Some users have recently had their accounts hijacked. It seems that the now defunct EVGA forums might have compromised your password there and seems many are using the same PW here. We would suggest you UPDATE YOUR PASSWORD and TURN ON 2FA for your account here to further secure it. None of the compromised accounts had 2FA turned on.
Once you have enabled 2FA, your account will be updated soon to show a badge, letting other members know that you use 2FA to protect your account. This should be beneficial for everyone that uses FSFT.
The two issues that MiniLED is likely to lose out on (having both MiniLED and OLED IRL here) that you won't really see on most videos is viewing angles and motion clarity. Viewing angles matter less for a TV, but for a monitor where you are up close, the edges are at quite an angle to your eyes...
I mean at this point I'm basically counting on it not coming out. First because this shit was announced in like, 2012. It has been over a decade of work on it, and it isn't out. I get it is a passion project and as such isn't going to get the kind of focus a commercial game is, but still this is...
Can afford, as in "I am willing to pay 6-7 figures for a big multi GPU server to run larger models." Not something home users can do, perfectly within reach of businesses.
They can't surveil you though. LLMs don't have any way to access the Internet on their own, they have to have a tool to use, and you can very easily monitor (and restrict if you want) tool calls on a system you control and it is the platform you run them on that provides the tools, not the...
I haven't seen anything from them any time recently.
Gemma is a bit of a gotcha. It is kinda like Qwen: Open... but only small models. Gemma has small models for single GPU systems. None of the Gemini models, the actual big frontier ones, are open. Qwen, Alibaba's, is similar: For the latest...
SSD speed is a huge bottleneck. These things are fucking massive and access needs to be extremely fast. Even if you had a perfect swap algorithm, it will still slow shit down a ton. I mean let's say you have the fastest nVME PCIe 5 drive you can get and it really does manage to push data to VRAM...
Also with open weight models running on your hardware: You can re-train them. If there's an issue you can make sort of a "training patch" that changes behavior. You see this with image generators all the time with LORA files where someone makes an addon that changes how it works in a certain...
Yep and this is something that you see with other models as well. Any mixture of experts type model and can have only the active part loaded in to RAM... but you will pay a performance penalty as you swap it in and out. It is the same thing you see with system RAM vs VRAM: You can totally let...
Pretty much :D. If we want high rez and high refresh, we have to accept technologies like this, or we have to accept games with lower detail. There's no magic technology that is going to let us just brute force the frames. I mean a 5090 can't do it and that chonk is pushing 600w... How much...
I tried to do bikes but since my driving style can best be described as "Psychotic drunken jackass wearing a blindfold" I ended up getting tossed off them too much. So cars it is for me.
I don't know that it is all quite as grim as people want to pretend. I see a lot of "Oh young people do this because they have no hope of a good future otherwise, they can only get one by gambling!" However that doesn't seem to be borne out by the amount of money that many make and then blow...
Even if they aren't, it wouldn't be worth the cost. Core design is expensive. It isn't like you sit an intern down and say "Hey that P core? Make one like that but bigger, have it done by Friday." It takes a lot of work. With the P cores and E cores, it makes sense, because the work is already...
Ya it was more Intel I was thinking about since they already have lorge, smol, and very smol cores on their chips. My thought was "What about extra lorge?" Like an X core that is the size of two P cores and isn't even used for normal multi-threaded, apps, just things like games that really peg...
That could work well for games. Even with some of the best coded multi-threaded games, there's always a main thread, that everything else has to wait on. Give that higher clocks, you are better performance when CPU limited.
It is actually something I've wondered about for hybrid core designs...