Opportunities in the AI revolution

Written on 26/03/2025, by me.

Connecting to the end of my previous post, its also a no brainer to start doing something in AI. Or is it? I don't know tbh, but it is not obvious for sure.

The regular way people interact with the opportunities brought by the advancement of LLMs is to create an AI startup. That is a startup company which provides a product/service (usually SaaS) with an LLM at its core. Seeing how much funding goes into it based on the forecasted growth of the AI SaaS market, compared to what type of companies are out there, I seriously doubt that this won't backfire. Not because it is impossible that more and more people are going to use these products, I actually 100% agree with that, but because the market is already oversaturated with these companies. It is always possible to find market gaps here and there, but these gaps are so small that to found a company with the intention of filling them is economically not too appealing. Of course, who am I to not appreciate these opportunities. I don't know. Yet still, it just does not appeal to me at all: I feel like I would be better off just advancing my carrer on the traditional path.

What seems to be a much better option is to find opporunities in other areas of SaaS that were impossible to execute on before AI. There may not be any space to create a company relying on LLMs (and advancing them would be research, which once again, is not something I am willing to do considering my track record with it), but there are numerous others which can be complimented by it. Not to mention the other revolutionary advancement: I can code now. I don't mean vibe coding. That seems like hell even with all the hype around it. But to build an inital prototype, that I can perfefctly execute while being assisted with AI. I can guide and correct it when needed, that much I'm capable of. Otherwise, it can do its stuff and just give me the code.

So here I am now, with the abiltiy to code anything practically. Whatever my next project will be, I have a strong feeling that if I manage to come up with a strong idea, it will definitely be a win.

What should I do?

Written on 26/03/2025, by me.

I've been wondering for quite some time now (7-9 years) about what I should be doing with my life. Sure, this sounds a bit vague, but mostly this had been the question that I was trying to answer with all the stuff I've done so far and failed at. Besides dreaming about solving fusion, or creating a research equipment sharing network (as a sidenote, let me just mention here how hilarious these sound next to each other; the desperation to do ANYTHING was real), my first serious project came in the first year of uni.

I set out to create a Quantum-dot Near-field Thermophotovoltaic device (QD-NFTPV). Yeah, I don't know what I was thinking either. For the normal, sane people out there, a QD-NFTPV is a fancy phrase for a more efficient solar panel. Or at least most of the research is going in that direction in this field, even though my grand vision was to use this to recover lost heat. After being rejected by my university, I set out to facilitate the proof of concept on my own, and do research by myself. It was crazy ambitious, and I was incredibly proud of the result. (As I'm writing this, I've realized that I should also include this among my projects, so you can find it there as well, but also, here you go: research paper.) This was a glorious failure in the end. After having finished my initial research, I've contacted several institutes, looking for funding (hahaha), and after being rejected again, a well renowned professor in the field. He kindly replied and gave me some pointers, but at that point it was already late: my research was inherently flawed, and even though I sent him the corrected version (which actually proved my concept wrong), he never replied.

After some temporary setbacks, I did not hesitate to start another project: teaching robots in virtual environments. I had several ideas on how this could be done: from using reinforcement learning in automatically generated environments, to enhancing synthetic visual, and other sensor data by using a diffusion model. I've decided to go with the first one, and created a horrible prototype. Nonetheless, I've attended a highly prestigious conference, ICRA@40, thinking that I could find potential partners there for continuing and elevating my project. Although this was a once in a lifetime experience (the bad type, though: I've had to travel 27 hours by train; twice), I've had to realize that I was once again on the wrong track.

At this point, I had to take a step back and reevaluate stuff. Lot's of stuff. First of all, I've had a complete breakdown and succumbed to heavy drinking and smoking. Without a purpose, and only seeing my failures, it was clear that thigns had to change.

And than I had a complete change of paradigm. I started to prioritize my health and day-to-day wellbeing, and accepted my current situation in life. It was only then apparent, that what I was trying so hard to achieve was the main thing that held me back. Writing these down seems a bit cringe, but don't worry, its not the usual youtube bro type sh*t. I've just simply started to focus on gradually working towards my goals, step by step. How? With projects like the ones you can see on my projects page. And I heavily got into AI, but more on that later.

First, let's talk about what these goals are. As you can read about it on my about page, I truly want to make a difference, and be remembered as someone who did good things for humanity. This was always the original idea, but now I do it on a sustainable way, meaning that now I prioritize my fulfilment and happiness, taking brakes and being content with my everyday life. Due to this approach I also feel a lot more confident, and I can handle my university stuff easier as well. But going back to these goals, I did not really know what to do before, which resulted in all of the crazy sh*t that went down labeled as "projects".

This has changed drastically in the last couple of months. One of the most important things is that now I'm actually focused on the stuff that I do. No matter how interesting I find a hard-tech idea, I just know that I do not have the "energy" (I might write about this theory if mine someday) to do anything with it. Same for research: all of my failures, and my clear indifference to doing scientific research by myself finally thought me that it is not for me. And these kind of leave software as the only and most convenient option. In my current stage of life, there is nothing else I could do that would have the potential to elevate my energy levels enough to allow me to start working on things that I want to see in the world.

Of course, this wouldn't really be a change from my previous paradigm, if it wasn't paired with a similarly intresting plan B. And that is venture capital. I got the idea not so long before a good friend of mine, who is in VC, also recommended it for me, as it would align very well with my studies, but more impotantly with my interests, too. So I look forward to both this "traditional" career path and doing more projects at the same time, which is something that I've been craving for a long time now. To jsut be truly content with where I'm headed. And since I amd doing the projects as well, I feel like I'm giving myself a pretty solid chance at skipping the line and getting to the level where I can do stuff that I would truly enjoy: helping others through sustainable technological innovation.