It has been quite a while since my last post, and a lot has changed in the world since this random guy from the Midwest failed to pursue a hobby in game development. From politics to policy to technology, so much has been going on its hard to wrap your head around it at times. For this post, I want to focus a bit on the technology side of things.

First, a quick update on me. I still work at the same company in the same function. I’ve had to learn so much more about being an effective manager, but I am proud to say that I’ve managed to retain my same team despite fluctuations in the market abroad. I’ve also managed to secure a majority of my team promotions, which speaks volumes about their talent.

Outside of work, I’ve wrapped up a friends D&D campaign, started my own (running “Curse of Strahd”, a beloved classic!), and continued my journey to improve my cooking skills. I’ve managed to put on about 10lbs over the winter which I plan to work off ASAP. I’ve leveled up once again as an Uncle as I now have a grand total of one nephew and *two* nieces, all super cute. I’m still single but finding it harder and harder to mingle in my 30’s. I’ve also had to learn a lot more about home improvement, helped others with their home improvement, and put a big focus on trying to host or go to in-person events when I can.

I truly feel that since Covid, the desire for in-person interaction has skyrocketed dramatically. To put it lightly, many became stir crazy during Covid from the heightened social restrictions and in that time built up more reliance on social media, content creators, and other online resources. In doing so, I fear some people have forgotten the importance of in-person social interactions and how critical it is to focus on your local events and politics. Too broad a view and too much access to unverifiable information would drive anyone crazy on a day to day basis. Likewise, most of these online spaces tend to be giant echo chambers, where the minority of users share one view point and the majority are just bots.

On the topic of broad availability of information online, one thing that has remained a dominant topic is the emergence of “AI” and large language models. I’ll be honest: I was a huge skeptic at first. You have to keep in mind the AI craze followed in the footsteps of the previous tech craze, block chain. That turned out to be used for more nefarious purposes than valid ones (at least to a random consumer like myself), and as such I could only view the emergence of AI and LLMs as some new money grubbing ploy (this may yet still be the case).

In practice, though, these tools have turned out to be truly revolutionary. From coding, to research, and even things like art and planning, these tools have an almost unlimited potential in what they can do to drive creative people toward their goals more easily. I cannot stress this enough just how enthusiastic I am about some of these things. To frame this in my perspective, allow me to give you an example that is very close to home and relevant for me.

In a previous blog entry, I briefly talked about dipping my toes into game development. There were parts I loved: Learning a new language, a new engine, and coding specific behaviors. Behaviors only get you so far, though, and I completely fell flat at the thought of creating sprite sheets, doing animations frame-by-frame, and other tedium that gets in the way of iteration. Realistically I could have ignored a lot of that and focused on the core game loop, but for me, I struggle to enjoy a game without seeing the cool stuff in real time.

This was just 2 or 3 years ago, mind you. Just the other week, I saw a person online ask an LLM to generate a 2D sprite sheet for a character performing an animation. It knocked it out of the park, and it wasn’t even a complicated prompt. I tried doing this myself back then as a beginner, and a task like this took me hours not just to imagine what a sprite could do, but actually coding and drawing it to do that thing in a loop. These tools are still in an infant stage too, and could yet still mature much, much more in the future.

There has been a lot of discourse on the subject of LLMs generating art and images. I empathize with the artists out there, whose works were almost certainly used to train on, but I also empathize with the people who are excited for the technology. I think it helps to frame the conversation around a few things I believe are fundamental truths.

Firstly, the best artists out there will always have an edge on coming up with unique and interesting content that is full of soul. I do not think LLMs will ever replace these people or their creativity, and I personally think physical works of art might see a large uptick in popularity. Secondly, people underestimate how much these tools can further augment independent artists out there today. Think how quickly your favorite animated series might get released if the artists working on could focus more on creative decisions rather than idling on tedious post production work or color correction. Similarly for independent filmmakers, image and video models can help small teams create state of the art effects without billion dollar studios.

This democratization of creative work is what really excites me about the potential of these models. The idea that, if I try to dive back into game development later this year or next year, I might be able to get a prototype out in 1/100th of the time. The idea that local artists, filmmakers, and software creators might be emboldened to create their own studios to produce world class works at fractions of the cost. The idea that the choke hold massive corporations hold over creative works will crumble and we might start seeing a swing back to bold, unique works of art that fit into niches that were previously impossible to make profitable.

It would be prudent of me to mention that the things I discuss above are thanks in no small part to the work of DeepSeek. They made an incredibly bold decision to open source a state-of-the-art model which laid the cards down for all other companies in the world. America, in standard fashion, had been shoring up all of its tools and knowledge under one or more large corporate umbrellas, closed for all. DeepSeek changed all of that, and slowly but surely these companies have been realizing that in order to compete they must follow suit. I’ll be honest, this has been an absolute pleasure for me to watch. I have grown tired of these large companies absorbing smaller companies until they become too fat and too arrogant to produce anything of value for consumers. Every day more and more people are starting to see the effects of enshitification in real time, both in software and physical products. If LLMs can allow independent studios and creators to take back large chunks of the market share, sign me up.

And that’s where I am at right now. I am sitting back and absorbing the news each day to see where things wind up. I remain optimistic for the future, and I truly do hope more of the AI revolution continues to be open sourced. In the mean time, I am focused primarily on my old home. It needs a lot of work, and there is a long list of things I’ve been putting off learning how to do to improve it.

For those who see this, thanks for reading. If you are someone I know and we haven’t spoken in some time, please reach out, I would enjoy catching up. For everyone else, you can find me on twitter if you search hard enough (though the site punishes those who don’t cough up money). I always enjoy meeting real people, and I’d be happy to chat more about anything. Stay safe out there, and support your local business owners.