Forgot your password?
typodupeerror

Comment Re:Meanwhile, back at the benchmarks (Score 1) 21

It actually makes RAM needs worse than a dense model, for a given quality.

That said, inference techniques like FreeToken are improving MoE swapping performance, so it's not as bad as it once was. Honestly, I would not be at all surprised if we start doing training in a swap-aware manner, where expert cache misses count as loss and models can queue experts to start loading before they're needed. Could even train in two tiers of loading - RAM and disk. Wouldn't that be great? Could run multi-terabyte models on consumer hardware. Training would tend to tend to concentrate key logic and reasoning on small number of very active experts, mixed reasoning/knowledge on less common experts that are usually kept in RAM, and rarer knowledge on experts that usually remain on-disk until needed.

Comment Re: Meanwhile, back at the benchmarks (Score 1) 21

There are well known, reputable abliterators on HuggingFace.

Also, censorship usually (not always) is something you care about for chats, not agentic work.

A .safetensors or .gguf model cannot run code unless you let it. No agentic framework = all it can do is write text. It can't hurt you directly. At worst, it can only lie to you.

Comment It's always flattening management (Score -1) 11

They've been doing this bullshit since the '70s and it always works. People hear it and think yeah fuck the manager is right?

I figured out the scam a while ago. The original purpose of management was to keep an eye on unions and prevent employees from unionizing. Seriously Google it.

America doesn't really have unions anymore so we don't really have management anymore. What you have are basically a handful of people to implement policy and then you have a shitload of line workers declared management for a variety of unpleasant reasons ranging from managers being exempt from overtime pay to using the manager title to get a lot more work out of someone for very little additional pay if any...

And every time it's layoff time they can jettison these "managers" and tell everyone they're just flattening the organization and making things more efficient so nobody starts poking around asking if the company is viable and if they should be selling off the stock.

It's also a very effective way to fire old people without getting sued for age discrimination. IBM fucking loves doing that and if you actually look at the history of their layoffs it's almost entirely age discrimination pretending to be efficiency

I am so tired of seeing the same fucking tricks work over and over and over again.

Comment Re:There are still VMWare customers? (Score 1) 45

For some customers, they looked at the resources needed to migrate, and decided it was cheaper to stay put and/or invest the resources that would be needed elsewhere in their business.

Yes, and the reason is that they stupidly painted themselves into a corner and never thought about an exit strategy. Had they been thinking about the possibility that they might one day need to migrate from the beginning, it would be much cheaper to do so now.

That's the thing about painting yourself into a corner, a little better planning at the start can always avoid it with little to no additional effort.

Comment Wake me when they actually shift (Score 1, Interesting) 45

VMware didn't do this in a vacuum. If they raise the prices that much it's because they have solid numbers that clearly indicate that their customers don't have anywhere to go.

I'm not a admin so I don't actually run my own VMs but from talking to other admins there is nobody who has tools for doing the kind of deployments you need to do in a large Enterprise except for VMware. I mean the tools exist but you're going to spend so much money paying people to monkey around with them you might as well just paid VMware.

There is a video from a YouTuber called Linus tech tips. His actual tech tips tend to be pretty shit but he does have one good video where he spent a few weeks trying to get all the people in his business off Adobe software and quickly found that yes he could do it but the man hours and the pay he had to spend doing it exceeded the cost of the licenses.

That's what Big Data gets you. You know exactly how much you can screw everybody over. You leave absolutely nothing on the table. Basically a giant fuck you to Leviticus 19:10.

Comment Re:Paywalled AI Marketing (Score 4, Interesting) 63

Nobody, I repeat, nobody who uses AI is going to see "OpenAI adding watermarks" and think, "Oh, I better use AI more, especially OpenAI".

And nobody, I repeat, nobody who doesn't use AI is going to see "OpenAI adding watermarks" and think, "Oh, I better use AI more, now that it'll be easier to see that what I did is made by AI."

It is not "an advertisement of the power of the technology". It's a regulation forced on them by the EU's AI Act, which neither companies nor users want.

Slashdot Top Deals

Some people have a great ambition: to build something that will last, at least until they've finished building it.

Working...