Call it "open weight" if you want, but it's not "fully open". The training data is still proprietary, and the model can't be accurately reproduced. It's proprietary in the same way that llama is proprietary.
You don't need port-forwading for torrenting. It's more like a nice-to-have. Mullvad works fine without it, so don't let that stop you if you're on the fence.
There are at least three other MFA methods that are not email based, and so no, you don't have to remember your email password.
Get an authenticator app. Get an authenticator key. Or hell, go use Duo for free (not recommended). And if none of those do it for you, use your 2FA recovery code. That's what it exists for.
And if all else fails, you can still shoot yourself in the foot and opt out of the change, but you're just begging to have your passwords stolen ¯(ツ)_/¯
That's fine if you think the algorithm is the most important thing. I think the training data is equally important, and I'm so frustrated by the bastardization of the meaning of "open source" as it's applied to LLMs.
It's like if a normal software product provides a thin wrapper over a proprietary library that you must link against calling their project open source. The wrapper is open, but the actual substance of what provides the functionality isn't.
It'd be fine if we could just use more honest language like "open weight", but "open source" means something different.
Absolutely, but this is one of the worst reasons to advocate self-hosting a service. The kinds of people that are upset about higher security standards should not be self-hosting anything.
They should be using 2FA then. There are at least 3 other multifactor authentication options available. Configure one of them, or you can be affected by the device verification change. Or, you can disable the feature, but without any secondary auth factor, you're just begging to have your passwords stolen.
The training data is the important piece, and if that's not open, then it's not open source.
I don't want the data to avoid using the official one. I want the data so that so that I can reproduce the model. Without the training data, you can't reproduce the model, and if you can't do that, it's not open source.
The idea that a normal person can scrape the same amount and quality of data that any company or government can, and tune the weights enough to recreate the model is absurd.
Mental Outlaw is a reich-wing freak, so that's par for the course. Unfortunately, there are a fair amount of these shitheads in the Linux YouTube space.
Is it actually open source, or are we using the fake definition of "open source AI" that the OSI has massaged into being so corpo-friendly that the training data itself can be kept a secret?
Well, my favorite is Helix -- a text editor. But it's a TUI in the same way that vim is a TUI, and being that it's a text editor, you're likely to have very strong opinions about whatever your current favorite editor is. Which is totally fine. I'm not looking to start any editor wars.
I use way more than just those. If you do find that some TUIs may be useful, there are a bunch more here. If not, I'm still interested in what behaviors you find off-putting about TUIs. Maybe I can incorporate that feedback for TUIs that I write :)
That's fine. I'm sure there are some sub-par TUIs out there. I've seen pretty great TUIs, especially the ones written in Rust (because of the excellent TUI lib they have).
GUIs are fine too (as long as they don't use Electron lol).
Electron is one of the slowest, clunky, memory-hogging ways to have a UI, and TUIs are the exact opposite. I don't care if (name of company that ships Electron slop here) can ship your software webpage masquerading as software to more systems more easily. If your messaging "app" has input lag when I type something, it's a dogshit experience.
Of course, there are ways to ship GUIs that aren't all of the things wrong with Electron, but comparing TUIs with those is less interesting and more a question of if the person likes to live in their terminal or not.
As someone that has never used Puppet, I also wonder this. Ansible is agentless and works on basically anything. What do you gain by requiring an agent, like with this?
Call it "open weight" if you want, but it's not "fully open". The training data is still proprietary, and the model can't be accurately reproduced. It's proprietary in the same way that llama is proprietary.