Comment by hintymad
1 month ago
It looks there's a difference this time: copying the details of other people's work has become exceedingly easy and reliable, at least for commonly tried use cases. Say I want to vibe code a dashboard, and AI codes it out. It works. In fact, it works so much better than I could ever build, because the AI was trained with the best dashboard code out there. Yes, I can't think of all the details of a world-class dashboard, but hey, someone else did and AI correctly responds to my prompt with those details. Such "copying" used to be really hard among humans. Without AI, I would have to learn so much first even if I can use the open-source code as the starting point: the APIs of the libraries, the basic concepts of web programming, and etc. Yet, the AI doesn't care. It's just a gigantic Bayesian machine that emits code that nearly probability 1 for common use cases.
So it is not that details don't matter, but that now people can easily transfer certain know-how from other great minds. Unfortunately (or fortunately?), most people's jobs are learning and replicating know-hows from others.
But the dashboard is not important at all, because everyone can have the same dashboard the same way you have it. It's like you are generating a static website using Hugo and apply a theme provided on it. The end product you get is something built by a streamline. No taste, no soul, no effort. (Of course, the effort is behind the design and produce of the streamline, but not the product produced by the streamline.)
Now, if you want to use the dashboard do something else really brilliant, it is good enough for means. Just make sure the dashboard is not the end.
Dashboard is just an example. The gist is how much of know-how that we use in our work can be replaced by AI transforming other people's existing work. I think it hinges on how many new problems or new business demands will show up. If we just work on small variations of existing business, then quickly our know-hows will converge (e.g. building a dashboard or a vanilla version of linear regression model), and AI will spew out such code for many of us.
I don't think anyone's job is copying "know-how". Knowing how goes a lot deeper than writing the code.
Especially in web, boilerplate/starters/generators that do exactly what you want with little to no code or familiarity has been the norm for at least a decade. This is the lifeblood of repos like npm.
What we have is better search for all this code and documentation that was already freely available and ready to go.