You mean the arson connector? Shout out to dawid

I was in a very similar situation not too long ago, and my solution was also to just have a spare ssd, install windows on it, make the update and then go back. I tried to use a pe version of windows, as well a full blown windows on a USB. These solutions just didn’t work as expected, somehow the update would always fail. Windows on a proper ssd worked

You do have a good point that at a certain point it is not possible to have assurances as in values. However, there are other assurances that can be even more important, such as, contractual ones. If they do not have the proper assurances at that level too up to a certain point, then everyone is highly blindsided on the c suite.

But my biggest point has to do with burning a massive bridge, the consumers, and take a dive into this nonsense.

You can always try second hand market if in your region is somewhat healthy. I am not entirely sure what is your use case, but for work related and studies I would advise to get a lenovo thinkpad of some kind. Durable machines overral and can be get for decent prices. Other enterprise level laptops also work

You are not wrong about the reallocation part. However, if you see the actions from micron (fuck you micron BTW), they are going all in and having a shit storm in PR on the consumer side. If they are taking these risks without proper assurances, then they are utterly deranged

For the dram unfortunately won’t be possible to use it in the consumer space, at least not in the current form. Hbm is really server stuff, and as is, you cannot repurpose it. As for the GPUs, maybe they can be used for the consumer space but I am not entirely sure the specs would be wise to use it at home, since they need some very serious cooling capabilities, as well electricity consumption. Biggest winners of this pop in my opinion would be anyone who need cheap server rack stuff.

As long am4 and 12th gen survives on these dark times we got chances of getting through this

I had issues with this as well, in my case it turns out he wanted to be warm, like very warm. I am not kidding, we are talking 3 hot water bags to setup the bed and leave 2 after he was in

There are tools to actually remind you to do this on a timely fashion… Also, some of them go as far as doing auto renewal. Is this such a hard thing?

There is certainly a very big amount of fuckery going on right now with nvidia drivers. I simply did not know it was getting this bad. Also, I find very interesting the nvidia “open source” bit got the criticism it deserves (is just not open. There was a transfer of responsibilities, and one small part got open)

I find great the devs of the respective projects took time to actually tell how things work for gn. Sure, they get the attention because it’s a known channel, but on the other hand, Microsoft would never give them this much detail and attention to actually understand what’s happening

And running some commands with sudo, you probably should! (Unless is not your stuff and you don’t care)

I do not use models in general online, but my needs are also much smaller. Max I use my local model for ollama is translations. I am always interested in seeing more focused models so we can use on lower end hardware

Did you try to do this workflow with local models? If so, in your experience what are the better models for this?

Fixed on bios, but from what I see, the dbx part is still missing in some models. They are working on it at least

Just because something is popular does not mean you will like it. Different tastes and all that.

As for me, more variety means more reach to gamers out there. What I don’t like to see is stagnation for years on established IP to the point is becomes boring (for me personally, Pokémon falls a lot into this category for most releases. The only Pokémon game I liked recently was arceus, simply because the whole battle and catch system got revamped to be something different for once)

Code your own engine. What a grown up response to all this. Expected nothing less from him.

I tried a couple of times with Jen ai and local llama, but somehow does not work that well for me.

But at the same time i have a 9070xt, so, not exactly optimal

So is not on this rack. OK because for a second I was thinking somehow you were able to run ai tasks with some sort of small cluster.

I have nowadays a 9070xt on my system. I just dabbled on this, but until now I havent been that successful. Maybe I will read more into it to understand better.

I have a question about ai usage on this: how do you do this? Every time I see ai usage some sort of 4090 or 5090 is mentioned, so I am curious what kind of ai usage you can do here