Anthropic released fresh research on AI's impact on the labor market

Spoiler: programmers are first in automation.

Researchers introduced a metric observed exposure — actual, log-confirmed use of AI for work tasks.

So, the absolute top-1 among all professions is Computer programmers. AI really covers 74.5% of our tasks. And the most popular automated use case across the entire economy is “Write, update, and maintain software programs.”

It would seem time to pack up and go to the factory?
Not quite.

1️⃣ No mass layoffs.
According to statistics, the unemployment rate among highly exposed specialists has not increased since the end of 2022. If you are already a confident mid-level or senior, the neural network won't fire you. It makes you many times more productive.

2️⃣ But hiring juniors is over.
Anthropic looked at statistics for young specialists (22-25 years old) in professions with high AI penetration. And there, a collapse was recorded: the probability of finding a new job after the release of ChatGPT dropped by 14%. Meanwhile, in professions where AI is not applicable, hiring is steady.

What does this mean? 🤔
The neural network does not replace a tech lead. The neural network replaces the need to hire a team of junior minions under a tech lead to write boilerplate, tests, forms, and basic CRUDs.

The entry threshold for programming has skyrocketed. It is no longer profitable for businesses to pay for your six months to a year of training at the company's expense, tolerate mistakes, and review trivial code. Claude or Copilot produce junior-level code instantly, don't ask for health insurance, don't burn out, and don't leave after a year for a 2x salary.

The conclusion is extremely simple.
If you are already in the industry and still fundamentally do not use AI in your daily work — you are your own worst enemy. You are simply refusing legal doping while colleagues speed up their work.

And if you are just trying to break in — forget the strategy “I'll learn Python syntax and go as an intern.” Now you will have to compete not with fellow course graduates, but with a machine that has already taken 74% of routine tasks. You will have to start your career at a level that, just a couple of years ago, was required of strong mid-levels.

Learning has become harder. Working has become more fun. 🌝