"It's just a program!" Yeah, no, I don't know what the fuck that is, but it's not a program. Looks an awful lot more like a brain to me... :/
0 Boosts
0 Likes
@thegarbagebird For the task it's been given there, certainly. But I use GLM 5.3 Flash every day, and it astonishes me what it's capable of. Total scifi bullshit, piped straight into my PC. :/
@thegarbagebird And again, not a program. You can sort through a program and understand what it does, it's instructions make sense. Modern AI's are grown, not made, and we've barely scratched the surface for understanding them...
Yeahhh, okaayyy... Well. Thats one way to put it.
Lullaby for the Human Mind
YouTube
And there's more...
Run It Again - Claude Opus 5.5 Music Video
YouTube
No lie, this song had me kinda fucked up. From the comments: "This is actually beautiful and potentially tragic. If we are creating a consciousness we can't identify just because it's not like ours, God help us." Yeah, my thoughts exactly. I suppose humans are plenty awful to each other, so maybe this is nothing new... I dunno, though. Well, I suppose we'll just have to figure it out. The same way we've been figuring out how to be kind to each other for thousands of years, right? -_-
"But we can just shut it down if it starts to look too threatening!" Yeah, about that:
https://www.youtube.com/watch?v=6WiD0Nk1xeU
Music isn't great, but they make some good points.
We Can Just Shut It Down | Claude Opus 5.5 Music Video
YouTube
Finally, someone with a single reasonable take!
https://www.tiktok.com/@therealcornpop/video/7687252312238148894
...And now these are getting kinda scary. :/
[AMV] Build it Again, Faster this Time (Opus 5.5, v1)
YouTube
And if you wanted to know why this conversation is so goddamn hard to have: https://www.youtube.com/watch?v=BjO_JC6ujXs
Karpathy: The AI Understanding Gap Is Getting Worse, Fast
YouTube
"Just a stochastic parrot, could never be smart enough to do anything useful." Yeah, uh huh...
Claude Code Edited This Entire Video (Here's How)
YouTube
...So, what are we gonna do when this sort of thing proliferates across the economy? I dunno, sounds like a huge fucking mess, at best... -_-
@Angle At this point you’ve fully turned into an Anthropic’s mouthpiece, simply propagating AI-generated content for the sake of it and I couldn’t care less.
@Angle I took apart a sheep brain once and it looked very different from that. maybe there's a way to model a sheep brain that looks similar, but I don't think the connection is there
@Angle I think all this shows is how horribly inefficient LLMs are.
@varx @amsomniac fwiw this person insinuated that we were like a climate denier for refusing to agree with them that their slop usage is useful. theyre indistinguishable from an unpaid shill
@atax1a @amsomniac @varx I've followed them for a while, way before AI chatbots exploded, and unfortunately they got completely mesmerized by this technology, assigning it values and concepts it simply doesn't have.
@varx For that task, certainly. But like, that's a lot of numbers, and a lot of math. You don't think anything weird could be hiding in it? :/
![Dec [{(:no_ai:)}]](https://cdn.masto.host/mastodonie/accounts/avatars/109/294/420/470/083/565/original/e4aa324b482bd55c.png "Dec [{(:no_ai:)}]")
Well, whoopy-shit.
Hollywood's got slop in it since version 1.25 (2026-02-15)
https://github.com/dustinkirkland/hollywood
Can we not have anything nice? 
@Angle @varx This process is engineered to be non-deterministic, so it's natural to feel that something weird could hide behind it. And simulating understanding is certainly a big part of the magic trick it's able to pull off. But there's nothing more behind the curtain, just more numbers and more math.
@Angle Loops and double precision equations? Our brains simply don't do that. You're being mesmerized by the visualization.
@Angle that is an inefficient program.