'You Won't Need to Learn to Code.' What I See as an Engineer Is Different.

A tech leader compares AI to the printing press. From where I sit, it looks less like freedom and more like a door closing.

Share

Earlier this year I read a conversation between two big names in tech about what AI means for programming.

One was Boris Cherny, a technical leader at Anthropic and the creator of Claude Code. He said: "In a year or two, you really won't need to learn to code." He used a grand historical comparison — the Gutenberg printing press. Scribes were freed from tedious copying to do more creative work like binding and illustration, and over 200 years literacy rose from about 1% to 70%. The implication: AI will free programmers the same way, from Git chores and dependency management, so we can focus on "the really fun product thinking."

The other was Lenny Rachitsky, the well-known Silicon Valley writer and product person. He was excited too: people without a technical background can now build things. You used to spend hours on Stack Overflow; now you just ask AI.

When I finished reading, something felt off. The gap between that picture and what I see as a software engineer was just too big.

What I see isn't "freedom." It's "clearing out."

Boris describes a nice future: senior engineers combine knowledge across fields and do higher-value work. But what happens to the people who aren't "senior enough"?

Look at what's actually happening in our industry:

  • Amazon cut about 30,000 corporate roles in two waves between October 2025 and January 2026 — reportedly the largest corporate layoff in its 30-year history, close to 10% of its roughly 350,000 corporate staff.
  • Block (the company behind Cash App and Square) cut about 40% of its staff in February 2026, from over 10,000 to under 6,000. Jack Dorsey wrote plainly that AI tools are changing how companies run, and that "smaller teams, using the tools we're building, can do more and do it better." The stock jumped almost 25% on the news.

Executives explain these cuts on earnings calls in plain words: AI lets us get the same output with fewer people. Translation: basic programming work can be automated, so the people doing basic work aren't needed.

The future Boris describes — senior engineers creating value across fields — rests on one assumption: that you get to become a senior engineer first. But entry-level roles are disappearing in bulk. Young people are losing the doorway in.

Was the printing press really that gentle?

The printing-press story leaves out a few things.

"The scribes were happy to be freed" is survivorship bias. Most scribes didn't gracefully move into illustration; their skills simply lost value. And the press took some 200 years to spread. Society had time to adjust.

Today? Tens of thousands of jobs gone in months. That isn't a transition. It's a generation's career path closing fast.

Then there's the idea that you don't need to understand what's underneath. Boris has said that about 80% of Anthropic's code is now written by Claude. He also shared a story of an engineer who built a Go service over a month but "still doesn't really know Go" — and he sees that as a good thing: if the result is right, you don't need the details.

As someone who writes code, I know the "tedious work" is exactly how you come to understand a system. When AI digs that path away, it looks like freedom, but it's deskilling. Your bargaining power moves to whoever controls the AI black box. The person who "can ship a Go service without knowing Go" looks efficient — and is easier to replace.

The split

That's the split I mean. Same industry, two completely different pictures:

What the tech leaders describeWhat I see as an engineer
AI lets programmers focus on "fun work"Entry-level roles disappearing in bulk
Cross-disciplinary work creates more valueYoung people can barely get in the door
Slow, gentle change, like the printing pressYears of layoffs packed into months
"Not needing the low level" is freedomDeskilled people are easier to replace

Boris's story has a hidden premise: technological change is always good, and if you adapt, you benefit. But technology is never neutral. When he says "you can do more interesting work," he assumes everyone gets a seat in that room, and that the market will make it up to the people who are replaced.

In reality, a Harvard study covering 62 million workers found that after companies adopted generative AI, junior developer positions fell 9–10% within six quarters. Big-tech hiring of new grads has reportedly dropped by half over the past three years. Some companies say it outright: "If an AI coding agent is cheaper, why pay $90,000 for a junior engineer?"

This isn't growing pains. It's a structural door closing.

The printing press didn't bring the Renaissance by itself. Without real buffers — real retraining, a social safety net, antitrust oversight — what it brings may just be faster unemployment.

And this time, the ones being optimized away aren't factory workers. They're software engineers, the people who used to think they'd found a safe harbor.


Notes on sources: Amazon layoff figures from 36Kr and Commercial Times reporting; Block figures from multiple media reports; the junior-role and new-grad hiring figures from the Harvard study and industry statistics. Originally written in Chinese on March 2, 2026. The author is a software engineer.