Thank you for that explanation. My regex impaired ass thought he wanted to hurt generation[x|y|z].

I’m like “what’d we ever do to you?”

At least the article points out that this is a Wall Street valuation, meaning it’s meaningless in reality, the company doesn’t have that much money, nor is it actually worth that much. In reality, Nvidia’s tangible book value (plant, equipment, brands, logos, patents, etc.) is $37,436,000,000.

$37,436,000,000 / 29,600 employees = $1,264,729.73 per employee

Which isn’t bad considering the median salary at Nvidia is $266,939 (up 17% from last year).

It sounds like the processor is the real limitation. Plenty of stuff from Windows XP era and before ran in less than 512MB.

… it should be possible to continue supporting those older cards by setting the compile time information to match your gpus.

I’ll admit that’s probably over my head, but still good advice. The 1080ti works fine for now, but whenever it gets dropped from support, I have a good home for it.

I have a Win7 machine so I can run my older games that don’t play well in linux. Nvidia Rocket Sled Demo on a 1080ti screams.

Awesome. I’d heard that Pat was one of Redd’s old friends from the “Chitlin’ Circuit” era of comedy, but I’ve never actually seen him do standup.

Probably better to ask on [email protected]. Ollama should be able to give you a decent LLM, and RAG (Retrieval Augmented Generation) will let it reference your dataset.

The only issue is that you asked for a smart model, which usually means a larger one, plus the RAG portion consumes even more memory, which may be more than a typical laptop can handle. Smaller models have a higher tendency to hallucinate - produce incorrect answers.

Short answer - yes, you can do it. It’s just a matter of how much RAM you have available and how long you’re willing to wait for an answer.

One thing I love about usenet is that it’s great if you’re just looking for one episode, song, etc and don’t want to download a whole collection.

Why replace torrents? Why not use both? It’s a bonus if your usenet provider includes a VPN.

Yep, I had been hoping for the same thing.

Also, to @[email protected], you might want to wait and see what gets announced at Computex next month. Hopefully they announce some new stuff and the current gen prices drop.

Ollama doesn’t currently support mixing CUDA & ROCm. https://github.com/ollama/ollama/issues/3723#issuecomment-2071134571

One thing to keep in mind about adding RAM your speed could drop depending on how many slots you populate. For me, I have a 5700G and with 2x16Gb, it runs at 3200Mhz, but with 4x16Gb(same exact product), it only runs at 1800Mhz. In my case, RAM speed has a huge effect on tokens/sec, if I have a model that has to use some RAM.

You can check AMD’s spec page for your processor, but they don’t really document a lot of this stuff.

This is why I want Intel Arc to succeed, because maybe it’ll make AMD take Radeon more seriously. It’s been a steady decline since the ATI days (2006)

Source

I love my Ryzen though.

I think it’s because architecture-wise, these are more in line with the 8000 series. The 7500F is a full featured Ryzen 7000, but with broken graphics. These are basically 8000 series APUs, so they have the lower PCIe version, lower cache and fewer PCIe lanes, and broken graphics.

I suspect the price will come down soon because no one should be buying these at this price. The 8400F is only $10 less than a full 8500G.
Edit: I just saw that the 8400F is a six-core, so maybe that helps justify the price.

I think 5 is what I’m looking for. Sometimes Civ (in general) gets out of hand with the micromanagement. I want something that’s kind of casual.

I’ve been looking to try Civ 5 or 6, but haven’t decided which one yet.

Did they give up on trying to put X3D into laptops?

That would be nice. And maybe a bridge for sharing memory between cards, since Nvidia got rid of theirs.

I’d love a 24Gb card, less than 270mm, without the cursed power connector.

Is this the generation where they’re only making mid-range and lower cards?

So it seems obvious that the US dept of Homeland Security would include a Meta rep on their AI Safety and Security Board. But they didn’t.

Don’t get me wrong, I hate Facebook/Instagram/Meta, but I appreciate their contribution to AI.

Even after the price cut, theirs is still 3x the price of Mercedes’ system which works better. I have a feeling Tesla’s earnings report won’t go well this afternoon. https://finance.yahoo.com/news/tesla-earnings-q1-175358835.html

I wonder how many offers he’s gotten for it so far.