I may be old-fashioned and look like an elderly person grumbling about how things are worse now than before, but I really don’t share the hype around AI.

Yeah, it can be used in numerous interesting projects. Where you’d need to hire a lot of people and establish and streamline a lot of processes to achieve what can be achieved by training a model. But these are rare cases and the whole thing is actually being supervised by people who know what they are doing. Who has full control and can fix or tweak the output. But that’s not how AI is being marketed right now.

They tell you it’s capable of anything, on any scale, in any field. And everyone is trying to jump onto that train in the hope of making profits. But the actual results are not that great, actually. Yeah, sorry, I don’t have any links to support any of the statements I make here. And I don’t care, it’s my personal blog, and you are free to stop reading it right now. Yeah, AI can help you with something simple, non-critical. But the moment you need any kind of precision, you better be careful.

If you need to supervise things and start being a reviewer of all the output you get, then what’s the point of even starting using such a tool? OK, when we talk about scale, I can see why, as I mentioned earlier. But if we talk about daily tasks or some routine tasks, then there’s almost no win. You spend almost the same amount of time on reviewing the output as you’d spend making things on your own. And this way of doing things doesn’t save you from making mistakes, overlooking something. So it doesn’t actually save time. And if you rely on it too much, it can even add more work at the same time, making the resulting quality worse.

And AI is being pushed into our daily lives from all directions. Everything has some sort of AI thing going on with it. Clothes produced by AI, health-related things processed by AI, your calendar is managed by AI, etc. And what’s bothering me here is that all this is driven solely by money, not by a sincere desire for a greater good or a brighter future.

And what difference does it make if it’s used to help you code or write an essay or perform surgery? You can die because a model mistakenly made a cut in the wrong place. You can lose your job/house/business because a model made a mistake in the calculations or did not get right some wordings and produced an output that completely contradicts the meaning of the processed data. You can lose your digital assets because the code contained a bug which was missed, because the team was too AI-enabled.

In all these hypothetical cases, people lost their ability to process information, to learn, to proofread, because they are being told that AI is the ultimate tool to resolve any kind of problem one may have. But that’s not true. Never in my life would I like an AI to operate on me. And neither do any CEO of an AI-related business.

Another thing is the amount of resources that are being shoved into this voracious monster. It’s insane. Imagine a person that has no idea what’s happening on the other side and innocently prompting to generate a raccoon in a funny hat, riding a bicycle. And then correcting this prompt several dozen times, just for fun. I can’t justify the amount of electricity and water wasted on this stupid shit. Are these just associated costs for the time being? Is it temporary while everything is not yet optimised properly? Ok, but still the total useful output of this worldwide AI usage is ridiculous.

AI, as it is, is here and there’s no point in rejecting it. But we could at least stop overusing it so blindly and stupidly.