Old Fucking Punk @lwriemen@librem.one

Social.‌coop meetings: https://link.social.coop/calendar
Organizing Circle
Thursday, Nov. 5, 2026
7:00 AM PT | 10:00 AM ET | 15:00 UTC

Social.coop members are welcome to observe this and all upcoming meetings.

More excellent news from The Juice Media:

https://www.youtube.com/watch?v=v_Lua2Xzg5s

:D

@rhelune I suspect a big part of it is sunk cost fallacy. "Waaaa! Nooooo! You have to use this because we've spent so much money on it! If you say it isn't helping then that means I'm not a big-brain money guy! Waaaaaa!"

Why are so many workplaces trying to force employees to use generative "AI" (chatbots, coding agents), even by threatening their livelihoods?

#AI #noAI #LLM #LLMs #vibeCoding #technofascism #workersRights

I don’t intend to call out anyone on here with this, but it’s important to reflect:

- I was told that both the slop PHP patch and the slop kernel patchset were high quality code with very few telltale signs of slop
- both of them had a bunch of telltale signs of slop (and again I’m really not trying to call anyone out for this, please consider this a postmortem)
- both were rushed through with very little to no understanding of the problem they were solving
- both were highly publicized but solved little to nothing: the linux patchset was flawed and decreased performance outside of microbenchmarks, and the php patch fixed an issue that was hard-to-impossible to exploit in practice with a gigantic number of slop-looking changes when another fix (abort on OOM now and gradually patch exploitable cases) was available, but nobody considered it because nobody considered anything

just some things to consider going forward. you absolutely don’t gotta hand it to any of these people.

@davidgerard LLM generated tests - they pass, so they must be good

@davidgerard i love qualitq

slop tests always work so amazing every time, like when torvalds slopped out a performance regression and the tests went “looks good boss”

PHP generate their fucking *tests* with a chatbot

that's how you know the software will be quality with a q

@rhelune I've been quite happy with IVPN since ditching Proton. Off the top of my head I think it matches what you're after, but I haven't double-checked.

@marcusxms I was born in Yugoslavia where workers' self-management was the norm. Factories to the workers!

"Employees navigating similar pressures at other companies should be watching closely, because whatever Proton works out here is almost certainly coming to their performance review next."

https://tech.yahoo.com/ai/articles/ai-unemployment-proton-interviewing-employees-145827415.html

In December my 2-year VPN subscription is running out. I had planned to switch to 22Mullvad, but Mullvad has come out as 50% nazi since.

Which VPN should I use, that works on Linux, Android, and GrapheneOS? I use VPN for everything I do online, except a few things that do not work with it, and German state services. I do not trust German ISPs and I do not want them to know what I do. Therefore the VPN has to be reasonably fast, private (no logging, independently audited), non-fascist (of course), no DNS leaks.

#AI #noAI #proton #VPN #technofascism #privacy #security

@Niklas @jlink @argv_minus_one I actually have a problem with the effective part the way they present it. Someone calculated that mosquito nets save the maximum number of lives per dollar donated, and most EA started donating for mosquito nets, before they switched to "AGI" of course. But that would result in an excess of mosquito nets while people in other parts of the world die of other causes. So what would be the most effective way to donate? Mathematically, calculate what is needed where, for example x USD for mosquito nets, x/2 USD for a certain vaccine, 3x USD for famine relief, and then, roll a dice, literally let a random number generator decide where your money is going to go based on those weights.

I boosted your toot in case anyone knows of such an organization. I myself am a very ineffective donator by their standards, I am more of a mutual aid person and support my Fedi comrades who need help (I do donate to NGOs, but mostly local ones, "think globally, act locally").

However, because there is enough money/resources in the world for everyone to have a good life, they are just being hoarded by the few, I am in favour of Piñata Economics: piñata the billionaires until their wealth trickles down.

Out of the ‘AI’ Crisis – Premium Newsletter launch and Halloween Sale

I'm trying a new thing and launching a premium newsletter in addition to the regular one.

I'm hoping this will help pay for the work that goes into the newsletter, videos, and podcasts, and times are genuinely a bit rough these days, so I'd very much appreciate it if people could spread the word 🙂

https://www.baldurbjarnason.com/2026/out-of-the-ai-crisis-halloween-sale-and-launch/

"Neglect a system badly enough, or long enough, or incautiously enough, and it will be at risk of failure. Any sufficiently advanced negligence is indistinguishable from malice."

Deb Chachra, How Infrastructure Works (pg 161)

I would be very grateful if people out there could stop saying that "every FOSS project is turning towards AI" just because some prominent examples are.

My projects are not. And for OctoPrint I'm actually doubling down on the "made by humans" point in the upcoming website relaunch and am also considering adopting CASA¹

But it's really exhausting to fight this shit at two fronts.

¹ https://destructured.net/casa

“Anthropic files its ‘IPO’ — but all the key numbers are missing – Pivot to AI”

https://pivot-to-ai.com/2026/10/27/anthropic-files-its-ipo-but-all-the-key-numbers-are-missing/

> As a serious business proposition, you’d be nuts to go for this.

What he said.

One thing I've noticed AI whistleblowers NOT saying is that a technology capable of autonomously destroying humanity is also a technology that could be used by its makers to destroy humanity, or some disfavored portion of humanity.

For some reason, they really want us to focus on that "autonomous" part.

RE: https://mastodon.online/@tomstafford/117517506699981465

important point

RE: https://sfba.social/@drahardja/117516108521286107

I stopped working on control systems back in my early/mid 20's when BAe/Rolls Royce couldn't promise the code I wrote wouldn't be used to test weapons.

I totally get the desire to work on cool shit but Boston Dynamics was always going to sell more systems to repress people than do search and rescue missions just look at the design of their robots. They are offensive tools.

So a #Tesla semi spontaneously burst into flames this afternoon and shut down a major highway in #Hayward #sfba #california until tomorrow.

“The CHP said at 7:45 p.m. that the fire was expected to burn for between 15 and 18 hours and that the southbound lanes of I-880 would remain closed until it was out.”

https://abc7news.com/post/tesla-semi-truck-bursts-flames-shuts-down-880-hayward/19988258/