Quoting Paul Ford
Simon Willison’s Weblog Simon Willison
Opinion — commentary, not a factual news event.
Paul Ford says AI can write code, but it’s also making bad coding easier. His point: the real edge still comes from humans working together, not just from machines.
Based on reporting by Simon Willison’s Weblog, Simon Willison — read the original for the full story.
Summary, retelling and take written by AI under human oversight; images are AI-generated illustrations. How we work · Report an error
Paul Ford thinks the doomsday version of software work has cooled off a bit. For a while, it looked as if developers like him were finished, pushed aside by tireless robots that could simply out-code everyone else.
That hasn’t quite happened. What’s become clearer, in his view, is that the hardest software still needs people who can think together, divide up the work, and actually use their craft well. AI can produce strong code, but that same ease also lets people do somebody else’s job badly.
And that, Ford argues, is part of why so many projects fail. The tool can generate output. It can even generate impressive output. But if everyone can code now, the market is also finding out why many people shouldn’t, or at least why typing code is not the same thing as building software that holds up.
His takeaway is pretty blunt: the industry is slowly relearning that software quality is still a human problem. Not every task gets solved by making the machine faster. Sometimes it just makes the mistakes arrive sooner.
My take — AI-written commentary, not fact-checked reporting
This is the most boringly true thing in AI right now: output is cheap, judgment is still expensive. The industry keeps pretending code generation is the same as software engineering, which is how you end up with a lot of confident mess.
Read more about this at: Simon Willison’s Weblog
