technology

cross-posted from: https://lemmygrad.ml/post/8137348

https://archive.ph/Rgbkz

Noticed this while I was going through bank statement .pdfs on my computer. What would Co-Pilot even do with those????

At least Co-Pilot turned out to be easy to uninstall. This was a real sneaky move though, since it's not like it showed up after a Windows update, instead it just silently installed itself on my fucking computer. Also, just the sheer fucking gall to make it the first option when you right click on anything

I foolishly assumed Microsoft wouldn't bother pushing AI as hard onto Win10 since they're dumping it soon. At first I panicked and thought this was Recall, but I wouldn't be surprised if that's coming next

Enough talk about watermarks and identifying which AI slop is real or fake. Regulate the power consumption, the water usage, the emissions, and/or the bandwidth. Either kill that shit in the cradle or let companies finally figure out how to proooooompt without killing our habitats even faster than we already were.

https://archive.ph/LJJpx

sadly, he still has too much faith in his fellow tech bros. his tactical analysis is right -- it would work, if pursued. but he's still missing the "material conditions" portion of the analysis. unless something drastic changes, this ain't happening any time soon.

AI companies claim their tools couldn't exist without training on copyrighted material. It turns out, they could and it just takes more work. To prove it, AI researchers trained a model on a dataset that uses only public domain and openly licensed material.

What makes it difficult is curating the data, but once the data has been curated once, in principle everyone can use it without having to go through the painful part. So the whole "we have to violate copyright and steal intellectual property" is (as everybody already knew) total BS.

cross-posted from: https://lemmy.ml/post/31250684

cross-posted from: https://lemmy.ml/post/31250679

"this morning, as I was finishing up work on a video about a new mini Pi cluster, I got a cheerful email from YouTube saying my video on LibreELEC on the Pi 5 was removed because it promoted:

Dangerous or Harmful Content Content that describes how to get unauthorized or free access to audio or audiovisual content, software, subscription services, or games that usually require payment isn't allowed on YouTube.

I never described any of that stuff, only how to self-host your own media library.

This wasn't my first rodeo—in October last year, I got a strike for showing people how to install Jellyfin!

In that case, I was happy to see my appeal granted within an hour of the strike being placed on the channel. (Nevermind the fact the video had been live for over two years at that point, with nary a problem!)

So I thought, this case will be similar:

• The video's been up for over a year, without issue
• The video's had over half a million views
• The video doesn't promote or highlight any tools used to circumvent copyright, get around paid subscriptions, or reproduce any content illegally

Slam-dunk, right? Well, not according to whomever reviewed my appeal. Apparently self-hosted open source media library management is harmful.

Who knew open source software could be so subversive?"

The project implements sparse multiplication and fuses up/down projections in the MLP layers through low rank weight activations. Work is based on Deja Vu and Apple's LLM in a Flash.

This approach avoids loading and computing activations with feed forward layer weights whose outputs will eventually be zeroed out.

It's a lossless approach as these weights anyway do not contribute in the current token prediction. It does however, need the predictors to be accurate in clustering the weights.

The result? 5X faster MLP layer performance in transformers with 50% lesser memory consumption avoiding the sleeping nodes in every token prediction. For Llama 3.2, Feed forward layers accounted for 30% of total weights and forward pass computation resulting in 1.6-1.8x increase in throughput:

Sparse LLaMA 3.2 3B vs LLaMA 3.2 3B (on HuggingFace Implementation):

- Time to First Token (TTFT):  1.51× faster (1.209s → 0.803s)
- Output Generation Speed:     1.79× faster (0.7 → 1.2 tokens/sec)  
- Total Throughput:            1.78× faster (0.7 → 1.3 tokens/sec)
- Memory Usage:                26.4% reduction (6.125GB → 4.15GB)

https://archive.ph/RKZSD

Best channel on youtube dropped another banger

cross-posted from: https://rss.ponder.cat/post/198503

Apple provided governments around the world with data related to thousands of push notifications sent to its devices, which can identify a target’s specific device or in some cases include unencrypted content like the actual text displayed in the notification, according to data published by Apple. In one case, that Apple did not ultimately provide data for, Israel demanded data related to nearly 700 push notifications as part of a single request.

The data for the first time puts a concrete figure on how many requests governments around the world are making, and sometimes receiving, for push notification data from Apple.

The practice first came to light in 2023 when Senator Ron Wyden sent a letter to the U.S. Department of Justice revealing the practice, which also applied to Google. As the letter said, “the data these two companies receive includes metadata, detailing which app received a notification and when, as well as the phone and associated Apple or Google account to which that notification was intended to be delivered. In certain instances, they also might also receive unencrypted content, which could range from backend directives for the app to the actual text displayed to a user in an app notification.”

From 404 Media via this RSS feed

https://archive.ph/VySZU

https://github.com/IRS-Public/direct-file

The researchers estimate that the lenses cost around US$200 per pair to make.

Researchers have made the first contact lenses to convey infrared vision — and the devices work even when people have their eyes closed.

Soon, it will not even be an exaggeration to say that China is beating the US with their eyes closed