I feel like the problem is that it's both. We're sanding off the long tail of human expression. It's not profitable this quarter, you see. Faster to let the AI do it.

It really feels sometimes like they were trained on too much short-form fiction or something. Really stunts their sentence and paragraph texture.

I don't see anywhere he was apologetic enough to fix any of the damages that he caused, and plenty of places where he was documented as doubling down on lies he knew were lies at the time.

The man was making money off of lying about dead children. Where's the comedy in that?

Some days I really wish the timing had worked out for me to build a serious private inference server before hardware prices spiked.

Given the LiteLLM Supply Chain Incident, threat mitigation and governance are definitely major concerns.

I feel like this is the kind of hack that made early radio tech exciting to play around with. The basic parts are incredibly simple to assemble from scratch, so it feels like magic. Speakers and microphones are the same thing in reverse. And so on.

More likely based in scale, in the sense of "they can't punish all of us."

I'm very mad about the RAM shortage.

It was quite rude of them to not wait until I built a new LLM server.

Getting dumped from "upwardly mobile middle class" to "unemployable underclass" does seem likely to be radicalizing . It's not clear yet how much it'll actually be happening, but it does challenge a lot of the traditional focus on blue collar workers as being the most up in arms about automation and labor.

One concern is that people in general might not have enough personal taste to care if something is slop. I have high standards and am familiar enough with LLM output to find that yet another repetitive pattern painful to read, but clearly that isn't universal.