But there's also DeepSeek, Qwen, Kimi, GLM, Jev, and countless services built on top of frontier or open source models. Those 3-4 companies might have the most numerous announcements that hit the front page, but there's way more than that going on in all directions.

The separation between the scientific and the mystical world is a very modern concept, and not everyone buys in, especially when they go down a rabbit hole. Science and philosophy have ridden in the same cart for a long time, long enough that America's oldest scientific organization is called the American Philosophical Society, because its name predates the word "Science" and the idea of that separation.

I'm an OpenCode user, but I'll fall back to Claude Code if I want to use Opus end to end for something, given my company has a subscription. But I'm not using yet another tool and subscription for a model that isn't even winning.

I don't have to imagine what it would be like under communism in order to see what it's already like under capitalism.

Jellyfin has no such solution that I can tell. Stuff like give them access over Tailscale is not the same user friendly option that Plex has. When there's an actual alternative for easily sharing with friends, I'll consider it. Til then I've had a lifetime Plex Pass for around a decade.

I feel like I've lived 3-4 completely different lives so far, but the constant is the ability to adapt to the next one, and still find joy in it while you're there. "Survival is the ability to swim in strange water."

Personally, the AI tools have been transformative for work, but haven't affected how I work much. I have always coded as a team. I'd often do the largest and most complicated parts myself, but work (both at work work and my hobby work) has always been about passing things between colleagues based on what our strengths were and how much time we had.

The AI tools are another colleague. They work incredibly fast, and I do less coding myself now, but my goal was always to solve the problem, not the code itself. The AI tools do a great job most of the time, but they sometimes screw up and need more guidance or me to step in to fix the thing (usually a very small error compared to the whole). If they screw it all up, we might need to start from scratch, or I might need to just do it myself. But that's not most of the time. And then I figure out the thing missing in my process to move them in the right direction, and improve it.

I feel like any software developer used to collaborating with, training, and mentoring other developers knows exactly how to work with AI tools, and the only main difference is how much effort I put into being really nice about it the whole time.

One main thing has changed. Before I would handle the most complicated problems people were having trouble with myself. Now I allow AI to work on them, even if I know I'll have to fix it. The difference is that I care about the time, the strain, and the morale of my coworkers. AI is just getting paid for iterating tokens, so I don't have to feel bad about what I ask it to figure out for me.

At Atlantic City airport I could park in an economy lot and just drag my luggage into the terminal. There were short security lines, and the airport was small enough that if anyone was running late for their flight they'd send them through first. I was able to trade having to rid an bus in the air for a much easier time going through the airport. IDK that anyone else will fill in there. Only other flights are American, just flying from Philly.

I'd actually be happy with this if they turned OpenCode support back on.