TL;DR:
I use AI, but when I don’t feel like using it, it’s because I enjoy the process and want to learn. I don’t mind taking my time.
Do I use AI?
Yes. I use T3.chat to access various models for only $8 a month, I use Deepseek with OpenCode for “vibe-coding” for a few projects, and I use Cursor to iterate on features or bug fixes.
Background
I have been programming since 2017 with Python, which is where my username “Rython” originated from.
I developed various projects and was seen as an expert at programming by my peers. In reality, I was a novice compared to industry expectations. I was able to hack things together, but I did not follow protocol; the code I wrote was not maintainable.
Before the public release/experiment of ChatGPT, I was already fascinated with generative AI after playing around with the Bloom API. It wasn’t a chat interface nor anything like a chat bot, it was closer to a fancy auto correct.
Since then, I’ve used ChatGPT, used it for coding work, and expanded to other tools like Cursor and OpenCode as mentioned in the first paragraph.
Where do I stand?
This question isn’t black and white. There’s many use cases of generative AI, and there’s a lot of ways people use it.
What I am against
I am against the encouragement of AI usage. I don’t believe people are falling behind when they don’t use AI, in fact, they are the most future-proof type of people, as they gain skills on how to build without relying on AI. The less AI-dependent someone is, the more skillful they are at planning, developing and debugging — the more useful they are in the industry.
I want to emphasise on the stream chatters who pressure streamers to use AI to get it over with. Even when a female streamer is clearly trying to learn, unhelpful chatters say “Just use AI!” and it is infuriating.
I am against video generation with AI (OpenAI Sora, Google Veo 3). I believe that it uses so much computation with so little reward, there is no valid use case of it, especially commercials.
I am against image generation with unethical source of data, or with the intention to replace artists. None of my emotes are AI generated, which is a bare minimum. Adobe Firefly can be used to save graphical artists’ time by easily removing something from the foreground of the image and having the background be seamlessly edited out. If AI is used to replace or disrespect artists, I highly condemn the usage of it.
Twitch requiring streamers to opt-out of using streams as Generative AI training is a very terrible move. It is especially embarrassing that Dan Clancy, the CEO of Twitch, refuse to speak on the livestream about it, and throw other people under the bus for it. It is a selfish move for the platform and a terrible image for Twitch as a whole.
In this section, I want to express my deep annoyance towards people who rely on AI for software development, namely, people who:
- Blindly do AI contributions to a project for the sake of creating a PR.
- Create portfolio websites with very little creative input.
In a capstone project, students are required to make contributions to a project, commonly through PRs. Using AI to spot a problem, then using AI to create a commit and pull request WITHOUT understanding the project is absurdly common in Universities and it’s wasting project leads’ time.
When you create a portfolio website, it is meant to be how YOU want to present yourself to other people digitally. Creating a soulless, AI generated website with generic texts and blue-purple gradients is the pinnacle of laziness and incompetence.
Environmental Impact
While AI does have an environmental impact, a lot of claims have been blown out of proportions.
A lot of water is used for cooling while both training and prompting. The water quality near data centres are poorer and residents nearby are affected. Data centres drive up prices of electricity (and possibly water bills) for nearby residents too.
There are claims that drinking water in the future will look polluted, that is blatantly not true.
What I tolerate
As for use cases where I tolerate:
-
Alphafold: Using generative AI (similar to the process of AI image generation) to predict and create protein structures, for medical purposes including vaccines.
-
Gemini Notebook: Formerly known as NotebookLM. This tool uses AI to tailor notes, generate a podcast to allow students to better grasp a topic. I was in a class where the professor didn’t speak clearly, and the slides provided were very vague and more visual than descriptive. As a result, I looked to this tool to achieve a good score in the class.
As mentioned, I use AI tools. I try to avoid using it when I’m learning.
I use AI tools when I feel desperate to have a certain process to be completed, so I can enjoy the fun parts. Such as developing the UI, figuring out integrations or just wanting to quickly get the project done so I can do my own side project for fun.
Conclusion, where I stand
As Tom Scott says: “Everyone draws the line of what’s acceptable just beneath what they’re doing themselves.” This heavily applies to myself, a software developer who often use AI tools.
Do I encourage everyone to use AI tools? No. I want people to learn.
Do I use AI tools? Yes, it is more productive when doing tasks I’m not enjoying.
I do not believe that my opinion is right, nor should it be a standard. I am just expressing my opinion,
I use AI tools, I am alright with people using AI tools, as long as nobody is getting harmed directly.