It reduces cycle time, if you can ask, “how long will this take” or “is this really hard to do”. The ai has much more context on the world than we do, because it’s basically indexed every book, web page, and video on youtube. Your time to iteration on learning drops significantly, because you can essentially guess, baed on the AI’s feedback, how long something will take to accomplish
E.g. One I ran into when trying to install Hyperland on PopOS. A cursory serach of all the major AI providers would show you, it’s going to require 10+ hours in setup time. That’s assuming your good at linux, and understand all the problems in setting it up. Grok AI saved me a bunch of time, by just pointing me to Sway instead.
It also expands your intelligence, it having so much more context, and from language perspective, can tell you to invest a lot of time into something, or just pivot to something else.
This also requires good prompting from the user side, but that’s not hard. Just tell it “ask me questions about the thing I want to learn, so you and I can get more context” and sometimes those questions save hours and hours of time going down a particular rabbit hole you hadn’t considered.
Summary
If you’re trying to do autodidactic learning, using AI as index for what is a good idea or not, can really expand your intelligence. It also serves as a great way to reduce cycle time on development.