Perpetual motion is everywhere in space. Using that motion for doing work will always change the motion, and it will eventually no longer be useful. This is what a perpetual motion machine tries to do but can not.

The great thing about foss is you can call it whatever you want, fork it even if you feel so inclined.

The best part of vba is you can record a Macro then edit it. I think you should start by adding a button that shows a message box when you press it, then record a Macro and throw some message boxes in it to see what values are used. Beware, office 365 has some other thing besides vba that I’m not familiar with, you may be better off learning that.

Phone=downtime Computer=work

Sometimes work is for the man, sometimes work is for me!

I’m all-in on oat milk in coffee, but there are many gross oat-waters masquerading as oat milk on the market. My favorite so far is the oatly barista edition, usually stocked near other coffe creamer.

Plus this oat milk is actually good, unlike what some people have experienced with store bought oat milk.

Nixos, meaning to try Gnuix but I got projects to finish!

I am a nixos fan, I do tell people to use nixos, but for some reason all these switch to nixos posts feel sus. Idk.

If you’re writing a program, definitely multiple threads or processes that each scan a chunk of the file, which basically means seek to the start of the chunk, read lines into the scan code until you hit the end of the chunk. For jsonl each chunk will need an alignment step to not break the jsonl.

For command line trickery, maybe the file could be chunked up by running multiple dd instances with an offset parameter piped into grep. This has many synchronization issues and all the outputs should be captured separately then combined afterwards. I can’t think of a good way to align this method to line edges but maybe you can put some fancy regular expression magic into the grep step to ignore malformed json at the beginning and end and overlap the chunks?

Grep is fast already, maybe test the simple approach and see how long it takes.

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

I wouldn’t say it’s the only sane way, but it’s certainly my favorite

I think nix just has name recognition, that’s why I use nixos not guix :/

Guix really should be called Gnuix, maybe that name was already taken.

What is this, open source software for ants!?!

This whole story is ridiculous. Put it behind a compile flag and merge it, we all know first across the finish line gets bonus 5 years of standardization.

It’s the law that’s a problem, not the software.

If you can see how the other machines on the network access the drives, like with net use in the windows command prompt, then you will know the protocol and address and share name to connect to.

If it’s something like \\computername\a\path\here that’s a windows file share which you can connect to with samba. On modern Linux desktop systems you can sometimes get away with opening a file browser and typing in a location like smb://computername/a/path/here and it may just work

Keep in mind computername could be an IP address instead, and some file browsers are sneaky about letting you type in a path (nautilus/gnome, which I think is Ctrl+L) if smb://… Doesn’t work smbfs:// may be worth a try

Wait, this makes no sense. Just ask your AI to write the libraries you need rather than clone an existing project

Now the next gpl needs to protect code written based on the docs

Stucco includes a layer of chicken wire like materiel, sometimes actual chickenwire

Violence

Yes, you’ll be fine

~/.local/share/user-age

Boom, UNIX