I wouldn’t doubt that LLMs got some special input to deal with the specific examples of this paper, or similar enough.
I wouldn’t doubt that LLMs got some special input to deal with the specific examples of this paper, or similar enough.
Probably good to add a /s somewhere here.
I suspect people are down voting without checking the piece.
I know I would, but I saw it shared on Mastodon in a cheeky way first.
Made me think this was the good news community.
People, shall we read the full article first?
Meanwhile, this is not the case with the Ryzen 9000 series desktop parts as the spec sheet of that says:
OS Support
Windows 11 - 64-Bit Edition , Windows 10 - 64-Bit Edition , RHEL x86 64-Bit , Ubuntu x86 64-Bit
So the new Ryzen AI chips that most people don’t care about won’t support Win 10, but Ryzen 9000 (the real deal desktop chips) will.
To be frank, the article title is misleading at best.
Same for the teenager part.
The Podcast (Dialogues instead of Monologues) is content for the older audience, though: https://youtube.com/playlist?list=PLet00UQnlQoUKqSB5-oFmrwpnnVc4C4A8
KeepassXC replied on that thread that it wasn’t just the privacy problematic networking that was removed:
that bug report is bunk. He removed ALL features, not just networking. That includes yubikey support, auto-type and browser integration.
Every 4-5 seconds? Yeah, logging.
You can either move the system dataset to your boot drive/pool or syslog to /var/log:
https://www.truenas.com/docs/core/coretutorials/systemconfiguration/settingthesystemdataset/
I’ve seem many users recommend a reboot after changing those settings.
Boy, are the example story and picture bad.
Out of curiosity, which is the launcher you chose?
I froze the stock launcher at the last version without Ads. It works, but I guess some entries like “continue watching” could be bugging out now because it is out of date/never patched.
Trying to make real and good use of AI generative models are cracks in the magic.
I haven’t used it, but it is on my bookmarks for when Feedly stops working for me: https://github.com/FreshRSS/FreshRSS
“cheat”, “lie”, “cover up”… Assigning human behavior to Stochastic Parrots again, aren’t we Jimmy?
Sounds like security through obscurity to me.
Highly susceptible to replay and man in the middle attacks.
If you’re gonna combine that with another authentication method (and you should), then I see little advantage over just going with the other auth method.
Is this you? 😜 https://www.youtube.com/watch?v=9kaIXkImCAM
Joke aside, I’d rather use a tool like MKVToolNix.
Blame Altman on that one, from the article: