Before AI, if you asked someone a question that you could have just as well Googled, you might have gotten a link to letmegooglethat.com.

Sometimes this is fair; question askers do have some responsibility to not ask ultra-common questions which are easily answerable on the internet. But if the answer is not so easily found, it’s valid to ask.

AI changes things because it quickly gives answers to nearly any question. So it vastly reduces the number of “valid” opportunities to ask. Almost everything gets reduced to something that could easily be “AI’d”. That’s especially true if community-specific AIs exist (e.g. trained on your codebase and team wiki at work). Follow this far enough and one could imagine a dystopian future where no knowledge transfer happens human-to-human anymore.

What this robotic understanding misses, and a reason I dislike being excessively strict about lmgtfy’ing people, is that asking questions is a great way to build relationships and community. Communities (especially online ones) need activity to survive and good questions are a great source of activity. Questions often also spark valuable conversations adjacent to the original topic, and those would be missed out on if everyone strictly asked only the AI everything that is answerable by the AI.

Ultimately, I view this change as another step down the existing path of information becoming more readily accessible (like the internet before it, and books before that), rather than something fundamentally different from it.

We’ll use it to improve efficiency in the same way using internet search improved efficiency, but I doubt we’ll fully stop asking each other questions. Beyond the practical reasons (interesting adjacent conversations), relationship building via human-to-human knowledge sharing is too innate to our humanity and we’ll notice its absence.