Laurie Voss: AI shifts software work from coding to product judgment
On 14 September 2026, Simon Willison posted a short entry on his blog quoting a passage from Laurie Voss's piece "We are all Product Engineers now". The quoted argument: the cost of writing code has already collapsed, and the cost of reviewing, fixing and operating that code is following the same path; Voss writes that they are assuming it eventually gets there too. What remains of making software, in this view, is finding out what people actually want, defining that precisely, and making the result pleasant to use. Voss argues this remaining cost is attached to each individual piece of software and does not transfer from one project to the next, so as the total amount of software produced grows toward infinity, which Voss expects because demand for software has no ceiling, that non-transferable cost of understanding and shaping a product becomes the entire job. Willison's post consists of the quote, the source title, and the byline.
Key facts
- Simon Willison quoted Laurie Voss's essay "We are all Product Engineers now" on his blog on 14 September 2026
- Voss argues the cost of writing code has collapsed and the cost of reviewing, fixing and operating code is following the same trajectory
- Voss states what remains of making software is finding out what people want, defining it precisely, and making it pleasant to use
- Voss frames this remaining cost as per piece of software and non-transferable, so it grows to become the whole job as the volume of software produced increases without a ceiling on demand
Why it matters
The excerpt captures a specific version of a widely discussed claim about AI and software work: that AI narrows the gap in the cost of production tasks, coding included, while leaving one activity comparatively untouched, working out what a piece of software should actually do and how it should feel to use. Voss ties that activity to each individual product rather than to a skill that scales across many products, which is why the argument singles it out as the part of the job that does not shrink.
Who it affects
The piece speaks to people who build software, described in its title as "product engineers": developers, and anyone else involved in deciding what a piece of software should do rather than only how it is coded. Willison's readership, a technical audience that follows AI and software development commentary, is the immediate audience for the quote itself.
How to use it
The excerpt is an argument, not a tool or a technique, so there is nothing to install or configure. The practical takeaway Voss draws is where effort belongs going forward: on identifying what users want and shaping it precisely and pleasantly, rather than on the mechanics of writing, reviewing or operating code, which Voss expects to keep getting cheaper.
How solid is it
The passage is Voss's own stated assumption, not a measured result. Voss writes explicitly "I'm assuming it gets there" about reviewing, fixing and operating code catching up to the collapsed cost of writing it, and offers no data, timeline or supporting evidence within the quoted text for either that assumption or the claim that demand for software has no ceiling. Willison's post presents the passage as a quote worth reading, without independently verifying or contesting it.
Risks and caveats
The quoted passage gives no timeframe for when the cost of reviewing, fixing and operating code is expected to reach the collapsed cost of writing it, and no mechanism for why the total amount of software produced would go to infinity beyond the bare assertion that demand has no ceiling. The source gives no background on Voss (role, employer or affiliation) beyond authorship of the quoted piece, and does not include the full original essay or a link to it, only its title and this excerpt.
“The cost of writing code collapsed, and the cost of reviewing, fixing and operating it is following, and I'm assuming it gets there. What's left of making software is finding out what people actually want, defining it precisely, and making it pleasant to use. That cost is per piece of software and doesn't transfer, so as the amount of software goes to infinity, which it will because there's no ceiling on demand, that cost becomes the whole job.”
— Laurie Voss, We are all Product Engineers now