For sure. The more specialized or obscure of things you have to do, the less LLMs help you.

Building a simple marketing website? Probably don’t waste your time - an LLM will probably be faster.

Designing a new SLAM algorithm? Probably LLMs will spin around in circles helplessly. That being said, that was my experience several years ago… maybe state of the art has changed in the computer vision space.

But this learning is also value.

Without playing around with it, you wouldn't know when to use an LLM and when not.

Honestly I think this is the primary explanation for why there is so much disagreement on if LLMs are useful or not. If you leave out the more motivated arguments in particular.