longer the main bottleneck: Logan says AI has made code generation fast enough that review, CI, test execution, and human coordination now slow teams down more than typing code does.
The IDE is not dying, but it matters less: He expects engineers to spend more time in agent managers and review flows, while still needing editors for debugging, direct edits, and deeper inspection.
Taste is becoming a core engineering skill: The best engineers know when to ask more questions and when to just build, cutting through the human "ping pong" that often delays product work.
Agent context may become as important as test coverage: Logan describes a future of "agent coverage," where repos include design tradeoffs, chat threads, meeting notes, and specs so agents can act with full context.
Your ambition should reset every few months: Because capabilities keep improving every two weeks, three months, or six months, he believes personal projects can now be scoped more like "mini companies" than side hacks.
Proof of work matters even more in hiring: Open source still helps, but now candidates can also build complete products themselves, as long as they do not confuse outsourcing intelligence with outsourcing understanding.
The Breakdown
You can outsource intelligence, but you cannot outsource understanding. Logan Kilpatrick argues that AI has changed the bottleneck in software from writing code to reviewing it, testing it, and giving agents enough context, which means ambitious builders need better taste, better systems, and a much higher bar for what one person can create.
Was This Useful?
Share
Keep Reading
Make Alcreon Yours
Tune your feedFive quick questions, and the feed ranks what matters to you first.Or just get notified
The weekly Echo. Signal worth keeping in your inbox.
Every new piece, announced on X.
Read Next
See all
Playbook
Tasteful Skills
“Tasteful Skills” argues that the best agent skills are not documentation or best-practice lists.

Playbook
The Art of Tasteful Prompting
Learn how tasteful prompting helps you move beyond generic AI output by shaping context, style, and judgment from the start.

Playbook
The Codex /goal Playbook
OpenAI shipped /goal for the Codex CLI. It turns a prompt into a persisted, self-continuing contract.