

Oof, OK. I have seen some S1 Andor deep-dives in the past which were genuinely good. Haven’t watched anything on S2 yet, because I didn’t like it.


Oof, OK. I have seen some S1 Andor deep-dives in the past which were genuinely good. Haven’t watched anything on S2 yet, because I didn’t like it.


Which one? Mind sharing?


Huh, not what I would have expected. I work for a company that has sadly shifted very AI-focused, with the exception of the actual engineers. Literally none of us likes or uses LLMs. Every other week someone from the C-suite reminds us that we are encouraged to use it, and get 300$ or some such in credits for AI tooling per month, and that they don’t understand why it hasn’t been claimed even once.


I’m so torn on this, because IN THEORY the argument “git blame should show the dunce who committed this” makes sense.
But then why not add the AI as a co-committer.
(All of this of course sidesteps the actual question, “why the fuck are you allowing AI contributions in the first place”.)


a number of these things are actual security exploits
yes, sure, but in the sense of “social engineering the slopbot”, not “there’s a vulnerability in [x] which does not apply proper sanitation to non-body text content before evaluating it”


Ah, makes sense. My mind was going more towards actual security exploits, but… yeah. Makes sense.
Anyways, thank you Flere-Imsaho, I will never not love seeing your username on here.


I’m apparently too out of coffee dumb to properly get the joke. What’s the implications if some system parses the PNG text chunks in the same way it does the body of the email?


You do have a point. TBH I only now realized that the video was posted from Doctorow’s personal account, and without a link to the “original”, which yeah, kinda weird.
The talk itself is still worth it (had the fortune of sitting in the audience), but probably a good idea to use the media.ccc.de link.


Originally/additionally hosted on media.ccc.de


Yeaaaaah I saw that on the schedule and decoded to not go, lol.


Really? Which ones? I didn’t notice any


Not the person you replied to, but I also switched to Librewolf a couple of months ago, also greatly enjoy it. Here’s a link to my nix config where Librewolf gets made to behave a little closer to default Firefox in terms of usability. Honestly not too bad!
Nah, both ways are fine. The first one just installs the package, the second one enables the module, which installs the package + does a bunch of additional setup and gives you super convenient configuration options (like setting up mail accounts declaratively from nix)
Don’t you mean
home.programs.thunderbird.enable = true;
?


Don’t forget the almighty:
journalctl -fu <servicename>
And yes, I am always reading that as “fuck you, service”.


Oh god yes
German?
My favorite math joke to this day is:
How many math professors does it take to change a lightbulb?
The answer is trivial and left as an exercise to the reader.


I’ve actually been having a great time with simple-nixos-mailserver.
Running with a dedicated ipv4 at a highly reputable hoster, to my knowledge, I haven’t landed in a spam folder yet!
oki bocchi