Is AI really the next big thing?

Day 54 / 366

I keep going back and forth in my mind about whether AI and LLMs are living up to the hype or not. Sure, on the surface it looks exciting, but I am sure that anyone who has tried to use them to build something substantial will share my doubts. Anything with AI these days starts as really promising but then fails to develop into anything useful.

Take AI-generated images for example. These models have been out for ages, and every time a new one comes out, there is hype online about all the lifelike images it can make. But then what? What’s a widespread application you know that uses AI images?

There were apps to make your own AI photos, which came and went. There are AI deepfakes out there which I don’t think is anything to be proud about. And then the most used AI pictures are used in low-effort posters and YouTube videos and reels, which are so clearly AI-generated. I sometimes use the latest DALL-E model to make banner images for these blogs and I am not that satisfied with the outputs.

A year ago, AI was supposed to wipe out most software developers. People said that AI will lead to 10 developers and that one person will be able to write code 10 times faster. Where is the new app that was built last year with AI?

ChatGPT, which led to the popularity of AI, is barely usable now. It refuses to do most tasks, and its code generation has gotten worse as well. As someone who is a ChatGPT Plus subscriber and heavily uses it to code, I can personally vouch for the drop in quality.

On one hand, I want to believe the hype and invest all my time in AI, since every big company in the world is heavily betting on it. But on the other hand, every time I try to use it for anything significant I realize that it’s just not ready yet. For now, it is just a cool trick to impress people.