It’s pretty remarkable to have an organization that China and the EU reject, that other big powers respond with “uh, I’ll think about it.” Most who accept are lead by demagogue attention seekers, like Argentina, Belarus, Hungary, Turkey, Israel, Kazakhstan and so on.
…To be fair, it is interesting that Qatar, Saudia Arabia, Pakistan, Jordan, the UAE, and such are onboard. I would assume they have a genuine interest in Gaza humanitarianism.
The problem with Davinci I encountered was that very common encodings like x264 or x265 don’t work with the community edition and I can’t justify getting the pro version from an economic standpoint
vapoursynth
It’s basically video editing in Python, that can be piped directly to ffmpeg or whatever encoder you want.
…It’s finicky. And poorly documented.
It’s not fast and does a lot on CPU, but it’s extremely powerful. As an example, I have a script for transcoding high ISO footage that, frankly, blows Davinci’s filters out of the water. And I have another for fixing up an old DVD that I just couldn’t have done with Davinci.
As a real life example, the Canon 600mm F11 telephoto lens should be awful, on, say, a 32MP crop sensor R7. That’s insane pixel density somewhere in the ballpark of this Fuji.
…But look at real life shots of that exact combo, and they’re sharp as hell. Sharper than a Sigma at F6.3.
The diffraction limit is something to watch out for, but in reality, stuff like the lens imperfections, motion blur, atmospheric distortion and such are going to get you first. You don’t need to shoot at F4 on this thing to make use of the resolution, even if that is the ideal scenario.
See: https://www.omnicalculator.com/other/hyperfocal-distance
https://en.wikipedia.org/wiki/Hyperfocal_distance
But TL;DR: for distant landscapes on a wide field of view lens, you can shoot at F5.6 and everything is in focus. I even do this at F1.4 on my lowly aps-c camera.
Put more concretely, the hyperfocal distance for a 35mm f5.6 lens on a medium format camera is 5 meters. Everything in the distance can be in focus.
For portraits, you want background blur anyway.
And if you’re doing anything else on a medium format camera, you’re kind of insane, heh.
Davinci resolve has acceleration support on Linux, but it’s not for the faint of heart.
Doubly so with vapoursynth.
Honesty, Windows sucks too. As sad as it is, I color grade some video on my iPhone because it just works with HDR, in basically any format.
A good hijack, thanks.
I’m in your boat.
I don’t play anything anymore, basically. Even “nostalgia” doesn’t feel fun because the dopamine hit from figuring the mechanics/lore out has passed.
My advice: co-op.
Playing a cooperative game with mates is fun. As an example, I got into Baldurs Gate 3 with family; never would have gotten so engrossed without them, and it sped up combat.
We started an Age of Wonders 4 game. One of them bounced off because it was too slow waiting for each other’s turns, but it was enough to get me engrossed with its systems and lore.
As for real life?
Gamification is a freaking menace.
Organizer apps are great. Apps/reminders that beat you in the head are excellent. But I’ve had just about enough manipulation from my phone, thank you.
Sony’s modern OLEDs are sick. There are a few between my family, and they have the best processing I’ve seen, they decode massive bluray rips no problem, and native options for a clean ad-free UI.
Why TF aren’t people buying them?
It would be embedded. Not on sticks.
It costs more to carefully cut the memory ICs off a motherboard and repackage them than to just buy RAM.
In rare cases it can be done (like how Chinese shops resolder GDDR to make 48GB RTX 4090s), but I don’t see EV RAM being worth it.
But to answer your question… some self driving systems or even infotainment systems do have a somewhat substantial amount of RAM.
Pro tip: go for the motor. Or the battery, if you can swing it.
It’s because social media is toxic.
Look at the view counts below conservative posts. That is why there is no counter; the game is rigged.
Good!
Even if you’re a climate change denier, fact is Ford has to figure out cheap EVs anyway. ICE costs pretty much bottomed out, unless you’re talking about a generator for an EV drivetrain.
Yyep. The Nazi gambit.
Just a little nationalist fascism to get the economy going.
Voters seem to always think “well, unlike the last hundred times, it might be alright this time?”
And, in practice, take forever, don’t really promote it, and make it look like no one at Facebook has ever laid eyes on a video game.
Oh and add to that stupidity the fact that almost every company refused to keep working from home after the pandemic
Including Facebook.
Zuck loves a bandwagon.
“How does this keep happening!?”
It’s one of those “reporting on social media without actually adding anything” articles, but in this case it’s pretty cool.
You mean an Nvidia 3060? You can run GLM 4.6, a 350B model, on 12GB VRAM if you have 128GB of CPU RAM. It’s not ideal though.
More practically, you can run GLM Air or Flash quite comfortably. And that’ll be considerably better than “cheap” or old models like Nano, on top of being private, uncensored, and hackable/customizable.
The big distinguishing feature is “it’s not for the faint of heart,” heh. It takes time and tinkering to setup, as all the “easy” preconfigurations are suboptimal.
That aside, even you have a toaster, you can invest a in API credits and run open weights models with relative privacy on a self hosted front end. Pick the jurisdiction of your choosing.
For example: https://openrouter.ai/z-ai/glm-4.6v
It’s like a dollar or two per million words. You can even give a middle finger to Nvidia by using Cerebras or Groq, which don’t use GPUs at all.
Yeah, accessibility is the big problem.
What I used depends.
For “chat” and creativity, I use my own version of GLM 4.6 350B quantized to just barely fit in 128GB RAM/24GB VRAM, with a fork of llama.cop called ik_llama.cpp:
https://huggingface.co/Downtown-Case/GLM-4.6-128GB-RAM-IK-GGUF
It’s complicated, but in a nutshell, the degradation vs the full model is reasonable even though it’s like 3 bits instead of 16, and it runs at 6-7 tokens/sec even with so much in CPU.
For the UI, it varies, but I tend to use mikupad so I can manipulate the chat syntax. LMStudio works pretty well though.
Now, for STEM stuff or papers? I tend to use Nemotron 49B quantized with exllamav3, or sometimes Seed-OSS 36B, as both are good at that and at long context stuff.
For coding, automation? It… depends. Sometimes I used Qwen VL 32B or 30B, in various runtimes, but it seems that GLM 4.7 Flash and GLM 4.6V will be better once I set them up.
Minimax is pretty good at making quick scripts, while being faster than GLM on my desktop.
For a front end, I’ve been switching around.
I also use custom sampling. I basically always use n-gram sampling in ik_llama.cpp where I can, with DRY at modest temperatures (0.6?). Or low or even zero temperature for more “objective” things. This is massively important, as default sampling is where so many LLM errors come from.
And TBH, I also use GLM 4.7 over API a lot, in situations where privacy does not matter. It’s so cheap it’s basically free.
So… Yeah. That’s the problem. If you just load up LMStudio with its default Llama 8B Q4KM, it’s really dumb and awful and slow. You almost have to be an enthusiast following the space to get usable results.
The problem is being “anti AI” smothers open weights ML, doing basically nothing against corporate AI.
The big players do not care. They’re going to shove it down your throats either way. And this whole fuss is convenient, as it crushes support for open weights AI and draws attention away from it.
What I’m saying is people need to be advocating for open weights stuff instead of “just don’t use AI” in the same way one would advocate for Lemmy/Piefed instead of “just don’t use Reddit”
The Fediverse could murder trillions of dollars in corporate profit with enough critical mass. AI is the same, but it’s much closer to doing it than people realize.
Yeah. Lemmy is kind of crazy. For all we complain about Reddit being filled with clickbait and echo chambers, we kind of replicated it here.
Like, I feel I’m pretty extreme/progressive, yet I don’t feel like I’m in the same reality as /c/politics.