A Remark from Laurie Voss

Software Cost Evolution

On 14 September 2026 Simon Willison posted a quotation from Laurie Voss on his personal weblog.

The post appears on “Simon Willison’s Weblog” and is sponsored by WorkOS, which promotes its auth.md product for agent‑based user registration.

Voss’s statement reads: “The cost of writing code collapsed, and the cost of reviewing, fixing and operating it is following, and I'm assuming it gets there.”

He continues: “What's left of making software is finding out what people actually want, defining it precisely, and making it pleasant to use.”

He adds a further observation: “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.”

The quotation captures a view that the traditional expense of hand‑coding applications is rapidly diminishing.

Interpretation of the quote suggests that as tooling, AI assistants, and low‑code platforms improve, developers spend less on raw code creation.

Consequently, the remaining expense shifts toward activities that cannot be automated, such as user research and experience design.

This shift aligns with the broader industry narrative that “we are all product engineers now,” a phrase Voss uses to describe the new focus.

The blog entry lists recent articles, including pieces on GPT‑6, Astra, and an OpenAI agents incident involving RubyGems.

These surrounding links indicate that the weblog regularly discusses cutting‑edge generative‑AI developments.

The page also shows tag statistics: “careers 83 ai 2,234 laurie‑voss 6 generative‑ai 1,980 llms 1,946 agentic‑engineering 63 deep‑blue 12.”

These numbers reflect the site’s categorisation of content and the relative prominence of topics.

The presence of a “Colophon” spanning the years 2002 to 2026 underscores the blog’s long‑term publishing history.

While the quotation is brief, its implications are amplified by the context of rapidly falling development costs.

Interpretation: As code generation becomes cheaper, organizations may produce a far larger quantity of software products.

Interpretation: The per‑product cost that remains is the effort to discover genuine user problems and craft pleasant solutions.

Interpretation: Because this cost does not amortise across multiple applications, it becomes the dominant labor component.

Interpretation: Voss predicts an unbounded demand for software, suggesting the market will continue to expand without a natural ceiling.

Interpretation: In such a scenario, the skill set of product engineers—combining empathy, definition, and design—will be the critical differentiator.

Implications for Product Engineering

Companies that invest in user‑centric research and experience design are likely to capture the most value in the emerging landscape.

Product teams will need to integrate rapid prototyping with continuous feedback loops to stay aligned with user intent.

Tools that automate code synthesis will free developers to focus on higher‑level problem definition rather than syntax.

Voss’s observation that “the cost of reviewing, fixing and operating it is following” hints that maintenance overhead will also decline over time.

Interpretation: As operational tooling improves, the total cost of ownership for each software piece may converge toward the user‑research component.

The quote’s reference to “infinity” of software underscores the expectation that niche, specialised applications will proliferate.

Interpretation: This proliferation will increase the importance of modular, composable architectures that can be customised per user need.

Businesses that treat every software initiative as a product‑engineer effort will likely achieve higher adoption rates.

The blog’s inclusion of related AI topics, such as GPT‑6, signals that generative models are already influencing the cost dynamics Voss describes.

Interpretation: As large language models become more capable, the gap between code generation and code quality narrows.

Ultimately, Voss’s forecast suggests that the competitive edge will rest on the ability to translate human desire into precise, delightful software experiences.

Why This Matters: As code becomes cheaper, the primary value in software shifts to understanding user needs, making product engineering the central activity.