I absolutely agree. The issue is mostly unskilled people who attempt to do what skilled people do except they do not possess the discipline, curiosity, patience, and cognitive capacity to actually learn. It is what annoys me by far the most in today's era. Too many retards who are fundamentally mediocre and below average, or have no knowledge whatsoever and yet claim with utmost confidence that they know what they're doing. All they do is take shortcuts every step of the way instead of doing that which is inconvenient, but necessary, in order to actually become good at something and understand it. We are seeing the exact same phenomenon occurring in the Monero community with too many users vibe coding AI slop and services not lasting more than a few months. AI is useful, if used in certain ways, but it also enables spoon–feeding on every level. Before AI, you had users on stackoverflow holding people accountable at least to some degree by discouraging it, but in the present day era, AI just gives users what they want at any time. It's like a parent just giving their young child candy whenever the child wants it. The child learns absolutely nothing and just becomes more and more entitled as everything is just handed to it.
As a community, I think we should take a stand on vibe coded AI slop and begin to discourage it. There are times when it is useful when it is about saving time or automating processes, but it should not replace logic. In the end it's the logic that really matters and that requires critical thinking skills, precision, and experience, which is something AI simply can't do. Initiation, intent, and logic need to be handled by an experienced dev.
Given the logic behind AI, I like to visualize it in four quadrants. You have an axis labeled "precision" and an axis labeled "cost" with the latter including money, time, and other resources.
Low Precision / Low Cost: This is where AI excels. Precision isn't needed and the cost is lower than what it would otherwise be if you hired a person or did the job yourself. Examples include rapid prototyping, proof of concepts, branding for small business, generic stock photos, copy editing.
Low Precision / High Cost: Some things have a high cost associated with them, but don't require high precision. A good example would be actual art, not the generic shit. It's not precise and doesn't have to be, but it costs a lot of time to produce. Using AI, you would have to roll the dice over and over again until you finally get what you want and even then it can be a long shot. At this point, you are better off just doing it yourself or having another human do it. AI can assist at times with ideation or retouching, but it can't complete the work in its entirety.
High Precision / Low Cost: Some thing are simple, such as looking up drug dosages and yet precision is required. AI being prone to mistakes can easily mess up the dosage, which could be fatal. AI can be useful here as an aid, but that's about it. Experience and education are absolutely needed and proper data management isn't optional.
High Precision / High Cost: This is where the money is, especially in fields like engineering, security, and the sciences where precision is extremely important and the cost is high. Unfortunately, this is where AI is not only not useful, but counterproductive and will create severe risks of its own.
Or summarized:
Low Precision / Low Cost: Significant use cases for AI
Low Precision / High Cost: Some use cases for AI
High Precision / Low Cost: Some uses cases for AI
High Precision / High Cost: No use cases for AI