While valve has a lot of deserved goodwill, that’s always the problem - they’re well-behaved, but set up in a way in which the customer has no leverage if they where to change their approach tommorow.
Good thing drm-free games run just as well on the steam deck.
At the moment, LLMs just aren’t very good at writing anything that is interesting. I experimented with it a bit for shits and giggles, and tried out several different local Models and online Services.
I’m not saying that it’s impossible it’ll improve, but for me as someone who enjoys writing, having your writing done by a tool just misses the point. I like to write because it allows me to express myself, and off loading parts of that process to a tool makes it less personal, less me.
I won’t judge anyone with a different opinion, but for me, part of the enjoyment of reading also comes from seeing how the author and their experiences colour their writing, which usage of such a tool, in a way, also diminishes. At the moment, I just can’t see an avenue to the prevalence of LLMs making creative writing better.
Is the VR streaming in the Local net (PC to Headset)? Just run the WiFi router without plugging it into the wall. Connect only the pc and the headset.
Also, appart from that, to use more wired devices, maybe use an unmanaged switch. Don’t think that does anything forbidden here.
Full disk encryption always seemed the most sensible to me, but I’m not sure whether that needs to be decrypted after hibernation.
That’s pretty much my ThinkPad’s Specs. Fine for almost all stuff I have to do on the go (expect CAD, don’t try to run BricsCAD on the thing, it’ll make you go crazy.)
I use full disk encryption on it, as on all my other devices, and it’s fine, speed-wise. The SSD is NVME, not SATA, but I doubt the performance impact would be noticeable on a SATA SSD if that’s what you’ve got.
It’s a MX Master 2S, funnily enough. I still have a over 10 year old working M705 Marathon, on second thought, that I had once bought for my laptop. Had to open it up and bend the mechanism for the left click back into shape once, but no Problems besides.
Uh, what would I be paying for, exactly? I don’t really see what Software support a mouse really needs, as long as it doesn’t ship buggy. Also, I’ve been using my (Logitech, funnily) mouse for 6 years now, and if you ignore the few scratches it has gathered, it still works pretty much perfectly.
Also, if their solution for a longer lasting mouse really is repairability, isn’t that just their way of saying “we designed our other products to be thrown away”?
I think that assuming that editorial decisions are never influenced by financial interests would be naive, but they’re such a big organisation that covers such a breadth of topics that it would also seem foolish to assume a douplicitous intent behind every story. It might just be journalist covering a currently relatively widely discussed topic.
Also, Reuters generally does quite well in remaining relatively neutral in their coverage (though that impression might of course just be based on my biases).
Well, producing illegal drugs seems to be generally rather high risk, high reward. You’d also need a lab, possibly employees, a distribution network, and might encounter potentially rather violent competition, though, so I’d say there might be a few more cost centres other than the raw materials.
I mean, we all knew it was quite easy, but I still think that it’s journalistically valuable to go through with it to see, and show how easy it actually is.
It sounds like you don’t necessarily like the idea of using a container (I tend to use podman, but most guides are for docker, so that’d probably be easier for you). From my experience, containerising things actually makes things a lot easier, especially in the long run, and getting started is a lot easier than it seems. You can probably find a ready-made guide to set up a plex or jellyfin container on Debian.
I wouldn’t even say that. Even if they had a truly unique LLM that ran partially locally with a custom co-processor, Android might still have been a good choice. It’s just hard to beat an open source base that’s already compatible with most mobile hardware, and relatively easy to find Devs for.
Using Android as a base was honestly the most reasonable thing they did. No reason to reinvent the wheel. What they made with it is admittedly really shit, though.
Huh, I’ve been in that train. Sudden, random hit of Nostalgia.
Well, shit, there goes my vote.
A modern replacement for OpenScan. It’s workable, but some features don’t work on Modern Android, and a good Scanner app is probably something most people could use. Could look at Adobe Scan and Office Lens for feature inspiration.
Logseq uses a bit of a different paradigm though. It is cool, but I wouldn’t say it’s a drop in replacement.
I want to buy the whole of Switzerland.