Hello World
Hello there!
This first post will serve as an introduction to my blog and website as a whole, as well as my plans for the future.
How we got here
First, I’d like to introduce you to my Notes and a bit of the story behind the site.
Notes are cheat sheets I write while learning new things, to help me quickly remember what I have studied after a long time has passed and reference it when needed. They are impersonal and direct, meant to be an introduction to the topic with some examples.
Before creating this website, I used to keep these notes to myself, but as I started to share them with friends, I realized that they could benefit other people, too. This is why I created trude.dev, so I could hopefully help others with topics I struggled with in the past, and also encourage myself to have a set of polished notes I can share as soon as someone needs them.
As my collection grew, I realized that not all notes are created equal. Some will never be outdated, since the fundamentals of computer science will always be the same, while others are more fleeting. For example, the way we use AI today is completely different from how we used it even just 6 months ago. I keep my notes constantly updated, as I reference them for my own work, but for some topics, I didn’t want to delete those old notes just to replace them, since there could still be some value in them.
I created the Journal section to host these temporary notes. The archived notes still follow this philosophy, and this worked for a while. I kept my notes updated and added new notes and posts whenever I had the time, between projects and after work hours.
As time went on, though, I noticed that I wasn’t enjoying writing these notes anymore. They are helpful, but I always try to make them as short and compact as possible, and I found myself wanting to share the projects I’ve been working on, and more personal things, too (like an actual blog), instead of just tutorials. And with the rise of AI tooling, people often default to asking an LLM for help instead of browsing websites. The traffic on my site fell to a tenth of what it used to be, and all my notes, together with thousands of others, were used to train new models without permission.
My stance on AI
I don’t blame people for using AI tools; even I use them quite a bit. They are an incredible technology that empowers us to do things we wouldn’t otherwise do on our own. LLMs are great at processing large amounts of text, prototyping, creating websites quickly, automation, programming, and of course, learning. An LLM can teach you almost anything, so long as you can fact-check its output. With agents, you can even have an LLM control your computer and do some real work for you: product research, email parsing, news feeds, maintenance, just to name a few.
There are also many things I don’t like about AI. The unethical way how LLMs are trained, scraping the internet while destroying websites such as mine in the process (be it overloading them with traffic, or copying our content), harming open-source by flooding maintainers with bad PRs, the massive power demands to train and run these models… the list goes on. AI is used for harm just as much as it is used for good.
Still, no matter if we like or hate AI, it won’t go away anytime soon, so we must adapt. LLMs, in their current state, are just tools. Like any good tool, they can be used to enhance one’s capabilities to do great things, or to cause harm. I see AI as a ‘force multiplier’. If you were going to do good things, then AI will help you achieve better results, and vice versa. The LLM isn’t to blame if it is used incorrectly.
LLMs have a price, though. Tokens, hardware, power, and attention. Not every task needs to be assisted by AI, but sometimes it is worth it. So, if a task is either cheaper or more efficiently done by an LLM, I think you should use it; Not using it at all would limit what you can achieve. But I don’t think AI should be shoved into every product, or overused to the point where one stops using it to learn, and instead uses it as an excuse to stop improving. AIs cannot replace you, because you are human. There are things that only you can do, because you have a unique perspective that an LLM does not have.
With that out of the way, here’s how I use AI, as of the time of writing: I use AI to help me research, prototype, and write code. I use it as a learning tool, and fact-check AI’s claims, often visiting the source websites to support the authors who wrote them. I have also set up my own AI agent, testing lab, and local models, and love to research the technology. I use AI for automation extensively. However, NONE of the content on this website is AI-generated. I do not, and will never, use AI to write blog posts, notes, or any other content on this website, as that would go against what trude.dev aims to be: a curated and polished list of notes and a way to express my own story, projects, and experiences. Every note and post takes me hours to write, but that’s the point. If AI wrote my blog, then it wouldn’t really be mine anymore.
Plans for the future
Earlier I mentioned using the Journal section for “fleeting notes”, but also wanting to make the blog something more personal. To accommodate both, I decided to start over, archive my old posts, and write it more like an actual journal.
Every time I work on a new project or research something worth sharing, I’ll write a blog post about it. There won’t be a fixed theme or schedule; I’ll write about whatever I’m working on when it’s ready. Most posts will be about programming, AI, and computer science, but I may include posts related to 3D printing, health, electronics, and anything else I happen to research and want to write about. I will also continue to keep notes up to date and write new ones.
I hope you find some value in this site. If you have criticism or suggestions, please send them my way at [email protected].
Thank you for reading.
