It’s a ploy from Microsoft to push Copilot: you have to ask it to open the programs or folders you want. /s

FFS… I’m never going to buy anything hyped by LTT before doing more thorough research on it.

Have you tried Piper’s speech models? You can use Pied to install a model and set Orca to use it by default. The French one is very good.

I actually switched from Action Laucnher to Nova because the former kept crashing.

Does anyone have recommendations to replace Nova?

While I agree with you, I wished that Mehta would entertain the idea of spinning Chrome into an independent company, even though it is still unclear to me how a browser company can generate revenue, especially to pay the salaries of the army of engineers working on Blink.

Aw, chucks

The author did not mention that you as a end user can customize rankings for every website and build custom searches (lenses), which are in my opinion the features that make Kagi unique and more useful than other meta search engines. And to my knowledge you cannot replicate those in SearNXG.

I think the more damning part is the fact that OpenAI’s automated moderation system flagged the messages for self-harm but no human moderator ever intervened.

OpenAI claims that its moderation technology can detect self-harm content with up to 99.8 percent accuracy, the lawsuit noted, and that tech was tracking Adam’s chats in real time. In total, OpenAI flagged “213 mentions of suicide, 42 discussions of hanging, 17 references to nooses,” on Adam’s side of the conversation alone.

[…]

Ultimately, OpenAI’s system flagged “377 messages for self-harm content, with 181 scoring over 50 percent confidence and 23 over 90 percent confidence.” Over time, these flags became more frequent, the lawsuit noted, jumping from two to three “flagged messages per week in December 2024 to over 20 messages per week by April 2025.” And “beyond text analysis, OpenAI’s image recognition processed visual evidence of Adam’s crisis.” Some images were flagged as “consistent with attempted strangulation” or “fresh self-harm wounds,” but the system scored Adam’s final image of the noose as 0 percent for self-harm risk, the lawsuit alleged.

Had a human been in the loop monitoring Adam’s conversations, they may have recognized “textbook warning signs” like “increasing isolation, detailed method research, practice attempts, farewell behaviors, and explicit timeline planning.” But OpenAI’s tracking instead “never stopped any conversations with Adam” or flagged any chats for human review.

I use unbound as an upstream resolver for Pi-hole, not directly on my machines. Be aware that resolving new domains can incur some delay (about 1s), but once cached, it’s pretty quick, even for additional entries on the same domain.

1: I’m assuming (hoping) this wouldn’t be a full wipe and start over? It should just upgrade right?

Yes, you can upgrade in-place. If I recall correctly, you just have to change the release channel from LTS to non-LTS in the Sources app, then trigger an update (I don’t quite remember how to do)

2: Do I need to do the whole USB route, and if so is there an option to keep everything (I’m hoping, I put a LOT of work into this so far and I don’t remember if that was an option on first install).

No, as I said above. But it’s always a good idea to have backups if you ever need to wipe and re-install.

3: I remember a few apps I installed were specific to Noble, will this break those apps?

Hard to be 100% sure without knowing which apps and how they were installed, but most likely yes as their dependencies might no longer be available on Plucky.

4: It seems like there should be an option to upgrade from the desktop, but I don’t have that option. If I run plasma-distro-release-notifier I should get an update notification right? In which case I can just say “hell yeah!” and it’ll do its thing?

Refer to my first answer

What originally started as a git repo for storing backup scripts and a list of GNOME Shell extensions now contains dot files, systemd units, Pipewire and Wireplumber configs, scripts for installing new software from Brew and Flatpak, and a systemd service that pulls and apply the latest changes on session startup.

https://github.com/axeleroy/setup/

My reaction when I read this article

Also this interviewee sums it up quite perfectly:

“If I know from looking at company reviews or the hiring process that I will be using AI interviewing, I will just not waste my time, because I feel like it’s a cost-saving exercise more than anything,” Cobb tells Fortune. “It makes me feel like they don’t value my learning and development. It makes me question the culture of the company—are they going to cut jobs in the future because they’ve learned robots can already recruit people? What else will they outsource that to do?”

I’d pick OpenSUSE over Bazzite because I don’t like the idea of updates possibly overwriting anything I install myself that isn’t flatpak/distrobox/homebrew

In atomic distributions you would install non-sandboxed programs in a layer that is applied on top of the base system. When your system is updated, that layer is applied back on top of the updated system. The only possible breakage would be if what you installed depends on a dependency in the base system that has been removed or which is no longer compatible.

AI bros are trying really hard to convince people that their parrots can be useful in business settings.

Telegram are being paid by xAI to use their product? Isn’t it supposed to be the other way around?

You can’t, it just part of how Fedora works now. Maybe Fedora should patch Dolphin to take /sysroot into account instead of /

Fedora Atomic Desktop 42 switched to composefs, which has a small full partition mounted to /. Your “real” filesystem is mounted on /sysroot

https://fedoraproject.org/wiki/Changes/ComposefsAtomicDesktops

I remember quite well burning an Ubuntu 9.04 Live CD, and before that trying an ancient Knoppix Live CD that my dad had laying in a drawer. I must have been 15 back then.

Most probably Microsoft has set objectives for how much LoC are from LLMs and developers invented numbers to match that metric (because they probably have things more important to do than counting LoC)