懐疑者へ、AIは確かにプログラミング業界を変えている

Twitter co-founder Jack Dorsey is now the CEO of Block, which runs payment services like Square and Cash App. On Thursday, he announced plans to lay off more than 4,000 workers — 40 percent of the workforce — and Block’s share price soared.

“Something has changed,” Dorsey wrote in a tweet. “The intelligence tools we’re creating and using, paired with smaller and flatter teams, are enabling a new way of working which fundamentally changes what it means to build and run a company. And that’s accelerating rapidly.”

Block CEO Jack Dorsey. (Photo by MARCO BELLO/AFP via Getty Images)

The announcement hit a nerve because it seemed to confirm public fears about the impact of AI on white-collar work. A widely read essay from Citrini Research last weekend predicted that AI-driven progress would drive wave after wave of layoffs.

Earlier this month, author Matt Shumer made similar claims in a viral blog post called “Something Big Is Happening.” Shumer argued that disruption has already started in the software industry. Here’s how he described being a programmer today:

I am no longer needed for the actual technical work of my job. I describe what I want built, in plain English, and it just... appears. Not a rough draft I need to fix. The finished thing. I tell the AI what I want, walk away from my computer for four hours, and come back to find the work done. Done well, done better than I would have done it myself, with no corrections needed.

He predicted that AI agents will soon come for other white-collar jobs.

“AI isn’t replacing one specific skill,” he writes. “It’s a general substitute for cognitive work.” In Shumer’s view, this means that lawyers, financial analysts, writers, radiologists, customer service representatives, and many others can expect their work to be automated.

“Nothing that can be done on a computer is safe in the medium term,” he concludes. “If it even kind of works today, you can be almost certain that in six months it’ll do it near perfectly.”

It’s hard to predict what models will be able to do in the future, so I don’t know how soon LLMs will automate the work of lawyers or financial analysts. But as a journalist, I can talk to programmers to see if their experience today matches Shumer’s dramatic description. For this story, I talked to more than a dozen software industry professionals — programmers and their bosses — about how AI agents are changing their work.

AI really is making programmers more productive

I learned that Shumer is exaggerating the pace of progress in software development. It’s not true that AI agents consistently produce production-ready software from a single prompt. Human programmers are still needed to make big-picture architectural decisions, write detailed instructions, and verify code after it’s generated.

But Shumer (and Dorsey) are right that something big is happening.

“I worked at Google for years and managed lots of people,” said Understanding AI reader Jim Muller. In his post-Google life, Muller has been writing software for two small companies he co-founded with his wife. He has made extensive use of Claude Code, which he likened to “a particularly reckless and nutty junior-level engineer.”

Despite that unflattering description, Muller believes Claude Code has dramatically increased his productivity. Even a reckless and nutty engineer is pretty useful.

I also talked to a manager who oversees a team of 20 programmers at a non-profit organization. He estimates that over the last year, coding agents have helped his team more than double their productivity — at least as measured by the number of software updates (known as pull requests) they submit each month.

But he also pointed to some downsides of the new approach.

Read more