I was at the gym yesterday, talking with a relatively new CrossFitter named John, and naturally he asked me what I do for a living. I said I write software all day. But it hit me—I don’t actually write software anymore. I told him that, nowadays, Cursor writes all my code—I just tell Cursor what to do.
I was at the gym yesterday, talking with a relatively new CrossFitter named John, and naturally he asked me what I do for a living. I said I write software all day. But it hit me—I don’t actually write software anymore. I told him that, nowadays, Cursor writes all my code—I just tell Cursor what to do.
I’ve written about how code isn’t important anymore and that we won’t be writing code much longer, but somehow that truth finally hit me. And of course, when a truth hits me, someone else has already spoken that truth. Ruby creator David Heinemeier Hansson just a few weeks ago told us to “put down our pencils” at Rails World 2026.
The simple fact of the matter is that Cursor, Claude Code, Codex, etc., are better coders than we are.
This dawned on me over the past week at my day job where Cursor found a startlingly large number of threading issues in some code that I had written. I’m not ashamed—except for a handful of savants, we all have a very hard time reasoning clearly about complex multithreading. We just don’t have enough RAM in our heads to hold all of the possibilities in our brains.
But you know who does? Cursor. We use Cursor at work, and we have it wired into our pull request process to examine the code we are reviewing and see if it can find any problems. And boy, does Cursor find problems. In that code I wrote, Cursor found more than 20 different scenarios where race conditions could occur. And then, Cursor fixed them. In minutes. Although I felt a little dumb as a result, I was pleased that our agent found these issues before our customers did.
I think we all know that multithreading bugs are all but impossible to track down and fix when a customer reports them. Cursor? It found them and fixed them no problem. It was humbling, to say the least.
In context
- Topic: Desarrollo con IA — Copilotos y agentes en el ciclo de desarrollo de software.
- Source: InfoWorld
- Published: 07/10/2026
Continue reading at the original source →
Excerpt published automatically by the site radar. The full text belongs to its publisher and is linked above.
Why it matters
What is changing in engineering is not typing speed, it is the relationship between whoever produces and whoever answers for the result. An assistant can generate in minutes something nobody on the team fully understands, and that carries a cost that shows up in no productivity metric.
I distinguish between accelerating an experienced developer and accelerating someone just starting. In the first case the tool multiplies judgement; in the second it multiplies the lack of it. Adoption policy has to acknowledge that difference or the codebase fills with plausible, wrong code.
What usually goes wrong
What usually goes wrong is that review does not grow at the same rate as production. The team generates twice as much and still has the same two people reviewing, who get tired, start approving faster, and three months later quality has dropped without any indicator showing it yet.
What to watch
- What happens to code review: if the assistant produces more, somebody has to read more.
- What happens to tests — generating code without generating tests only moves the problem forward.
- Where the source code travels when the assistant is used, and what the contract says about it.
How I read this entry
If I were rolling this out, I would do it with written rules: what can be generated, what always gets human eyes, and what is off limits to an assistant — typically anything handling money, identity or personal data. Rules give freedom, they do not remove it.
This entry is an excerpt from the original source, selected by the site radar. The commentary above is the site's own and does not belong to the cited publisher.
Comments
No comments yet. Yours can be the first.
Living through this in your own team?
Open the chat and tell me how you're handling it. I'm interested in comparing notes.
