I’m in a similar situation; I built my computer a few years ago, and now, 3+ years later, my machine would cost like ~50% more for similar specs. It’s nuts.

Like, I would be the target demographic for this, if hardware prices weren’t fucked right now. I fully intend to buy one, but only once the price drops in a couple of years, once the AI bubble bursts and chips get re-prioritized back to consumer hardware again.

If my computer died today, I’d probably try to buy a Gabecube to replace it, and install CachyOS on it to use as a desktop computer and Steam Big Picture mode for gaming with the controller. (Nothing against SteamOS — I love it on my Decks! — but I’d want something more focused on the desktop experience as my main computer.)

Exactly. And the few things GenAI does well aren’t that compute intensive and can run on local models.

LLMs are terrible at writing new text that matters, i.e. anything at all technical or important. But they’re good at reformatting content into properly structured English, editing text for grammar, and explaining everyday concepts. But they’re so good at those things, that a tiny model running on consumer hardware can do it.

So, I don’t understand how anyone can possibly expect LLMs to be a profitable business. Where are the moats?

In my experience, the trick to this is to get a list of changes to make yourself, rather than trusting the machine to actually get it right. (Unless you’re in a dev/test environment and don’t care much if it fucks it up.)

I’ve been thinking about this, and I think it’s likely mostly to do with cashflow. They won’t say that, of course, since implying any cashflow challenges is a massive red flag for analysts, but it’s the only thing that makes sense.

Basically:

1. Microslop is all-in on AI.
2. AI infrastructure is incredibly expensive to build, requiring intensive capital investment over years to build out capacity.
3. AI models are incredibly expensive to train.
4. Microslop’s gaming division also requires intensive capital over years to deliver games.
5. Microslop has finite cashflow available to reinvest in long-term bets.
6. Microslop has to choose where to allocate its capital investments, and gaming is less attractive to upper management than AI.

Because AI is the “it thing” right now, the reputational harm to Microslop for, well, being called Microslop because of their self-inflicted sloppening and the reputational harm from writing off all their investments in gaming is less than the perceived reputational harm they would face if they weren’t seen as being technology leaders in the AI space.

So, it all comes back to the AI bubble; investors are pumping up any company “doing AI”, so it’s become the target that Microslop cares about, and is redirecting all available cashflow in a mad gold rush to establish Copilot and their agentic OS, regardless of all consequences.

On the plus side, the world will be a better place with more investment in open software, so Microslop imploding will be fantastic… But that’s not going to do anything, at least not now, for the tens of thousands of laid off employees as they crash and burn.