AI Has Made Developers Lazy

I love wasting things. Right now I’ve turned the heating on even though it’s spring (and around these parts we don’t need it). I put boiling water in the Freezer for fun. I buy food and watch it expire.

When you use things to excess, you help the unit cost reduce. That’s why I’ve just smashed up my MacBook Pro, as I’d like them to be cheaper in the future.

That’s also why I just asked Claude what day it is. In 13 CMUX tabs. Don’t worry, I got an orchastrator to do it as I’m not a monster.

The Present

I just needed to make a single line config change. That means I decided to spin up a new agent, and got everything moving.

Now I’m in the same space as many software several software developers.

“I don’t really remember how to do half this stuff anymore”

I bet you’re coming to that conclusion too. The problem is that we aren’t reducing the cost of tokens, we’re reducing the duality of our thoughts.

Need some regex? AI writes it.

Need a SQL query? AI writes it.

Need to remember some obscure syntax from a framework updated every three weeks by people who think breaking changes are a personality trait? AI writes it.

The productivity gains are real. AI tools absolutely make many developers faster and pretending otherwise is denial at this point.

But there’s another side to this.

We’re quietly losing the friction that forced people to actually understand things.

I learned enormous amounts from debugging problems I didn’t understand. Most developers did. You’d spend hours reading documentation, experimenting, failing repeatedly, searching old forum posts written by someone called “xXDarkCoderXx” in 2009, and eventually the solution would click into place.

Now there’s no struggle. No deep understanding. No memory formation. Just throughput.

It’s a disaster

Side Effects

Interviews are difficult. Developers can build products all day long using AI but are then crumpled into a ball of insecurity when their intellectual crutches are taken away.

No wonder.

In the few AI-powered interviews that exist there are few who can pass them as the questions have been developed so as to invade the easy solutions presented by LLMs when wielded by the inexperienced.

AI is simply the logical conclusion of a process for the industry where we have degraded and reduced thoughtfulness over a long period. Software development has been degraded and reduced. Now software developers are doing it to themselves, outsourcing intelligence to an LLM — on a meter.

The Danger

Senior engineers exist because they spent years making mistakes. They developed intuition after breaking production systems, misunderstanding APIs, deploying bugs on Friday afternoons and suffering the consequences.

You can’t shortcut experience.

An AI can explain why your code crashes. That isn’t the same thing as the sinking feeling you get after causing the outage yourself while 40 people sit silently on a video call waiting for updates.

That emotional damage creates engineers.

I also think developers are becoming strangely passive.

I increasingly see people accepting AI output without challenge. If the code compiles and the tests pass then apparently that’s enough. Nobody asks whether the architecture makes sense, whether the code is maintainable or whether the generated solution resembles something created by an exhausted raccoon.

The AI confidence effect is real. Developers trust authoritative sounding output even when it’s nonsense. That simply can’t be a good thing.

Conclusion

The developers who thrive over the next decade probably won’t be the ones who memorize the most trivia.

They’ll be the developers who can:

  • Understand systems

  • Verify correctness

  • Spot bad assumptions

  • Ask better questions

  • Integrate messy business requirements into working software

  • Recognize when AI output is catastrophically wrong

In other words, software engineering may become less about typing code and more about judgment.

That’s what we should prize. That’s what we should be thinking about. That’s not the direction we are heading in (unfortunately).

About The Author

Professional Software Developer “The Secret Developer” can be found on Twitter @TheSDeveloper and regularly publishes articles through Medium.com

The Secret Developer suspects some engineers would outsource brushing their teeth to AI if somebody built the API for it.

Next
Next

AI Turned Every Developer Into QA