From my observation, the people that are the most excited about AI are low skilled/unskilled people in that domain. If said people treated AI as a learning tool, everything would be great (I think AI can be a really effective teacher if you're truly motivated to learn). The problem is those people think they "now have the skill", even though they don't. They essentially become walking examples of the Dunning-Kruger effect (the cognitive bias where people with limited knowledge or competence in a particular domain greatly overestimate their own knowledge or competence)
The problem with being able to produce an artifact that superficially looks like a good product, without the struggle that comes with true learning, is you miss out on all the supporting knowledge that you actually need to judge the quality of the output and fix it, or even the taste to be able to guide the agent in good patterns vs poor patterns.
I'd encourage people that are obsessed with cutting edge AI and running 5000 Claude agents simultaneously to vibe code a website to take a step back and use the AI to teach them fundamentals. Because if all you can do is prompt, you're useless.
>I think AI can be a really effective teacher if you're truly motivated to learn
This! I use simple freebie gemini queries and actually read the code produced because thats the actual intent of me asking gemini a question.
Now that doesn't mean that I mind folks running these vibe coding agents and I bet in the hands of an invested individual aware of risks, some percent of those agents could be tests on results, "devils advocates" type things, like "search for security problems in this code", "break into this system" etc, don't get me wrong I haven't ever actually used agents so my assumptions of what they can do are probably pretty naive lol but theres no denying that neural networks aren't going anywhere.
My biggest problem with this seem to be slanted more toward the messed up economics and politics of it all which is really just the usual clueless people encountering and using tech from the clued people with a dash of actually psychopathy tossed in for that extra spicy spicy hit. Witness my favorite bad guys suggesting an actual ban on states issuing regulations regarding AI, and then when I see the money being borrowed and the ramifications of budgets required to repay them etc and the fraud, social media AI slop and actual deceptive possiblities that are most certainly being used, well its just all so tiresome
From my observation, the people that are the most excited about AI are low skilled/unskilled people in that domain. If said people treated AI as a learning tool, everything would be great (I think AI can be a really effective teacher if you're truly motivated to learn). The problem is those people think they "now have the skill", even though they don't. They essentially become walking examples of the Dunning-Kruger effect (the cognitive bias where people with limited knowledge or competence in a particular domain greatly overestimate their own knowledge or competence)
The problem with being able to produce an artifact that superficially looks like a good product, without the struggle that comes with true learning, is you miss out on all the supporting knowledge that you actually need to judge the quality of the output and fix it, or even the taste to be able to guide the agent in good patterns vs poor patterns.
I'd encourage people that are obsessed with cutting edge AI and running 5000 Claude agents simultaneously to vibe code a website to take a step back and use the AI to teach them fundamentals. Because if all you can do is prompt, you're useless.