Raw LLM Responses

​@EmmyEmmyEmmyEmmyEmmyEmmyI disagree. You dont "need", true, but you should. Because it will give you advantage knowing how a particular system parts work internally. Low level skills are important if you want to perfect the higher level ones. Take for example something outside programming, like some sport, boxing for example. If what you said is absolute truth, boxers would not need to understand how to punch. Their coaches would right away show them how combinations are done, and they would train that. But they would not understand fundamentaly how to put power into those shots. How power generation works, because a good punch consists of many things you have to learn first and practice, now just if you can throw mutliple of those in a fight and knock someone out. So the more you abstract your skills to a system, the less control you will have over what happens with your project, and of the direction its heading. AI is fine as a tool, but you still have to explain to it what it should do, even in your coded modules in order to create consistency between those modules in your system, if you even insteucted it in the first place to apply any kind of SOC principle to your code. And another thing is that today AI isnt perfect, its context window isnt big enough, it halucinates sometimes even newer models, makes mistakes, and if you dont know how to setup unit tests, if you try to add upgrades to your systems in the future or make some fixes with your code, it might break other parts of your app if you initially didnt set up unit tests, logging etc. And how do you expect a new inexperienced programmer to do all that, and more, using only AI, if he doesnt know how to program, and not just code but many more things. I just mention a few. Its like 10% you need to know if you a building a bigger system, especially alone. AI cant do planning for, should not do that, you need the ability to do it, and you cant if you dont understand how programming of these big systems works internally.