Forgot your password?
typodupeerror

Comment can't take my phone in to class... (Score 1) 50

Their cell phones will rot a generation, but they've got an AI horror bot to really screw up the kids. Will it help them with their suicide note ? or will it help plan the mostly justified revenge killings of the school council ? How about instead of the billion dollar AI bot we pay the teacher a few bucks and hire 2 of the mom's as class aids ? I liked Woz, mainly because he kept his fat ass out of the media and didn't make stupid grandstanding gestures.

Comment It's has been a while but... (Score 1) 203

In the so-called fire-storm(2017) in the California Bay Area, I was out of power for 8 days. Several houses on my street were on an isolated grid and it blew 3 times in a week. Once while the PG&E guy was still on the pole. It was kind of like camping for a couple of days then it began to suck. Luckily we didn't have medical machines or such. We had a huge BBQ the 2nd day because the freezers were thawing in the extreme heat. I live in Yuma Az now and looking back 104 wasn't that bad :)

Comment Re:Ticket prices are insane. (Score 1) 31

THIS THIS THIS. Bands used to make money in several ways...That isn't so anymore. The record labels FSCK'd everyone when they signed blanket agreements with the online services. Amazon/Temu killed the merch marked with Chinese knockoffs. All there is left is the show and the stuff they sell there, and bands are getting squeezed out there by fee's and tax's from everywhere. Corporations will own everything eventually.

Comment Re:American Open Weight Models (Score 1) 109

Any quantization beyond Q4_0 goes mad (losing context, making category errors, malforming markup tags, mangling and inventing words) long before the context reaches a size I would consider "useful". I've tried running 31B at IQ3_XS and it looks alright from a distance, for a few prompts at a time. Unfortunately the errors compound and semantic drift goes into runaway. I have no interest in playing with quantizations any smaller than that. The only thing "massively wrong" is the suitability of a 2-bit quantization for my actual use. I've investigated it, and it can be an amusing toy, but it is far from usable for any meaningful work.

Even Q4_K_M was too error-prone for meaningful use, and I was using Q5_K_M as my baseline of "usable for a while". Fortunately, most of the heavily quantized models are based on QAT releases now, and a 4-bit QAT model is pretty much on par with a 5-bit non-QAT model.

Comment Re:American Open Weight Models (Score 1) 109

DeepMind has released Gemma 4, which doesn't include their largest models, but 31B is a creditable model. I wish I could say the same about the 26B-A4B model, which is just smart enough to be entertaining for a few days or weeks, then everything it does starts to sound the same. Unfortunately, 31B on my hardware (i5-8500, 48 GB DDR4, 12 GB RTX 3060) runs at 0.3 to 1.2 t/s. So while I'm not a big fan of Alphabet's business practices at large, they aren't regressing. Switching to the MIT license basically means they're abdicating all control over derivatives.

The only thing Grok ever really did for open weight AI was show up for the party a couple times. At first, this helped establish a baseline that could never be retracted, but it hasn't proven to be particularly important. Everything since has been far off the bleeding edge, but they collect their participation trophies. I think their subsequent actions have gone a long way toward demonstrating their purposes, which are wholly selfish. They'll do as little as they can to contribute while retaining the benefits of being perceived as open and competitive.

Comment Indirectly imaged ? (Score 2) 70

So if someone posts your image to an Instagram account without your permission Meta suddenly owns your likeness and the ability to derive images from it ?
What if you are in media and your image is your income ? Does Meta get to remake the images of "Models" for ads on their platform ?
Will you find your own image in some ad ?

Slashdot Top Deals

One picture is worth 128K words.

Working...