It is opposite for me than what the author describes. With AI I can explore new areas far more effectively. For example, I have never been great with front end web stuff, but asking LLMs to generate something and then looking at the code they built is pretty cool.

Of course there are errors and gaffes and weird bits all over, but that helps keep me in the loop and understanding it better.

I can now (more) boldly go where I feared to tread before.

With all due respect, what you're describing doesn't quite spell learning in my opinion. Sure, its better to read the code of an application you envisioned to learn a programming language rather than an unrelated convulated program; but you're still offloading your thought process to the LLM. You're not quite seeing why it chose A instead of B, and why it uses that struct instead of the other. And honestly, the LLM doesn't quite know why it chose that either.

I feel the same empowerment. LLMs allow me to tap into my second-order knowledge.

I agree with the exploration aspect of AI. I don't have it create code for me but I have certain longer threads about topics that I think are interesting but would not have approached without it. Previously you would've had to find an entire textbook or class on a subject that may not have been quite what you were looking for if you were just interesting in poking around in a topic to start. Like I have a long running chat about making databases and WALs that I don't think I'd have started if I had to dive fully into a course on writing a database from scratch (if that even exists).

Being able to ask the dumb questions but then get focused on something more technical is so nice. Get examples on demand.

It helps me pickup and understand new things pretty fast.

Same here. It really helps me get deep into things that would have been daunting to start due to the boring research work required. I have heavy ADHD and high intelligence (not to brag, it is often more of a burden than a blessing!) so I can understand things really quickly but I have no patience for painstaking research or planning and documentation. I just want to build, to create and make my vision come to life but I'm really one-sided about that. I don't care about how it fits in the organisation process, the governance or where it should be documented.

And when I need to learn something new I hate doing courses because they're always too slowly paced for me. All courses are paced for the slowest possible participant. Especially those from the US where trainers have this "training voice" where they speak slowly and articulated and usually non-skippable which makes me scream in agony. My brain can't slow down to that level, I just lose focus then. It's like driving a ferrari at 5km/hr. Especially the non-IT courses are so slow and constantly repeat everything, as if I didn't fully understand them the first time, argh. So I tend to spend my time on these awfully designed mandatory HR courses trying to fuck around in developer mode and re-enable the skip segment features :) The more boring the training the more I learn about moodle and its frameworks.

I just love to dive right in and just do something but I often skip that first analysis of what method would be best. And I tend to stick with what I know. Which is often not the best way in a greenfield project. And often the boring first steps are so awful for me that I just put it off. If I do start I often do it half-assed. Then I run into a wall, have to regroup etc. I learn from that (mistakes are always a learing experience), but it's slow going because I take the wrong direction first.

AI really helps me get past that first step meaning I actually start it instead of procrastinating the research forever and move onto the next thing. And I can do great things because I know what I want and how I want it. It's a real enabler for me especially something like cowork. It helps me widen my horizon efficiently instead of wasting my time and energy like online courses do. It picks up things that my weak points are (patience) and feeds me information in a highly dense format I can easily digest. Helped by custom prompts of course ("don't tell me I make great decisions, evaluate everything on its merits, just tell me the meat of the content, don't elaborate"). I still work along with it and take ownership of its output, I'm way too much of a control freak to let the AI do everything without understanding it. But it gets me there quicker and it avoids procrastination by skipping the boring steps. I ask it to suggest methods and techniques, compare their pros and cons so I fully understand what I'm selecting for.

I have to say this was not really the case with early simple question/answer LLMs. The hallucinations were too bad, the searches too limited. The knowledge too hit and miss. It would come up with wrong stuff or overlook the best options. But the new high-context agentic technologies like cowork that really investigate the source material for the purposes I provide, they really are a game changer for me.

For example, I recently asked it about an obscure error I'd been having. Previous simple LLMs would just collapse into a directionless troubleshooting mode, "Did you try...?". Cowork actually went and retrieved all the relevant code from github without being asked to do so, and evaluated what would have happened in the background for that message to appear. Really good to see because it's exactly what I would have done. But I would have taken much more time to do it.

So no, for me it's not making me lose savviness. It's helping me grow it in exactly the way that's efficient for me.

I think the difference is that you read the code and get curious about it rather than glancing at the output and immediately telling it what to correct.

Is reading example code uncool?

How large are your projects? What kind of?

Just curious. No right or wrong answer possible.