Thanks a lot for having taken time to respond so thoroughly. I must say I haven’t thought about things from this angle.
The part about the book and the mediocre comedian definitely rings a bell. Getting stuck in stupid local extrema (like in optimization) more often than necessary is definitely a thing with me.
What of many of the things you’re “supposed to do” are things you don’t actually want to do and therefore you don’t do? What if it’s external circumstances define what’s called a mental disorder?
Would you feel like you have a mental disorder if you lived in some completely different context?
I’m just wondering if we can call things a disorder that might mostly arise because society is built around working better for more neurotypical people (it at all).
Would you call it a disorder being tall if for some reason most people were short and all our infrastrucure were built for short people?
I’m not questioning the difficulties many people have with their lives. I’m wondering what to do about it and where the threshold is.
I often wonder as well. Then I think: is this not just the human condition. In any case I seem to score pretty high on those online questionnaires.
Nice work! What does the optics look like? Do you have a picture of the whole thing?
KDE Connect is amazing. Also works without KDE.
Not true about xmpp in general. There are modern clients out there.
What’s your problem with xmpp?
Had to install it now. Never played before. It’s a tough game.
Thanks for the interesting point! I learned something today. I guess it all depends on your use-case, whether flatpaks make sense or not.
A floss project’s success is not necessarily marked by its market share but often by the absolute benefit it gives to its users. A project with one happy user and developer can be a success.
I’m not against probabilistic models and the like. I merely try to capture part of the reason they are not always well received in the floss community.
I use LLMs regularly, and there is nothing rivalling them in many use cases.
Flatpaks won’t get their libs updated all at once by just updating a library. This can be very bad in cases like bugs in openssl. Instead of just updating one library and all other software benefiting from the fix, with flatpaks, you need to deal with updating everything manually and waiting for the vendor to actually create an update package.
I’m not 100% sure about this. Flatpak has some mechanisms that would allow to manage dependencies in a common fashion.
This and on top of being inexact, it’s not understandable and un-transparent. These are two of the top reasons to push for free software. Even if the engine executing and teaching models are free, the model itself can’t really be considered free because of its lack of transparency.
I think it’s a short term vs long term debate. In the short term snaps are nice. They might help you get that software you want right now. In the long term though, it will only take away some of your rights and make you into a product.
There are also some interesting things to say about wording. Specifically consumer vs user. Software is not consumed, it’s used and depending on the specific software, the user might be abused by the people producing and controlling the software.
I totally agree about rate limiting, mostly against bad passwords that you are not in control of. But banning failed attempts is mostly not interesting if you ask me. It feels like the right thing to do, but IP addresses can change and other measures are better.
It’s debated whether software like fail2ban actually helps or if it just makes attacks visible that would anyways fail if you have up to date software. Oftentimes, defensive software adds attack-surface because it adds more software that can be targeted by attackers.
Fail2ban might help with protecting against exploiting of bad passwords though.
Same for me
I like the looks of them
I’m interested to hear how it turns out!