xcjs @ xcjs @programming.dev

Posts

0

Comments

100

Joined

2 yr. ago

That's not how distillation works if I understand what you're trying to explain.

If you distill model A to a smaller model, you just get a smaller version of model A with the same approximate distribution curve of parameters, but fewer of them. You can't distill Llama into Deepseek R1.

I've been able to run distillations of Deepseek R1 up to 70B, and they're all censored still. There is a version of Deepseek R1 "patched" with western values called R1-1776 that will answer topics censored by the Chinese government, however.

The client is open source and can be administered using the open source Headscale server. I use it with Keycloak as an auth gateway.

It is! It's a port of OpenSSH. The server has been ported as well, but requires installation as a "Windows Feature".

Windows now has an SSH client built in.

They do, actually!

It does!

I've had this issue on several Pixel devices - I most recently had this exact issue on my Pixel 6 Pro, but I think recent-ish Firefox updates resolved some of the issues.

In my case, Firefox would lockup, but I could easily switch to another app without an issue. I would have to kill the Firefox app to get it responsive again.

Getting Keycloak and Headscale working together.

But I did it after three weeks.

I captured my efforts in a set of interdependent Ansible roles so I never have to do it again.

I still have it on my Pixel 8 Pro. It requires a double tap to occur in less than 300 milliseconds.

This bug has been the bane of my existence for almost four years now: https://issuetracker.google.com/issues/204650736

The Android version of the app still has the zoom/cursor offset bug when using a software keyboard from when they sunset RDP 8. That has been a severe usability bug for over three years now.

CSV only exports data, not formulas. I don't really consider it a proper spreadsheet interchange format.

It would be extremely barebones, but you can do something like this with Pandoc.

Not the person you're debating (and I'm on your side here), but what's up with all the revisionist history going on lately?

"This thing you're arguing for was never the intent."

Then what was the intent you dimwit?

And they never have an answer aside from acting like it was some grave oversight that was only recently caught as a mistake.

That is exactly a function of a jury.

"What you say disagrees with my world view, so I'm just going to pretend you're crazy and your words don't make sense."

I've had this exact tactic used against me - it's very transparent when used and weakens your position.

I get what you're saying, and yet it exists and a term exists for it.

I know there's no "nullification" verdict and the binary guilty/not guilty are the only recognized options, but nullification is used to describe the not guilty verdict despite any charges and evidence in a trial, which I'm sure you understand.

I don't think her decision to take the deal took into account whether jury nullification exists or not. The way you explained it sounds like retrocausality, though I don't know if that's the way you meant it.

Jury nullification isn't about fair outcomes, I should clarify, but about whether the law itself is lawful, representative of the people, or applied lawfully. Maybe that fits into the definition of fair I had in mind, but I was thinking on it more objectively, not subjectively.

There are proponents and opponents within the United States, true, but if a legal system does not permit punishment of jurors, then jury nullification is a logical byproduct of the system. And an important one I would argue. It fits into why trials by jury are important in a democratic legal system - the people have the final say, whether they realize it or not.

Whether a jury feels a charge is fair is the whole reason trial by a jury of peers exists.

It's a feature of the system, not a bug.

It'll probably be there, but at least it can be disabled in the settings now. It won't go away on its own.