ThorrJo

joined 2 years ago
[–] [email protected] 13 points 1 week ago (3 children)

But what makes this AI model unique is that it’s lightweight enough to work efficiently on a CPU, with TechCrunch saying an Apple M2 chip can run it.

An Apple M2 can run bigger, higher-precision models than this FWIW. More important than this is perhaps whether older CPUs can run it with acceptable performance.

AI models are often criticized for taking too much energy to train and operate. But lightweight LLMs, such as BitNet b1.58 2B4T, could help us run AI models locally on less powerful hardware. This could reduce our dependence on massive data centers and even give people without access to the latest processors with built-in NPUs and the most powerful GPUs to use artificial intelligence.

This is definitely relevant to my interests especially with NPU support for such models coming. Dirt cheap ARM-based PCs based on e.g. the RK3588 are shipping with small NPUs

[–] [email protected] 3 points 1 week ago

Love me some Bouguereau style

[–] [email protected] 1 points 1 week ago

I'm planning to keep a shitphone just for such apps.

 

So, like it says on the tin: I want to try GrapheneOS on my next new-to-me phone, which practically limits me to the Pixel series. And my price point will likely limit me further.

Anybody got thoughts on the Pixel 6a vs Pixel 7, those exact models (e.g. not Pixel 7 Pro etc)?

I currently have an old phone with Snapdragon 845 and 6GB of RAM which still performs well enough under LineageOS with microG to keep me happy.

Fears: that I'll miss the microSD slot; that the 6a's battery life might suck even on GrapheneOS; that the battery on a Pixel might be hard to replace.

Nice-to-haves: a camera that works well in low light.

Thoughts?

[–] [email protected] 6 points 2 weeks ago

Although purported interest has never been higher, all the soft white libs I know are still deep in their full-of-shit phase and not doing a single fucking thing to prepare. It's only the far leftists and people in the center that are on the uptake thus far from my vantage point.

[–] [email protected] 3 points 2 weeks ago

The bsky vibe is rancid as hell. It's full of those kind of people. I'm in favor of whatever keeps as many of them as possible out of the ActivityPub milieu.

[–] [email protected] 3 points 2 weeks ago (1 children)

Crap wallet for bewildered noobs who want to pay eye-watering fees for convenience

[–] [email protected] 3 points 3 weeks ago (1 children)

the problem lies within the underlying protocol.

The problem lies with Gargron doing what Gargron does, implementing whatever the f he wants for "the Mastodon network" and not giving a crap how it affects the health of the overall fediverse.

Hell, this isn't even the first time there's been drama over Mastodon's advisory post scopes, not by a long shot. I kinda wish I'd saved receipts from the last couple times, some highly experienced devs have chimed in in the past.

[–] [email protected] 4 points 3 weeks ago

Mastodon implemented a new feature in a way that would break (in a really jarring privacy-violating-to-some-extent way) until everyone else copied their implementation exactly.

You ever notice how Gargron refers to the fediverse as "the Mastodon network?"

He's been doing things this way since 2017 at least. At this point, any longtime observer really has no other choice but to consider the behavior deliberate.

[–] [email protected] 2 points 3 weeks ago

Then when I see people asking questions of his code or how to do things he usually jumps down their throats - or completely ignores them.

And honestly the biggest thing that irked me was that I didn't feel he gave credit to the hundreds - thousands of other people who work to make the fediverse work.

Anyone who's ever touched the Mastodon dev process knows that Gargron is much the same, FWIW, minus getting angry in public. These days I just have to shake my head at all the bright-eyed bushy-tailed noobs updating issues on the Mastodon repo, because those of us who've been around since the start know exactly how far that's gonna go in nearly all cases - and in the cases it does go anywhere, it'll be because Gargron implemented something similar with zero discussion and no credit where credit is due.

But yeah, follow Dansup long enough and you are guaranteed to see some regrettable behavior on main.

[–] [email protected] 1 points 3 weeks ago (1 children)

And the main reason Bluesky can have that is because it's not actually decentralized.

[–] [email protected] 7 points 3 weeks ago

The furry engineer who writes cryptography posts is working on this. Turns out it is not so simple in practice.

https://soatok.blog/category/technology/open-source/fediverse-e2ee-project/

[–] [email protected] 3 points 3 weeks ago (1 children)

If any dev should be getting roasted, it's Gargron, for his many bad decisions over the years.

 

I've had an xmrig process running on Linux terminate after printing "Aborted." twice in the last week, never seen this before. Anybody know what might be the cause? System was recently upgraded to Debian 12 and had an NVMe stick put in.

view more: next ›