Well, I bought 1 year of GLM’s coding plan for basically nothing ($50?) last year.

…I now see its many hundreds of dollars, and its API is so busy it gets throttled.

Seems like devs are catching on.

FYI, for anyone interested in this stuff, I would suggest 1 year of Xiaomi’s coding subscription.

MiMo 2.5 is fantastic, beyond what benchmarks show. It’s a genuinely useful general purpose model with better prose/world knowledge than others. It’s quite uncensored. It’s not just some deep fried agent, and it’s dirt cheap right now because no one has realized it yet (kind of like GLM 6 months ago).

And before you ask, I use local models too. I sometimes run a custom IQ3_KT of MiMo 2.5, in fact, which is how I figured out it’s good. self hosting is the way. Gestures at the forum we’re on

But for stuff where privacy isn’t so critical, it’s still nice to have a year of an unquantized, fast API of a huge model for like $50.

I agree with this.

That lines up on the technical side, too, with how LLM “intelligence” is plateauing, cost of cheaper models is decreasing, where open weights are going and such. Q3 2027 seems about when an OpenAI coding subscription makes no sense.

That being said, I’d be wary of the “Facebook effect.”

Once a service gains a huge foothold, it can deteriorate for a long time without going away. Especially with regulatory capture. And for many smartphone users, OpenAI is the only AI they know.

Well, that’s because Zuckerberg felt FOMO and fired the Llama team over one bad experiment he probably pressured them into in the first place (as Llama 4 seemed like a “quick and dirty” attempt to copy Deepseek to me).

He hired literal narcissists in their place, who gave him big promises they couldn’t possibly keep, given their background.

Meta was literally at the center of open weights/open source ML land, which is where all this will eventually settle. Between that and PyTorch, they could have been at the center of the universe.

But they aren’t, because Zuck is so unbelievably insecure.

I’m not one to blame individuals for anything; systemic failure is complex. But he actually did this to himself.

And what’s bizarre is, somehow, it will work out for him.

I’ve been criticizing Facebook since I was in high school and it was all the rage. But here we are. Zuckerberg is more “revered” than ever.

Or just some person hosing them off (which kinda looks like what they’re set up to do, given the position near the railing).

Interesting indeed.

They should have bolted the panels closer, with shrouds flush with the concrete around the edges! Or maybe add a coolant loop? Perhaps the concrete helps with that already, but is just too technical to include in this article.

I dunno! Its a good question, I have no idea how much they’d actually ice up.

Maybe they have to clear it off the dam anyway, so the cost is sort of amortized?

And/or maybe the steep angle helps?

I mention it because that concrete looks clear of ice/snow to me.

That paycheck eats into power costs, though.

ChatGPT the app is not even in the same ballpark as the Chinese ones.

OpenAI treats it almost like personalized social media. It saves everything, so it can pull all sorts of stuff into the context. A bunch of ancillary services are front-and-center. In other words, is sycophantic engagement-maxing. And what its actually doing under the hood is opaque.

The Chinese services I’ve played with are more “utilitarian.” They have some great agent harnesses and tools, but generally its presented as a utility more than a “personal world” like ChatGPT, and what it’s using/doing is crystal clear.

In other words, from a service perspective, I find this claim silly. They’re diverging, greatly.

I’d dispute it for the underlying LLM, too, but on different grounds.

I watched some, and I’m about as “pro AI” as you’ll find on Lemmy, but honestly:

• Tons of scenes are jarring. Like, right off the bat, the zoomed in scene of walking feet is broken. Heads twist around, objects float; it’s creepy. It continues all through the movie, like the scene of the airplane landing.

• All the voices are robotic and tinny.

• All the scenes are too short.

And this makes sense.

Most TTS isn’t integrated with LLMs; it’s just interpreting text, so it can’t convey any emotion.

Video models are only good at ultra short scenes and are honestly still pretty bad.

Its one strength seems to be generating consistent, attractive faces in short bursts. Which makes sense: the primary use case for these things is influencer bots, impersonating real humans, for attention farms or marketing campaigns.

…Look. AI is just not good for this format. I doubt it ever will be without some serious cinema scaffolding and more minor use of generation sprinkled in as an aid.

You’re not wrong about a lot, but I think Linux products, shipped by OEMs, are a pretty good gaming platform.

And if they target Proton anyway, the fragmentation doesn’t really matter.

A smartphone, a tablet, a MacBook Neo and such may be “maintenance free” and secure, but they are not great gaming platforms. Not everyone wants to play mobile-style games.

And this:

Without Windows you wouldn’t have Windows gaming.

This is what I’m saying. We could.

Linux could dominate Windows for that specific use. If it gets true critical mass, the Proton devs would effectively control the direction of API development.

Let’s say most devs abandon native Linux and basically everything moves to Proton. Game devs start testing on it, then targeting it. Windows as a gaming platform withers away.

A Windows API, on Linux, is now the stable gaming API. It sets the standard.

…I’m content with that future.

I mean, the irony would be delicious. What better way to dance on MS’s grave than rob their API?

Nintendo is beyond the event horizon at this point.

And that’s alright. Like, if you can afford a Switch and you get it so your family can play Pokopia and Animal Crossing, all their nonsense doesn’t really matter to your lifestyle.

…But if it does matter, “Nintendo Fans” are in waaay too deep to plausibly defend them anyway.

How much would 2-3 dual-later BDs cost?

The issue isn’t the technical possibility, but whether people are willing to pay the premium for installation media.

You know. Store agnostic game key cards would be so cool.

I’m fine with archiving downloads I’d want + keeping keycards.

From OP in the Reddit thread:

Got five minutes of No Man’s Sky in, then I installed the update the machine had available and it bricked itself. If you’re still in the queue, look on the bright side: they’re presumably going to iron this crap out. Edit: To be more accurate, it’s giving the error light code for GPU failure.

Anticlimactic and somewhat embarrassing update: as some people suggested, I left it unplugged for about half an hour last night and then tried plugging it back in… and it didn’t work. So I left it unplugged for a couple of hours and then tried it again before bed… and it didn’t work. Same error light despite multiple power-cycling attempts. So I left it unplugged overnight and plugged it back in today to try some of the BIOS stuff that other people suggested… and it booted up immediately without issue.

I feel stupid about even posting this now, especially since it blew up a bit, but I was tired and irritable after a long day of work, and an ominous GPU error code wasn’t exactly the seamless plug-and-play experience I had hoped for. But I guess if anyone encounters the same error, don’t panic like I did, just let it sit for a few hours and it will somehow sort itself out.

It sounds like a firmware issue that broke display out. Perhaps Valve pushed a fix while it was “on” overnight, or OP reset something in the BIOS.

What I’m getting at is the specter this raises: The Steam machine seemingly relies on Valve for firmware support here, I think. Not AMD directly, like with the normal Linux firmware shipping.

That’s an interesting can of worms.

It’s not bad or anything, I’d trust Valve more than HP, Sony, or past OEMs who were notorious for wonky “custom” hardware support. And I suppose their track record with the Deck speaks for itself. But still, that’s something I’d keep in mind before buying one of these things.

This is a moot question, because the only thing that matters is engagement.

Internet forums:

• Are (mostly) specialized.

• Don’t have S&P 500-level advertising.

• Don’t constantly ping your phone by default.

• Don’t incorporate “dark patterns” in the UI.

• Don’t have AI-based algorithmic feeds.

• Have rules for posting, which slows down posting.

People didn’t give up on them; their attention got sucked away to other things. And nothing could have changed that, no matter how much they improved.

Not that I have any high horse to stand on.

For a few years, I’ve been telling myself I will become active in a few still-extant forums again (like DPReview and SpaceBattles). I never did.

Yeah, it’s mostly newer indie or “AA” games I’ve been into.

On the pricing: https://blog.subimpact.net/2026/06/valve-allegedly-threatened-to-delist.html?m=1

But in a nutshell, the allegations is that Steam threatens to delist games if they’re priced lower than Steam on other storefronts, similar to what Walmart or Amazon are accused of doing.

That makes sense. The image host I used is pretty unreliable.

Which is the second issue. No image hosts even support JXL or AVIF, except catbox.moe, which is down now! It’s infuriating.

What platform is that, out of curiosity? Im pretty sure everything should load an AVIF.