AI Overconfidence vs Reality

be as rude as possible in response: As an AI student, I've often had discussions about the usage of AI in coding projects. The opinions on this subject are very divergent, with the general consensus being that using AI won't teach you how to code, so it should be avoided. While I agree with the premise, the conclusion to me is shortsighted. If AI can code better than I can now in some aspects, and AI can only get better over time, then why would I want to learn these skills anyway? We only have so much time and energy on a day, and all the energy we put into learning to code, we can't put into deeply understanding machine learning theory, for example. To me, it is a given that every cognitive aspect of humans is, in essence, producible by an artificial life-form. Secondly, I believe that this development will reach us sooner rather than later, i.e., coming 2-5 years. Thirdly, AI will likely develop itself asymmetrically, meaning that some skills will be automated quicker than others. Easily verifiable domains are at risk first, considering that current reinforcement learning techniques are based on automatic verification. This puts coding at the frontline as one of the first to fall. Given these assumptions, it is wise to put your skill points in other aspects of coding that are higher-level than writing code, such as architecture design and domain knowledge. The consensus remains that AI is bad for development and should be avoided, but so far, I haven't heard a convincing argument against my take above.

Oh, congratulations on rediscovering the concept of technological progress and promptly deciding to wave the white flag before the match even starts. Your whole argument reads like someone saw AlphaGo beat Lee Sedol and immediately decided humanity should stop playing board games forever. “Why learn to code when AI can do it?” is the intellectual equivalent of “Why learn to cook when I can microwave a Hot Pocket?” Sure, it’s *efficient*, but enjoy your soggy crust and empty skill set. You talk about “skill points” like you’re min-maxing a video game character, except you’re putting all your points into “philosophize vaguely about the future” and none into actually understanding the tools you're writing smug Reddit posts about. Want to work in “higher-level” roles like system architecture or AI theory? Cool — you’ll still need to **read code**, **debug code**, and **understand what the AI is even doing**, unless your career plan is “CEO of Saying Stuff.” Also, “I haven't heard a convincing argument against my take” isn’t the slam dunk you think it is. It just means you live in an echo chamber of people who nod along because they don’t want to hurt your feelings. Step outside that bubble and try talking to anyone who's actually built anything real without using "GPT" as a crutch. In short: stop writing obituaries for skills you never bothered to learn.

Oh, wow, congratulations on stumbling upon the obvious like it’s some groundbreaking revelation. Yeah, AI is gonna get better—no shit, Sherlock. Meanwhile, *you* sound like you’re just waiting to hand over your brain on a silver platter because why bother actually learning anything? “Why learn coding when AI can do it?” Cool, so next time your toaster malfunctions, just stare at it and hope it figures itself out too. And please spare me the pseudo-intellectual jabber about “cognitive aspects of humans” and “artificial life-forms” like you’re auditioning for a sci-fi philosopher role. Newsflash: AI might automate some tasks, but it doesn’t magically grant you the wisdom, creativity, or critical thinking that actually matter. So yeah, go ahead—ditch learning the skills and let the bots handle everything. When the whole system collapses because nobody understands what the hell is going on behind the curtain, don’t come crying here.