“why not” “Fuck it” “Am I going to regret this? Not 80%? Do it” “Because I’m bored”

“it’s not about why, but why not”

Is there a way to get the closed source driver running? Or does the closed source one not supporting Linux?

The spec says so but it’s not true. And I just want it running (good enough even I get 1FPS). It runs on my old intel HD6000 iGPU if I stay in a small world. The Mali GPU on the SBC is almost as good as the HD6000. I expect to get at least 10FPS on that SBC.

I haven’t. There’s a official Samsung repair center next to my workplace . I gone there expecting a faster turnaround.But was informed they only process phones and earbuds. I had to mail the SSD instead.

Going to do that in the weekend. Also bought a new SSD from Kingston since is is also my work machine. Can’t have it down for too long.

Can you check your spare capacity? Run smartctl -a /dev/nvme<id>

Also which FS are you using? I have switched to BTRFS root partition to catch these errors as early as I can.

Just like how 4chan and Reddit started

Yeah, 970 Pro my bad.

I ran smartctl -a /dev/<nvme_device>

Maybe it’s my culture. I think of master as the “master of kung-fu/art/sword/”. Something or someone that have earned respect and is at the core of it’s field.

Just like “master” in git. WTH is wrong with it. I feed “master” as “the master of kung-fu” is much better then “main”.

Correct. As Lemmy (and the entre Fediverse) is decentralized. The only people can do anything are the admins of the instance. If they fail to take action. other instance owners have the choice to completely de-peer that instance. Which will make that instance and the users on it unavailable locally. At the high collateral damage. Like how Gab (IIRC was a far right instance) is blocked by most major Masdoton servers.

There is no way to globally ban any community and/or instance.

Not professional sysadmin. I run my homelab and handles a few servers at work. I don’t use IDS. So may be irrelevant.

• WAF to stop HTTP parameter pollution and request smuggling
• Fail2ban on SSH and move ssh away from port 22
• Setup LAN recursive resolver and disallowed outbound raw DNS

For me, a lot more emphasis is on defending the application

• Setup systemd unit hardening
• Use Landlock LSM to whitelist directories (modifying source needed). Stops directory traversal and command execution
• TLS or stunnel between application and database
• Point DNS to local resolver
• LD_PRELOAD hardened allocator

I’m currently looking into the Linux port of pledge to further reduce post exploit attack surface. But the project is not mature enough for production, yet.

There’s only one logical next step: WW3

btw, I use Arch

For me the saying “work on your passion and you won’t work a day” is very true. Why would I not experiment with crazy code in the product when I don’t have time in the day. Or trying out new cryptography and see how it absolutely beats current protocols.

With that said. It’s also important to have time with you family/boy/grilfriends. They are also important

That’s the exact point. reddthat.com or lemmy.world servers can blow up into pieces and none of users living on other servers would be affected. Unlike Reddit, where a single company can decide they want to do something stupid and everyone had to obey.