HTMX creator Carson Gross posted a judgment this week via Simon Willison's blog: even if AI tools continue to explode, programmer remains a career you can do until retirement. But we noticed his reasoning isn't the usual "AI isn't smart enough" — it's something more structural. Programming was never about typing code.
What this is
HTMX (an open-source framework that lets web pages deliver interactivity without writing JavaScript, expressing dynamic behavior directly through HTML attributes) creator Carson Gross is viewed in developer circles as a standard-bearer for "anti-over-engineering." His central argument splits into two things: using computers to solve problems, and managing complexity while solving them. The former AI can already help with a lot; the latter, he believes, AI is still nowhere close to handling.
Industry view
The remarks drew polarized reactions on X. Supporters cite the Stack Overflow 2026 survey: 72% of senior engineers say AI tools have made them more efficient but haven't changed the nature of the work. The opposing voice comes from Anthropic CEO Dario Amodei's earlier prediction — within 3 to 6 months, 90% of code will be written by AI. What's worth flagging is that Gross's argument is entirely framed from the programmer's perspective; for product managers, designers, and operations roles — people who "don't directly write code but live off technology" — this logic doesn't automatically hold.
Impact on regular people
For enterprise IT: the hiring logic for in-house software teams won't change short-term, but pure-outsourcing vendors and low-code providers may accelerate being displaced by AI.
For individual careers: separate "writing code" from "solving problems." The former's barrier is dropping; the latter's capability is becoming more valuable. This is the watershed for deciding whether to send your kid to a coding bootcamp.
For the consumer market: those "learn Python in 30 days and land a big-tech job" courses out there — you can mentally cut the price in half before enrolling.