echo b > /proc/sysrq-trigger

For all this trouble I would rather give bad credentials inside the container and replace them with real credentials in the proxy.

Local LLMs and learning CUDA on a machine with 6 1070Ti GPUs… Some of it works out of the box… For other stuff I’m having to explore what will work on the sm61 cuda architecture.

Also getting better at running a pair of old 3d printers.

I’m not one to buy new hardware if you can’t tell :P

Many fancy motherboards have a button to reset these settings as well. Often the bios settings are referred to as CMOS settings as well.

Definitely friend, especially the spineless variety

Five syllables then
Seven syllables and then
Five more at the end

And just like that gold is a proof of work currency. Too bad those economics will change as gold becomes less scarce. Buy mercury now!

Don’t be afraid of the command line, breaking Linux is how you end up learning how to use it!

I haven’t done this tutorial but if that kind of thing helps you this one looks pretty good.

My best guess is you need to do something like:

(In the shell, one line at a time, enter runs the command)

mkdir /mnt/tmp
mount /dev/sda2 /mnt/tmp
nano /mnt/tmp/etc/fstab

Nano is a text editor that uses your whole terminal, so you will see the contents of /mnt/tmp/etc/fstab (the file that controls where disks are mounted) and replace ‘sdb’ with ‘sda’ on the line starting with /dev/sdb2. The bottom of nano’s screen shows you the keyboard shortcuts, I think Ctrl W will make it write the file, asking for confirmation of the filename, which should stay the same. Exit nano (Ctrl+x maybe?) then reboot with the command ‘reboot’

If you get any errors about access denied or permissions, run ‘sudo bash’ to get a shell with more power and try again.

Good luck!

What most likely happened is your disk order switched and, as others have mentioned, using /dev/sda1 or something similar to point to partitions is unstable and can’t be trusted. Once your system is back up, look up how to specify partitions in /etc/fstab using UUID (something like /dev/disks/by-uuid/xxxx-xxxxxxxxxx-xxxx instead of /dev/sda2)

When this happens to me I mostly assume Linux shutdown automatically because of a critical over temperature event. I’ve seen it in the logs a few times but I don’t usually check anymore.

There’s an example of this here https://unix.stackexchange.com/questions/502226/how-do-you-find-out-if-a-linux-machine-overheated-before-the-previous-boot-and-w

Pretty sus :P

I would start by removing the graphics card if you have integrated graphics available (or disable the PCI port in your bios)

This reminds me of the kinds of issues I would get when setting up overclocking and getting just past the limit of stable operation. If you have overclocking set up definitely try disabling it.

If removing the GPU does nothing don’t forget to check removing each ram stick separately, or make sure your bios runs a full memory check.

Let’s you turn sideways for a little dip in the pool when you’re done

Ah mirroring Wikipedia has been on my to-do list for too long, time to buy some storage.

Ya and taking off the serial number doesn’t really matter when it’s less than 10 😅

… I said to myself as I meticulously removed all traces of the serial number.

I didn’t believe it at first 🤣

😅 what year is this guy living in? SQL has scaled, probably better than any other software paradigm.

I wouldn’t trust current models to do any real work. Aaand I think humans will be cheaper than LLMs for a long time to come. Ultimately all costs are labor and if you need to give the power plant people (running the plant, mining the fuel, building the plant) sandwiches to get them to provide power for your llm, you’re probably better off giving a human programmer sandwiches instead.

The ai bubble pops when investors decide they want dividends instead of speculative gains.

At least it’s not SpaceX

Any electronic thing