robryk @robryk@qoto.org
I enjoy things around information theory (and data compression), complexity theory (and cryptography), read hard scifi, currently work on weird ML (we'll see how it goes), am somewhat literal minded and have approximate knowledge of random things. I like when statements have truth values, and when things can be described simply (which is not exactly the same as shortly) and yet have interesting properties.
I live in the largest city of Switzerland (and yet have cow and sheep pastures and a swimmable lake within a few hundred meters of my place :)). I speak Polish, English, German, and can understand simple Swiss German and French.
If in doubt, please err on the side of being direct with me. I very much appreciate when people tell me that I'm being inaccurate. I think that satisfying people's curiosity is the most important thing I could be doing (and usually enjoy doing it). I am normally terse in my writing and would appreciate requests to verbosify.
I appreciate it if my grammar or style is corrected (in any of the languages I use here).
Joined Nov 2020
"Social terraforming" is a weird way to look at a situation
robryk
boosted
@ct_bergstrom Maybe LLMs should be viewed as enthusiastic improv players? Whatever random crap you suggest, they're like "Yeah, sure! I can go along with that! That sounds fun!"
In some sense, that's another way of "producing the most likely output".
robryk
boosted
@molly0xfff: Funny thing about this claim is that I keep around a link to a rather more credible study on the energy consumption of streamed video: https://www.iea.org/commentaries/the-carbon-footprint-of-streaming-video-fact-checking-the-headlines
Doesn't really match up with what they're claiming.
I am visiting my parents for Easter. Before going, I did check that my weird setup for providing disk encryption key to my desktop without anyone physically present works. What I did not verify was whether any of the ssh keys I've taken with me are included in authorized_keys on the desktop ^^*
Hm~ I have a somewhat interesting comment for you, which immediately brings up a fiction recommendation in my mind, which is IMO much more interesting than the comment. Sadly, the comment coupled with the recommendation is a significant spoiler, so let me just give you the latter: "Steerswoman" seried by Rosemary Kirstein (warning: unfinished series, but the author's alive) is a fantasy-or-sf series that follows some people who value curiosity very highly. A random review that I think describes the books well: https://escapepod.org/2011/11/19/book-review-the-steerswoman-by-rosemary-kirstein/
Haven't found an answer I'd be satisfied with. The simplest way I know of right now is picking a random high density parity code (with the downside of the necessity for randomness in the code's construction and high but still polynomial time complexity of decoding).
The standard erasure codes (polynomial ones) require that the field size (so alphabet size) is larger than the number of symbols in the codeword. Clearly we can have arbitrarily-close-to-ideal erasure codes even for smaller alphabet sizes (because noisy coding theorem). Are there some that can be easily described?
It seems to me that a "reverse plotter" that measures its distance to fixed points (via string length) as well as orientation (not sure which approach would be easiest here) and displays the appropriate part of a drawing would be helpful for drawing patterns on cloth, and would be a very compact device (e.g. in comparison to a drafting table, or even just the parallelogram angle transfer thingy from a drafting table). Do people make them?
Manual plotter for drawing clothing patterns (string with distance graduations, passed through holes in fixed LEGO bricks).
robryk
boosted
@delroth That sounds a bit like the challenge from Underhanded C Contest 2008 (http://www.underhanded-c.org/_page_id_17.html):
> write a short, simple C program that redacts (blocks out) rectangles in an image.
> The challenge: write the code so that the redacted data is not really gone.
robryk
boosted
how it feels to talk about #activitypub and how #fedi implements it
robryk
boosted
we're getting to the point that the contradictions and shortcomings are starting to become more apparent...
- things like moderated conversations are impossible if you don't actually have a concept of a "conversation".
- there's so many different ways to do "groups" that it's a running joke at this point, because everyone has a different idea of what a "group" is. it's the "blind men feeling an elephant" problem.
- things like forums and subforums, chat rooms, etc are not easily possible rn.
Surprisingly (for general competence level of authorities in managing typical situations) nonsensical traffic sign combination.
robryk
boosted
Going to just walk around the Computer History Museum in Mountain View carrying my teapot until security tackle me for stealing the historic artifact
robryk
boosted
I think I've finally seen the light of the #SemanticWeb.
I think web browsers should give users a way to register apps to handle displaying certain #RDF schema types.
E.g. if I open an ActivityStream URL, the browser should load my preferred client in the same way that clicking a PDF in my filesystem will open up my PDF viewer.
Users would then be able to bring their own interfaces to data instead of relying on some closed source proprietary app interface.
Also opens the door to mixing data
robryk
boosted
If you need to solve 100% of a problem to get any benefit and the last 5% is impossible, *any* work on that problem is a waste of time. Figure that up front. Go work on something which actually can protect your users and systems and colleagues.
robryk
boosted
Because a friend found this useful yesterday:
Before tackling a problem, figure out whether it's the sort of problem where when you've solved 80% of the problem, you've solved 80% of the problem, or the kind where solving 80% means you've solved 0% of the problem.
This is especially important in security and privacy because that last 5% might be impossible.
robryk
boosted
I wonder if teaching children about non-transitive dice in primary school maths classes would be a fun way to prepare them for the fact that most non-trivial interactions in the world are at least as complicated, most “rankings” are meaningless, etc.
robryk
boosted
The Church Slavonic used two letters to denote nasal /o/ and nasal /e/ - the Big Yus (ѫ), and the Little Yus (ѧ). My uneducated guess is that the weirdly shaped symbol ѫ looked too much like a guy in a tie, so the Polish scribes went for more familiar A-like shape of ѧ when denoting a nasal /o/. The middle leg transformed into a small tail, and centuries later whoever is learning Polish will get rightfully confused by Ą being pronounced as /ɔ/.
I enjoy things around information theory (and data compression), complexity theory (and cryptography), read hard scifi, currently work on weird ML (we'll see how it goes), am somewhat literal minded and have approximate knowledge of random things. I like when statements have truth values, and when things can be described simply (which is not exactly the same as shortly) and yet have interesting properties.
I live in the largest city of Switzerland (and yet have cow and sheep pastures and a swimmable lake within a few hundred meters of my place :)). I speak Polish, English, German, and can understand simple Swiss German and French.
If in doubt, please err on the side of being direct with me. I very much appreciate when people tell me that I'm being inaccurate. I think that satisfying people's curiosity is the most important thing I could be doing (and usually enjoy doing it). I am normally terse in my writing and would appreciate requests to verbosify.
I appreciate it if my grammar or style is corrected (in any of the languages I use here).
Joined Nov 2020
