htmx Creator Carson Gross Believes in the Future of Programmers: Two Pillars of the Profession

Carson Gross, the creator of the htmx library, has boiled programming down to two points and built a forecast on them: AI tools will not put an end to the profession. Simon Willison pulled a quote from Gross's essay "Yes, and..." on htmx.org into a separate post, and it spread quickly through professional circles. The statement is short, but it fits neatly into the debate over who needs developers when models write the code.

Two tasks instead of a list of technologies

Gross defines programming through two things. First: solving problems with computers. Second: learning to control the complexity that arises along the way. The definition mentions no languages, no frameworks, and no "code" in the sense of a pile of lines. He describes the craft in terms of outcome and discipline.

A cautious conclusion follows. Gross writes that he finds it hard to imagine a future in which the ability to solve problems on a computer and keep the complexity of those solutions under control is valued less than it is today. In his view, the profession will therefore remain viable after the arrival of AI tools. This is a bet on the durability of a skill, not a prediction about the number of jobs or salary levels.

Why the focus on complexity

The second point in this formula is more interesting than the first. Language models handle local tasks well: a function, a test, a migration, a configuration. But every generated snippet adds to the codebase, and any codebase over time accumulates dependencies, conventions, and exceptions. Who is responsible for keeping the system understandable a year from now, when half the code was written from prompts? Gross's answer: the person who knows how to manage complexity. As long as a model does not take on that responsibility entirely, demand for such a person remains.

The author's background matters too. Gross is known as the creator of htmx, a library that builds interfaces around hypertext and deliberately reduces the amount of client-side code. For him, complexity is a constant enemy, and his position looks like an extension of that philosophy. The original essay is not reproduced in full in the source material, so we limit ourselves to what the quote says.

Key points

• According to Gross, programming consists of two components: solving problems with computers and managing complexity.
• By his estimate, the value of these skills will not decline in the foreseeable future.
• He treats AI tools as a complement to the profession, not a replacement for it. The essay's title, "Yes, and...", points the same way.
• The thesis is framed as the author's personal judgment and is not backed by labor-market figures.
• Simon Willison published the quote with the tags computer-science, careers, and ai.

What developers and teams should do

The main practical takeaway concerns training and hiring. If the value lies in framing the problem and controlling complexity, then it makes sense to invest in architectural thinking, reading other people's code, decomposing tasks, and the ability to say no to what is unnecessary. Models suggest syntax and APIs quickly, but the decision about what to build in the first place and how to maintain it is still left to a human.

For managers, this is a reason to revisit their metrics. The volume of generated code says little. It is more useful to look at how quickly a new person gets to grips with the system, how much time a change takes, and how many dependencies the last release added. For beginners, the advice is simple: use AI tools, but work through every decision so that you can explain it without a prompt.

Gross's quote does not prove that the labor market will withstand the strain, and he does not insist that it will. He offers a criterion: as long as a developer is required to solve problems and keep complexity in check, the profession is needed. That criterion will have to be tested in practice.