My stance on Generative AI

TL;DR: I like it as a tool, but I don’t like the way most people are using it.

Generative AI. Generative AI. Generative AI. It’s the talk of the town, it’s what people are talking about, whether that’s hating on it or using it to do everything for you is another story. This blog post will be very opinionated, normally as a personal blog post would be, and my option is leaning towards the hate trend. (Note: This post is mainly about LLMs)

Why I would like Generative AI

The current way that I use it, I use it when I want help designing technical stuff, such as for database schemas and comparing various design choices. And of course, I’ll still be making the final calls. I also try to minimize the use of generative AI whenever I can.

I can understand how using generative AI can be addicting because it’s easy to get comfortable with doing less, with the fact that you have someone to help you think about things. But I’d love it more if I could completely get away from generative AI and to think, research, and ruminate about everything clearly.

Why I don’t like Generative AI

A lot of people rely on it too much, and just stopped doing research on their own entirely. I have met a person with this exact behavior and it’s not exactly plesant to see. I was helping them set up their network, which I also said “Feel free to ask me any questions”, but he mostly did not ask me and kept asking Google Gemini everything.

I was also working on a group project, where the members did not understand the concept of feature creep and kept adding features. Instead of reducing, They just implemented it all with generative AI. The website worked but… The HTML file was 7000+ lines long. And if they want to modify it, they can’t, because “I have to wait 6 more hours for my Claude rate limit to reset.” Meanwhile, I also made a design in Figma and implemented the prototype in Nuxt, but it was, depressingly, thrown into the drain and the group moved to using Canva and Anthropic Claude as their main tooling. The group members see me as a person who likes to work alone, but from my perspective as a person who started doing things first, it is rather that they chose not to work with me.

Some may argue I should’ve just join their train of vibe coding, but it simply goes against my values. All they care about is putting subscriptions into their app and violating licenses by using content from Wikipedia without attribution and locking a feature based on OpenStreetMap behind the paywall, while I make my content all available in CC BY-SA 4.0, code licensed under the GNU GPLv3, and make them question how will I ever be rich.

For people outside this whole AI hype thing: vibe coding refers to using purely generative AI in making projects. See Wikipedia.

So, I don’t actually hate Generative AI, but mostly the way people are using it to think less and trusting it like it’s their only source of truth in the world. If you want to learn how to code, please, please, please. Please don’t end up using AI to write all your code. Actually learn how to it properly and it’ll be worthwhile. Believe in yourself. If you keep asking AI to do things for you, please try to reflect what you’re getting out of it, and what you learned.

Accusations

Maybe I scroll on Twitter too much, but I really don’t like it when people accuse others of using an LLM to make their stuff — such as for using em dashes. Sometimes it’s true, sometimes they’re just accusing people who knows how to properly use these symbols. (I certainly am not one of those people)

Anything other than text

I just don’t like AI-generated images or videos. It just looks uncanny in some way.