A study of 52 junior engineers found that AI-assisted coders finished tasks 8% faster but scored 50% on a follow-up quiz versus 67% for those without AI. The gap came from worse debugging skills, suggesting AI speeds up familiar work while eroding problem-solving ability.
For thousands of years, humans have built tools to make hard things easier. A hammer does not make you worse at building. But is AI different? Because AI helps your brain, not your hands. And that raises an uncomfortable question: what happens to the skills we outsource? Is AI brain rot real? What should we do? Well, I see a big difference in the media headlines and actual real-world studies. So, let's see an actual experiment with a group of 52 junior software engineers split into two groups, one that used AI for coding, and one that didn't. I have two questions here, and the results on both kinda surprised me. One, did AI make them faster? Well, the AI group finished their tasks a good 2 minutes, or about 8% faster. So, they had time to watch a Two Minute Papers episode, which is never 2 minutes. But then, wait a second, the result is not statistically significant. I'll downgrade this a bit, so did AI make people code faster? Maybe. Now, second question. Did the AI make them dumber? Now, hold on to your papers, Fellow Scholars, because the answer is kinda yes. After having them take a quiz, we get 50% for the AI people, and 67% for the no AI people. Oof, that is almost a 2-grade difference. Huge difference. And this is statistically significant. So this one would be much harder to dismiss as noise. I gotta say, this is a fairly large effect for such a smaller study. So, where is the largest gap? The AI group really underperforms? When exactly? Well, in debugging. Yes, you get an AI to write a piece of code, and at some point, there will be a problem. And if you rely on the AI too much, sooner or later, you will lose the capability to fix something that goes wrong. I would be careful with that. And I find this study informative because it tells you how much worse it is, and in what areas. So, what should you do going forward? Well, in my opinion, 3 things. Dear Fellow Scholars, this is Two Minute Papers with Dr. Károly Zsolnai-Fehér. One, if you use AI, use it mainly for automating and speeding up things you already understand. It is excellent for that. Two, for things you don't know yet. Ask questions. Keep your mind sharp. That is the way of the scholar. Three, when something breaks, don't just ask the AI to fix it. No! First, try it yourself. Then, have it explain to you what you missed. You will learn so much more. We are scholars here, so let's be measured. This is not an ultimate study. No! This is not the final word on this question. This is 52 mostly junior developers, one Python library, one short task, and one quiz. Oh, and it has a chat-style assistant, not a full agentic coding system, which would probably make this kind of difference even larger. So, plenty of limitations apply, but as a pointer, I feel it is quite valuable. TL;DR: if you use AI as a tutor, it will sharpen your mind. I use Lambda to reproduce AI research papers, often in minutes. It's also great to train your own models or fine-tune an existing one. Run inference or text to image or video? Easy peasy. Running a Deepseek chatbot or agent? Super fast, super reliable. Lambda gives you powerful NVIDIA GPUs to run your own experiments. I test ideas from the papers I cover, and moments Later. Results. Love it. Seriously, try it out now at lamda.ai/papers.
