2026-10-29

This is a thought-provoking read:

Is our aversion to AI idealogical or practical?

TL;DR of the following wall o' text: I use AI, and I actually like it for some things. I also have concerns about how it was created and how we keep it sustainable in the future.

My thoughts, triggered by the article rather than in direct response to it:

I do have an ethical distaste of how the AI models were created, and I wish I saw a path where the creators whose content was fed into them were compensated. But I'm not idealogical at heart, so I can't find it in myself to turn my back on tools that are genuinely helpful just because their creators are shitty.

The idealogical purity test approach to AI feels doomed to fail, as AI is inserted into everything tech, even places where you may not identify it as such. (Like your phone's camera or swipe-typing) That ship has sailed.

We can work towards a world where we get models that are built on ethically-sourced content. We can work toward a world where models run more efficiently, on more efficient hardware, reducing the climate impacts of their use. I want these things! But I really think we are past the point where the AI genie can be shoved back into its bottle. It's too damned useful.

I was a huge AI skeptic. The first time I tried it at work, it wrote shit code and lied to my face. No way I would ever want to use it for anything that mattered, and using it for something that didn't matter felt like a huge waste of resources.

A couple of years on, the tooling has improved to the point where I find it genuinely useful, and it really does help me get more done in my day. (I'm a software developer.) And my company is really pushing us to find ways to use it better (not really more, just better). This work feels like a worthy use of the computing power it takes up. It's not frivolous or creating content that will only hurt people. I've long found real meaning in my work (I love building tools); now I can spend my time on the parts I really enjoy (figuring out how to solve a problem) and less time on the parts I just accepted (actually writing the code to turn the plan into reality).

I have concerns about how this will play out long term. How do you train up a new generation of developers who have the skills and knowledge to be able to properly guide their tools? How do you keep a company afloat in an environment where every customer suddenly has someone on their staff advocating for a home-built app to replace their business software? How do we manage the expansion of computing resources in a way that is fair to communities and investors both?

None of those concerns go away if I just refuse to use anything that is "too AI". All that would do is make it harder for me to keep this job and make me less qualified for the next one I will need to find.

2026-10-28

I made this chicken and rice skillet, and it was delicious! A tad oily, probably because I used more oil than needed (I added more before sauteing the onions, which in hindsight I really didn't need to do). The chicken cooked through perfectly, and the rice and onions were delicious. I'd dice instead of slicing the onions next time.

https://cooking.nytimes.com/recipes/777998043-scallion-chicken-and-rice-for-two

The rice cooks to the bottom of the skillet, but the food has a lot of moisture so it will stay warm while you clean the pan. Hot water into the hot pan, scrape everything loose, and then wash and dry.

This one is a keeper!

2026-10-25

Y'all, I've been driving for decades, and I don't think I've ever seen one of those semi trucks that cars slide under in the movies. I feel cheated. All the trailers around me sit way too close to the ground for fun sliding shenanigans.