Reason to use Bluefin is to not have any proprietary software (aka Steam) included while still having all the open source bells and whistles from Universal Blue (Distrobox, Just, brew, devcontainers, fzf, ripgrep, etc).

Though I guess Bluefin does include the proprietary nvidia drivers.

That’s fair, I have definitely learned more when I add my own weird constraints. Though I also find it takes 10x longer, and what’s frustrating is that even after years of “learning” from this process, I still constantly run into these week-long time sinks troubleshooting various issues in my bespoke configurations. I still have to sift through documentation, forum posts, wiki articles. Do small scale experiments to break things down and figure out what’s going on. Just takes so much time. Perhaps not exactly tutorial hell but still feels like a hell of its own.

either it’s a kink, or a 5 year old asked an AI agent to find some people to play tag

Got any examples of some of the issues LLMs can’t solve? I still feel like an “advanced beginner” and want to know what’s beyond

Ok well there’s probably a dozen different ways to tackle this issue but I’ll let you explore them since that’s part of the fun (and I’m lazy lol)

Beautiful thank you, I’ll check them all out

maybe just “if velocity < 0.01, find which side is closest to the floor, and if its within 1mm of the floor, take the opposite side (the one facing up) as the answer”

The only edge case I can think of is if it gets close to balancing on an edge, and then finally tips over to one side or the other. But in that case I don’t think the face closest to the floor would be within 1mm of the floor, so it should still work.

What extensions do you recommend? Aside from tray icons

Can you predict the resolution when it is close enough to settled? Worst case maybe have an algorithm that estimates the probability of each face ending up on top, and if one face is overwhelmingly more probably, choose it. Otherwise let it settle using your physics engine

After years of using SO, then years of going to docs first, I’ve started going back to SO. It just tends to be faster. You can find a question that is 90% similar to yours, rather than piece together scattered bits of information across pages and pages of docs.

The government could force Huggingface to implement KYC. This would probably push people to start uploading torrents of models, but the problem is that there are dozens of quants for each model. Depending on what GPU you have, how much VRAM and RAM you have, the perfect quant for your machine might not be available on these third party download sites. So might as well grab them now while you can.

Nope, I never watched him back when I had a youtube account.

I believe you can use the command-line tool “adb” for this. It’s been a while since I’ve done it since I use syncthing now, but here’s a quick intro:

1. on phone, enable developer options
2. on computer, install android tools, which should come with adb (Do not install full android developer toolkit, it’s huge. You only need adb)
3. on computer run ‘adb devices’, which should currently print that no devices were found
4. plug phone into computer, and accept authorization prompt on phone
5. on computer run ‘adb devices’ again and it should list your phone with the status “authorized”

Now you should be able to use adb pull -p -a <folder name> to back up folders to your computer, for example adb pull -p -a /sdcard/DCIM.

Note that you will have trouble backing up the /sdcard/Android folder, which contains a lot of app data. Unfortunately I think you need root to back up that folder. Alternatively you can back up individual apps using adb backup ....

Personally I think in the long term, automatic backups like syncthing is better. Manual backups are a hassle and because of that, you won’t do them as often.

I admire the drive. And I am not here to discourage you. The privacy world could always use more driven people. There are tons of projects that can use more supporr, like I2P, Tor, Monero, Internet Archive, Whonix, Qubes, etc. Keep at it and I’m sure we can win.

I’ve seen Markiplier a few times on the Invidious feeds btw. I think I got introduced to Markiplier by some Invidious instance actually

If you believe hardware attestation is inevitable, then why bother poisoning the well for fingerprinting? None of that will matter once hardware attestation takes hold

If you use Tor Browser in the “safest” security setting, websites will actually load pretty fast since javascript will be disabled. Well, at least for the websites that load at all

I want the whole fingerprinting industry to just go away.

Be careful what you wish for. As hardware attestation continues to spread, companies won’t need fingerprints to identify you. Though if you want, hardware attestation can be thought of as an unique and unforgeable fingerprint. Poisoning the well won’t work for it.

Just to clarify, hardware attestation is not bad in itself. It can be used by the user to secure their device (aka secure boot). It’s when third-parties try to use it, where it becomes a threat to privacy. Like a website asking for your hardware-attested device id.

Of course we don’t want to give our id to these third parties. But we need to convince everybody else to do the same. Because if everybody else gives up their id, then websites won’t care about the few “privacy extremists”. And those people will simply be locked out of the web.

But anyways, back to fingerprinting. I feel like the best way to get a sense of the scale of the issue, is to simply dive into the developer discussions around it. Look into mailing lists and feature discussions for Tor browser, Mullvad browser, and Brave. I guarantee you there are discussions around trying to make the browser look identical to the average user, and the challenges of doing so.

But there’s strategy to this. What I’m saying is that it’s more productive for privacy advocates to spend their efforts convincing others that privacy is important, or pushing for privacy-friendly regulation, rather than trying to fight a war of attrition with big tech.

Adversarial methods like that won’t work. Google will know when their data starts getting poisoned. Their ad targeting will lose effectiveness, their profit margin impacted. And they will tweak and tighten their trackers until they can squeeze out a good fingerprint again. In the end Google simply rolls out Web Environment Integrity, and consumers won’t be able to use the web unless they prove their identity. If you want to fight big tech, this is the endgame they will push toward. If you want to play a cat and mouse game with big tech, big tech wins in the end since they have all the money.

Tor Browser recognizes this. So Tor Browser just erases any existing identifiers, while accepting that websites can identify Tor users and block them if they want. It’s a very explicit signal that “this is a user that cares about privacy”, and it’s up to the website to accept that user or not. If a website doesn’t care, then usually it will work fine in Tor Browser. If a website does care and tries to block Tor users, then Tor Browser doesn’t bother fighting it. There are better places for them to focus their energy.

Ultimately if you want the web to be private, the only solution is to convince everybody else to want the same. Websites won’t block privacy-seeking users, if everybody is a privacy-seeking user. Then, even big tech will have to concede.