Dear Linux community,
In these shite times, we need to remember to see the human in all of us. So I address the humans of this community:
-
To the Stallmanites. SCREW CORPORATE YEAHHHHH! Each day humanity grows smarter, no matter how many iPad kids or mass media diatribes there are, so celebrate how far we've fucking come. Look at your oldest machine. You're a million times better. My old Slackware was full of SHIT! FUCK YEAH
-
To the pragmatists. Linux doesn't get in your way. People get in each others' ways all the time. But most importantly, don't get in your own way. If you're feeling like shit -- the human capacity to hope beats everything to a pulp. You know what we do when shit fucks? We submit a fucking pull request. FUCK YEAH
-
To the hobbyists. Life is a bundle of fun so feel every moment. When you're old, and you'll be old as fuck, you'll be like "fuck yeah I lived every moment of my life up to now." So fuck another segfault up. Learn C, and call C the best language. Then learn Rust and call Rust the best language. Make a homelab then call homelab the best language. FUCK YAML HOLY SHIIIIIIIT
Flaming hot wishes,
Your fellow fucking human.
"Open Source" is a misnomer for LLMs. The weights are comparable to binary blobs.
The difference is nobody knows what the binary blobs are made of
You could recreate them if you had the training data, which is pretty much never disclosed.
You could of course use one of the models that are trained on open data sets. Maybe a little worse than those directly from Mistral etc but truly open source
I decided I don't care for imaginary AI copyright and will use Llama as if it is public domain.