LLMs might be making us better communicators

There's a real question worth sitting with: are LLMs making us dumber?

My honest take — probably yes, in some narrow ways. If you're leaning on models to write code, you're likely losing some of that low-level fluency. That's a real tradeoff.

But I think the other side of this gets ignored.

Using LLMs well forces you to communicate clearly. You have to articulate what you actually need — not a vague gesture at a problem, but a precise description of it. And then you have to read back what the model gives you, evaluate it, and respond. That's a feedback loop a lot of developers never had.

Many engineers I know lived inside their heads. Brilliant at thinking, not great at expressing. Working with LLMs every day is quiet practice in the thing they avoided.

Maybe that builds more empathy. Maybe it makes people better at explaining their thinking to teammates, not just to machines.

It's not black and white. Some skills atrophy, others sharpen. Worth being honest about both.