Article · 2026-08-19 · 6 moments

Conceptual integrity and counting lines of code

Last week I recorded an episode of the Talking Postgres podcast with Claire Giordano on the subject of "How AI is changing software development". We had a really great conversation. Here are a couple of my highlights from a lightly edited transcript (prompt to Claude: "very minor edits to remove disfluencies"). This is the latest version of an argument I've been trying to build about why sometimes it does make sense to talk about lines of code as an indicator of productivity with coding agents, ✦ AI generated

Read original ↗

HOST Simon Willison

timeline · colored by role

Claim◆

Lines of code can meaningfully measure productivity because there is a hard ceiling on how much production-ready code a human can write, and coding agents can substantially raise that ceiling.

Despite conventional wisdom, LOC is a legitimate productivity indicator because human output has a hard cap (~200 lines/day of production code), and agents that reliably produce 10x more high-quality code represent a genuine, measurable improvement.

transcript

Simon Willison: A lot of people will tell you it makes no sense to measure productivity in lines of code. I'd actually disagree, because there's a hard limit. In the before-times, a software engineer could produce a few hundred lines of production-ready code per day — and 200 lines of working, debugged, production-level code is an incredibly good day. Most days you'd produce 50 or 60. If agents let you produce a thousand lines of debugged code, that really is a very meaningful improvement — as long as the code is the same quality: maintainable, tested, all of that.

Claim

Achieving high-volume, high-quality output from coding agents requires significant skill, knowledge, and experience — the qualities that define senior engineers.

Agents can produce far more code, but only when wielded by engineers with the judgment to maintain quality — making senior-level skill the critical enabler of agent-driven productivity gains.

transcript

Simon Willison: You can get to that point with agents, but it takes a huge amount of skill and knowledge and experience. That's what senior engineers are made of.

Mechanism

The new bottleneck created by coding agents is cognitive capacity — an individual engineer can generate code far faster than they can reason about it, so teams remain essential for load-balancing cognitive load.

Even though agents let a single engineer produce far more code, no individual can maintain oversight of 100x the codebase alone, so engineering teams are still necessary to distribute the cognitive burden.

transcript

Simon Willison: I can do way more work as a single engineer than I could without agents. So you could argue, why should a company have more than one engineer? Beyond the obvious bus factor thing — a team of one is a very badly designed team — the answer is that the new limiting factor is cognitive capacity. I can churn out code a hundred times faster. I don't have the cognitive capacity to stay on top of 100 times the amount of code. So you still need a team of engineers, so you can load balance that cognitive capacity across the team.

explains mechanism · 1

Mechanism◆

Conceptual integrity — the quality where well-designed software has no surprises, covers exactly the right domain, and everything fits together — is much harder to maintain with coding agents because features are added so cheaply.

Referencing The Mythical Man-Month, Simon explains that conceptual integrity — coherence and fit in software design — erodes with agents because the low cost of generating features leads to uncontrolled, incoherent growth.

transcript

Simon Willison: There's a concept in The Mythical Man-Month — conceptual integrity — where well-designed software has an integrity to it: there are no surprises in it, it covers exactly the right domain of things, everything fits together and makes sense. That's so much harder with coding agents, where you can have an idea for a feature, run a prompt, and five minutes later you've got the feature. Your software grows little weird bumps in funny different directions.

provides context · 2supports · 1

Example

The Winchester Mystery House is an apt analogy for what happens when coding agents make it trivially cheap to add features: software accumulates endlessly without coherent design, and decision-making becomes harder as complexity grows.

Claire Giordano's analogy — a house with 140 rooms built endlessly by a widow haunted by ghosts — captures how cheap agent-generated features lead to bloated, incoherent software where conceptual integrity breaks down.

transcript

Simon Willison: It's got 140 rooms, because the woman who built it was the widow of the guy who invented the Winchester rifle, and her psychic told her she'd be haunted by the ghosts of everyone killed with that rifle unless she kept building the house forever. So for 40 years she kept adding new rooms. That's exactly the problem with coding agents and software: it's very easy to keep adding new rooms, because the cost of adding those rooms is so much cheaper. What you end up with is something where the conceptual integrity falls apart — and then it's harder to make decisions about it.

explains mechanism · 1

Mechanism

Previously, the time-cost of writing code served as natural discipline that filtered out marginal ideas; coding agents remove that friction, making it far easier to justify features that would once have been dismissed as not worth the effort.

When code took a week to write, engineers naturally self-filtered out low-value ideas; with agents, the same feature takes an hour, removing the friction that once enforced design discipline.

transcript

Simon Willison: It all keeps coming back to discipline. It used to be that the discipline was enforced on you by the amount of time it took. You'd come up with an idea for a crazy feature and think "yeah, but that would take me a week — I cannot justify that, so I'll forget about it." If it takes an hour, it's so much easier to justify.

Highlight slides

LOC as a Productivity Metric✦ from: Lines of code can meaningfully measure productivity because there is a hard ceiling on how much production-ready code a human can write, and coding agents can substantially raise that ceiling.The Hard Ceiling✦ from: Lines of code can meaningfully measure productivity because there is a hard ceiling on how much production-ready code a human can write, and coding agents can substantially raise that ceiling.The Quality Caveat✦ from: Lines of code can meaningfully measure productivity because there is a hard ceiling on how much production-ready code a human can write, and coding agents can substantially raise that ceiling.Conceptual Integrity in Software✦ from: Conceptual integrity — the quality where well-designed software has no surprises, covers exactly the right domain, and everything fits together — is much harder to maintain with coding agents because features are added so cheaply.The Agent-Driven Growth Problem✦ from: Conceptual integrity — the quality where well-designed software has no surprises, covers exactly the right domain, and everything fits together — is much harder to maintain with coding agents because features are added so cheaply.

Related episodes

Video · Lenny's Podcast

OpenAI’s Head of Design: This is the best time in history to be a designer | Ian Silber

→

Video · Lenny's Podcast

Chatbots are not the final interface: OpenAI’s Head of Design on what’s next | Ian Silber

→

Video · The Cognitive Revolution

Intelligence on the Edge: Liquid AI's Ramin Hasani on the Search for Device-Native Foundation Models

→