Sebastian Krzyszkowiak @dos@librem.one

All Your Base Are Belong to LLM

"The output from an LLM is a derivative work of the data used to train the LLM.

If we fail to recognise this, or are unable to uphold this in law, copyright (and copyleft on which it depends) is dead. Copyright will still be used against us by corporations, but its utility to FOSS to preserve freedom is gone."

https://blog.brettsheffield.com/all-your-base-are-belong-to-llm

#FOSS #OpenSource #FreeSoftware #LLM #AI

@maltimore Yeah, I should probably add that some app/stacks do better than others, which is one reason why you see vastly different performance, often also depending on the used HW.

For most devices, the work here will mainly impact battery life. But especially in the #LinuxMobile and low-end area the difference is quite visible. I.e. devices like the #librem5 or #PinePhonePro can play 4k videos just fine and shouldn't heat up when doing so, just like other phones.

Laptop died yesterday but thanks to the #Librem5, the Baseus Dock, an HDMI Screen, a USB keyboard and #Phosh's docked mode I have access to most of the things via my phone.

#nitrokey helps a lot as that means I have my ssh keys available by just plugging it in.

Took me about 15min to notice that I didn't even plug a mouse in.

#LinuxMobile

We are releasing the fix in staging and it should be available within the hour. Place make sure you upgrade to 6.6.27+librem5-2 if you are in staging (If you were using the trixie repo, no need to do anything!)

Very, very happy that our proposal for adding #cellbroadcast support to #ModemManager and #Phosh was accepted by @NGIZero : https://nlnet.nl/project/SMS-CB 🗼 〰️ 📱

#LinuxMobile

A new release of #feedbackd is out (https://source.puri.sm/Librem5/feedbackd/-/releases/v0.3.0).

Besides some cleanups the user visible change is that "blessed" apps can now override the global feedback level so e.g. your alarm clock can ring even when the phone is in quiet mode (https://gitlab.gnome.org/World/Phosh/phosh/-/merge_requests/1398)

#LinuxMobile

"If you make your own engine you'll never ship a game"

I'm here to tell you this is BS. Kitsune Tails uses a custom engine built on a custom framework all made from the ground up by yours truly. Neither of these were even remotely the bottleneck for development duration. Turns out that the thing that takes the most time when you're developing a game is all the stuff that is hyper specific to that game and can't be generalized anyway

Not to mention all the, you know, actual content that sits on top of the engine. I can write a json parser and serializer from the ground up in a day or two, but all the cutscenes I had to script for KT took many weeks

Then there's the fact that if I'd used an out of the box physics solution for Kitsune Tails I'm fairly certain I'd never have been able to nail the game feel it has, which is the most core thing to the whole experience

You don't *have* to make an engine but quit pretending doing it is the hard part of making a game. It fucking ain't

Updates on the xz situation:
https://tukaani.org/xz-backdoor/

https://lore.kernel.org/lkml/20240330144848.102a1e8c@kaneli/

And there's a new commit in the git repo:
https://git.tukaani.org/?p=xz.git;a=commitdiff;h=f9cf4c05edd14dedfe63833f8ccbe41b55823b00

Hey it's totally cool that #microsoft #GitHub blocked access to one of the repositories in the very center of the xz backdoor saga.

It's not like a bunch of people are scrambling to try and make sense of all this right now, or that specific commits got linked to directly from media and blogposts and the like.

Cool, cool.

I know the next 3-7 days will be filled with exaggeration and doomsday talk, but IMHO the #xz backdoor, though seemingly meticulously planned for a long time, failed miserably as it was caught at a stage where it wasn't widely deployed but only in testing/prerelease distros. Yes, it made it quite far in the supply chain but it ultimately failed. The mess is being cleaned up, no cases of actual use of the exploit in the wild are known thus far. The immune system of FOSS has worked. Again.

damn this lzma/xz backdoor is wild

https://www.openwall.com/lists/oss-security/2024/03/29/4

First known symbolic gesture observed in birds; scientists recorded the Japanese tit using repetitive wing movements to signal "after you" to others.

https://www.u-tokyo.ac.jp/focus/en/press/z0508_00339.html?utm_source=join1440&utm_medium=email&utm_placement=newsletter

I'm at an EU workshop for Apple to discuss and justify its DMA compliance plans. They've been forced into 3 huge u-turns by EU (killing PWAs, removing Epic Games' developer licenses, sideloading) so I'm interested to see their demeanour. Will it be humility, or the usual 'fuck you, we're Apple'? There are people from EU companies that Apple blithely planned to bankrupt by sneakily removing PWAs here. And regulators have human feelings too; no-one like "fuck you". I'll be here all day (try the baguettes!)

It always gives me joy when I see that people find useful the /sys/kernel/debug/devices_deferred sysfs entry that I only added to scratch my own itch 🙂

This time in a FOSDEM talk by @a3f about the Linux kernel drivers and devices matching mechanism: https://fosdem.org/2024/schedule/event/fosdem-2024-3222-linux-matchmaking-helping-devices-and-drivers-find-each-other/

#linux #kernel #FOSDEM

With a bit of help from #phoc the next phosh-osk-stub release can (optionally) hide when a keyboard gets attached.

#phosh #LinuxMobile