There was talk of establishing a national ID in the 90s but republicans revolted saying it would be the number of the beast. I guess their principles on the matter depend on whether the beast in question is Trump or Clinton. There would definitely be scope creep for such an ID, much like the SSN. I say just make it a national ID and eliminate the need to register to vote.

He was convicted for one murder so the serial killing is what he is accused of.

Tangentially related to the article: To get an idea of how this data helps aggregators, ask ChatGPT to develop a complete psychological profile of yourself.

My usage history consists of requests for basic programming questions, help wording certainty things, scripts, reviewing documentation (eg organization policy), and a couple things goofing off (eg put my cat in a tuxedo). Just based on the types of questions I’m asking, and how I ask questions, and how I redirect it’s output via follow up questions, it was able to put together a profile that was surprisingly accurate.

It had a few things wrong, notably related to how organized or self disciplined I am. But if it had my calendar, by browser history, my biometric data (eg watch info), my location, it would easily be able to fill in the gaps. If the system only existed to help me, it would be tremendously valuable. But since the owners of this tech are all adversaries I am terrified for the future. The idea of the US federal government using it to profile citizens is too much to contemplate.

I understood you and had the same issue. I solved it by using an Apple USBC to mini jack audio device instead of onboard. Not ideal. Not sure if it’s still a problem though.

Edit. My mistake I had the problem with Pulse not Pipewire.

I find chat gpt to be useful for deciphering error messages, generally noticing things I missed. It’s also helpful for boilerplate code, essentially a customized form of stack overflow. Beyond that it really has trouble knowing what it’s doing. I can’t imagine trusting it for completing a large project. Helpful for minor optimizations though.

In this case the web server will temporarily exceed 1GB RAM and crash. It’s usually hovering around 700MB. I configured 4GB swap but I haven’t seen it use more than 1.5GB. I honestly doubt they would even have a way to know, but thanks for the warning.

Swap is super helpful if you’re running a web server on a 1GB RAM 2 CPU Linode instance!!

The syntax is easy, but the options change a bit depending on what you want to do. My entire job I requires me to use my brain so I don’t mind cheating when it doesn’t really matter as in this case. In my case I wanted a SQLite database to store URLs and playlist IDs for recording attempts and to make sure I don’t download the same video multiple times. I think I also had songs run thru music brainz for audio fingerprinting and mp3 tags. ChatGPT doesn’t get it right the first time but often gives a reasonable boilerplate piece of code as a template to start from.

Try 1/1/1969.

I have zero interest in learning all the intricacies of ffmpeg so I find ChatGPT to be very useful. I’ve also used it for yt-dlp for downloading videos and converting the audio to mp3. Very useful. I personally save them as bash scripts so I can just input the file name or url as a command line argument. On Mac you can also wrap your bash scripts in AppleScript if you want to make applets for these functions. ChatGPT works great for apple script as well but I’ve had to feed it source code (eg from Apple Digital Mastering applets) to ensure it writes the new code correctly. You still must know what you’re doing.

Sounds like a job for JoMiran! Rooting for you!

I think a lot of the problem is every tutorial expects Fedora/RedHat/Ubuntu/Debian and it’s easy to figure out which instructions are compatible with your distribution, but there isn’t a good knowledge base for Fedora Atomic or related OS. I have a Bazzite VM. Normally I use Ubuntu and am familiar with RHEL compatible, but am constantly lost with Bazzite, trying to use the wrong instructions.

My kid has an Apple Watch but because you can’t play a game on it he often leaves it charging by his bed. He’s also responsible and sets alarms for his chores which means I’m often the only one around when they go off. Sometimes I think about turning it off and hiding it.

I’m all in on the Apple ecosystem but won’t get an Apple Watch for that reason. Besides the fact that I already get distracted enough, I can’t bear to have a wearable that won’t make it 24 hours without a charge. Resistance to that level of obligation is the same reason I never bought a Tamagotchi.

It should be encrypted by default because most people don’t take care to dispose of their machines responsibly. I picked up a few machines destined for ewaste and the hard drives were full of tax returns.

I needed to make a docker image based on Core OS (RedHat) and the docker host had to be RHEL compatible. My machine is Ubuntu. To get it to work, I installed Rocky Linux on LXC and docker inside that machine. Turns out there are a lot of security settings isolating LXC and restricting nested virtualization, but fortunately Canonical posts a 20 minute video explaining how to modify the permissions for that use case. I cannot imagine virtualizing much further without the machine refusing to comply!

Did you actually get that to run or is this a fun thought exercise? It seems like a lot of nested virtualization. If you’re clever enough maybe you could get Windows > WSL > WSL Wayland compatibility layer > Ubuntu Wayland session > LXC > Fedora > QEMU > macOS > Wine > Windows app

If you need to reinstall your OS you don’t have to mess with the home drive. I use Linux for work and some of my clients actually require all data to be stored on a separate disk or partition from the applications. It also makes your backup strategy simpler and is transparent to you as a user.

2TB is too much space for an OS disk, especially since you’re not going to dual boot, so might as well get a bigger data directory and speed.

My workstation is a PCIE Gen 4 Threadripper. I’ve got a multifunction card with a couple 2TB Gen 3 NVMe drives that I striped and the bandwidth is identical to a single Gen4 4TB NVMe. Obviously you’d need a backup strategy to handle the case of a drive failing but that is true no matter what.

My workstation runs Ubuntu 22.04 with an AMD GPU, but I use an NVIDIA GPU (A4000 which is basically a 3070) for VFIO virtual machines, mostly windows. I did try Debian 12 vm with VFIO and had zero issues getting the Nvidia card set up. My VMs have secure boot /TPM enabled so no problems there either. I don’t remember the steps I took but basically disable secure boot in bios, install the proprietary driver, update the kernel, reenable secure boot. Debian was the easiest Linux distribution I tried to get set up. I also tried Ubuntu 23.10 and that worked ok. I think Fedora was OK but cannot remember. Bazzite surprisingly was a fail.

Also when all else fails, check the arch wiki. Obviously not tuned to Debian but generally most things you can figure out and the documentation is top notch.

Also wanted to mention if you’re not striping those Firecudas, definitely assign one of them to your home directory. If you do stripe, I’d create a 3.5TB home directory and leave 500 GB for / and your swap file.

Good luck.

ETA: in my experience, drivers either work right away or not at all so good news is that if your setup fails, it should fail fast, unlike windows that tries to find a workaround for janky configurations.

I still don’t see how a standalone web browser survives financially. It seems like Firefox is always near death and has to make compromising decisions. Do you have any thoughts on how this ought to work?