brucethemoose

joined 6 months ago
[–] brucethemoose@lemmy.world 3 points 1 hour ago* (last edited 1 hour ago)

...But I believe Jaws can save the world

[–] brucethemoose@lemmy.world 51 points 17 hours ago (9 children)

Israel’s “Special Military Operation”

Even Russia learned their lesson in this part of the world.

[–] brucethemoose@lemmy.world 32 points 18 hours ago (2 children)

I mean, as long as the bots click on ads, everyone is happy? Riiight?

[–] brucethemoose@lemmy.world 2 points 18 hours ago

They'll probably mix it up, and randomize the results of different categories... I guess?

But the last thing AI search needs is spam.

[–] brucethemoose@lemmy.world 10 points 18 hours ago (2 children)

Wait until they put ads into the AI search results.

[–] brucethemoose@lemmy.world 1 points 1 day ago* (last edited 1 day ago) (1 children)

Yes, but this is an offline game, and I've never seen such a warning without some plausible justification. There's no basis for interfering with an online component here, so what would Larian even say as they sent warnings?

Using an legally purchased offline game "illegally" would be quite a precedent, no?

My guess is that it won’t get shut down because WOTC can’t make Larian bully people into shutting it down.

Yeah. Larian didn't seem very interested in blocking this capability (they left all this stuff in the executable), like they did the absolute minimum they were contractually obligated to do lol.

[–] brucethemoose@lemmy.world 3 points 1 day ago* (last edited 1 day ago) (3 children)

Thanks, this is exactly what I was trying to ask. What "motivation" potential modders have.

wotc would have to "admit" to doing what the community only suspects (deliberately restricting these tools in the contracts) to harass modders, right? That could be a PR disaster, hence I hope that means they'll turn a blind eye.

[–] brucethemoose@lemmy.world 3 points 1 day ago

This does benefit him if it gets him votes. He wants voters to like him, and he'd absolutely build this crazy pipe and slap his name on it if he could.

But like you said, he'd drop it like a rock if it's inconvenient.

Unlike other politicians, Trump accepted there's no real consequence for making fantasies up and almost lying, just like he did in business.

“Is he saying this because he thinks it benefits him to say it, or because he thinks it benefits him to do it?”

And anyone who's on the fence about Trump is not thinking critically like this, they are looking at a few things he's saying and pondering if its a good thing and benefits them.

And again, fact-based news journalism does not have the luxury of assuming "Here's what we think he's saying, and we think he's making that up because it benefits him, so it's probably nonsense."

[–] brucethemoose@lemmy.world -3 points 1 day ago* (last edited 1 day ago) (1 children)

It's not totally incoherent though, its vague and almost poetic.

This is kind of Trump's talent. He makes these grand statements that aren't quite lies. The crowd gets exactly what he's trying to say: all this water pouring out of snowy mountains into the ocean is a "waste" when it could just be diverted to LA, so let's fix that. It's worded almost like a dream. It's an attractive fantasy. But it's also vague, not quite enough to be a lie even if the implied facts are straight up wrong.

What can the news do? If they dig into it, he didn't really make any hard claims to roast. They can veer into opinion talk and say that sounds unpresedential and that his speech should be more clear, but making fun of his speech style at a rally is not supposed to be their job. So they do what they can, guess what he's saying and refute that.

Again, this was his talent before he got into politics. The Motley Fool did this great podcast on Trump (before Trump was big and political) where he sold massively overvalued real-estate from his private company to his public one, effectively "duping" the market, and it worked because he sold it as a vague fantasy just like this. He got plenty of criticism and it didn't matter, because he threaded the needle and what he's claiming is not hard enough to stick. This is what he does.

[–] brucethemoose@lemmy.world 15 points 1 day ago* (last edited 1 day ago) (6 children)

The difference is he could be the next president and try to turn whatever he's thinking into national policy, so it's worthwhile to try and dissect what he's saying.

But those experts are also (somehow, still) not really accustomed to Trump's bombastic language. He was like this long before he got into national politics, hyping real estate and business for the market (where it kind of worked). That's a totally different world, where half lies and crazy sales talk are the norm.

[–] brucethemoose@lemmy.world 5 points 1 day ago* (last edited 1 day ago)

What are people doing with these super expensive boards now? Like, I know there's always the "top 1% first-person-shooter" niche that wants that last sub-millisecond of latency, playing games that don't really respond to 3D cache, but... what else? That's not a big niche. Modern CPUs have like no overclocking headroom, and even at stock are pushed way too hard.

I'd only spend that kind of money on an embedded Strix Halo board, or HEDT with tons of PCIe lanes. I just don't see why you'd shell out for Arrow Lake like that when you can get 95% of the performance for a fraction of the price and power usage elsewhere.

[–] brucethemoose@lemmy.world 2 points 1 day ago

The motherboard doesn't matter AMD's 3D cache CPUs, which are king for these kinds of games. From what I've seen, you'd be crazy not to get either a 5700X3D or a 7800X3D with a cheap mobo.

13
submitted 1 week ago* (last edited 1 week ago) by brucethemoose@lemmy.world to c/localllama@sh.itjust.works
 

https://huggingface.co/collections/Qwen/qwen25-66e81a666513e518adb90d9e

Qwen 2.5 0.5B, 1.5B, 3B, 7B, 14B, 32B, and 72B just came out, with some variants in some sizes just for math or coding, and base models too.

All Apache licensed, all 128K context, and the 128K seems legit (unlike Mistral).

And it's pretty sick, with a tokenizer that's more efficient than Mistral's or Cohere's and benchmark scores even better than llama 3.1 or mistral in similar sizes, especially with newer metrics like MMLU-Pro and GPQA.

I am running 34B locally, and it seems super smart!

As long as the benchmarks aren't straight up lies/trained, this is massive, and just made a whole bunch of models obsolete.

Get usable quants here:

GGUF: https://huggingface.co/bartowski?search_models=qwen2.5

EXL2: https://huggingface.co/models?sort=modified&search=exl2+qwen2.5

 

Obviously there's not a lot of love for OpenAI and other corporate API generative AI here, but how does the community feel about self hosted models? Especially stuff like the Linux Foundation's Open Model Initiative?

I feel like a lot of people just don't know there are Apache/CC-BY-NC licensed "AI" they can run on sane desktops, right now, that are incredible. I'm thinking of the most recent Command-R, specifically. I can run it on one GPU, and it blows expensive API models away, and it's mine to use.

And there are efforts to kill the power cost of inference and training with stuff like matrix-multiplication free models, open source and legally licensed datasets, cheap training... and OpenAI and such want to shut down all of this because it breaks their monopoly, where they can just outspend everyone scaling , stealiing data and destroying the planet. And it's actually a threat to them.

Again, I feel like corporate social media vs fediverse is a good anology, where one is kinda destroying the planet and the other, while still niche, problematic and a WIP, kills a lot of the downsides.

 

Senior U.S., Qatari, Egyptian and Israeli officials will meet on Thursday under intense pressure to reach a breakthrough on the Gaza hostage and ceasefire deal.

he heads of the Israeli security and intelligence services told Netanyahu at the meeting on Wednesday that time is running out to reach a deal and emphasized that delay and insistence on certain positions in the negotiations could cost the lives of hostages, a senior Israeli official said.

 

HP is apparently testing these upcoming APUs in a single, 8-core configuration.

The Geekbench 5 ST score is around 2100, which is crazy... but not what I really care about. Strix Halo will have a 256 -bit memory bus and 40 CUs, which will make it a monster for local LLM inference.

I am praying AMD sells these things in embedded motherboards with a 128GB+ memory config. Especially in an 8-core config, as I'd rather not burn money and TDP on a 16 core version.

view more: next ›