Remote
_Έλλεν Εμίλια Ά. Arson ζ.
@fogti@chaos.social
mastodon 4.6.9languages, programming, type theory, math, control theory.
neurodiverse (AuSHD?), queer & trans, fem-leaning abinary / demiwoman 
Friend of the 69420*arson ppl.; anti-authoritarian, anti-Copyright, anti-LLMs,, hellenophile / anarchist 
anti-nationalist philhellenism. part of the scientific precariat.
punch Nazis. punch Child abusers. punch Freeze peach extremists. Authoritarians fuck off.
248 Followers
381 Following
50 Posts
Joined December 25, 2022
Pronouns:
En: they/them; she/her; Deu: sie/ihr; es/{ihr,deren}; Ελλ: το (αυτό), η
Git; semi-home page:
fogti (GitHub, Codeberg); https://fogti.codeberg.page/public-docs/
Languages:
Deutsch, English, λίγα Ελληνικά
Replying to
@unkx@icosahedron.website @cstross@wandering.shop @DJGummikuh@mastodon.social @david_chisnall@infosec.exchange It would be very funny to effectively have their opponents donate to their lawyers for making their lives probably easier in the future.
Boosted by @GroupNebula563@mastodon.social
LLM-derived information is the memetic/social equivalent of microplastics.
Replying to
@Mara@hachyderm.io This feels close to "we're shutting down the Internet, it was a nice time".
Replying to
@gsuberland@chaos.social @jackemled@furry.engineer @viq@social.hackerspace.pl @zl2tod@mastodon.online @hailey@hails.org Besides that, I agree, except for security fixes, I doubt anyone really expects new features beyond those already established, in rsync.
Replying to
@gsuberland@chaos.social @jackemled@furry.engineer @viq@social.hackerspace.pl @zl2tod@mastodon.online @hailey@hails.org It is a concern as afaik openrsync lacks a lot —in terms of features— behind samba's rsync.
Replying to
@ireneista@adhd.irenes.space @lcamtuf@infosec.exchange Maybe it's related to the tipping points in "groups of humans" sizes, where there's a point where communication becomes less efficient and a probably also a bottleneck.
Replying to
like, the most extreme form of this I could imagine would be "you need to obtain a certificate/license of successful training in Internet and perhaps pay a regular fee to be able to access most/all of the internet, otherwise you only get a very narrowed-down version (even if you could reach the servers, the servers would also be forced by law by default to deny you access w/o the cert".
That's not technically impossible to implement.
Replying to
People are joking about the "Internet access license" (like: driving license) in Germany, it's a common meme for decades. It could become an enforced reality very quickly.
Replying to
What I'm actually wanting to say:
Now is the time for overlay networks and a total splintering of the Internet as we know it.
Net neutrality discussions of the past will feel tame and mild compared to what's about to come.
Replying to
(I hope this version doesn't become a policy proposal as a "compromise"... it wouldn't even technically be a bad compromise, it would just actually imo be too easy to enforce, and shift the overton window massively)
Replying to
@ireneista@adhd.irenes.space @lcamtuf@infosec.exchange I would also assume that there is simply a tipping point in regards to the scale of a project, where a rewrite of the whole thing at once just doesn't make sense anymore.
For very mall things, it obvious makes sense, because it's easy to hold the entire architecture and important fixes in a single persons head (and thus also documenting them would be doable), but there comes a point where even a large team can't do that in a cooperative manner anymore.
Replying to
Also consider, that it wouldn't even be less secure than doing it in the end-user device, given that people can just hand an end-user device to another person, too.
Replying to
Sure, it would massively restrict access for most people. But what if you just would have to buy a device in a store to get access to a VPN to view "adult content", and you need to show an ID to buy it. (similarly, make SIM cards age-gated; they often already require ID).
It's easy (both in implementation and policy, and easily enforceable after a grace period), devastating in impact, and most critique of age-gating or whatever curr. considered wouldn't apply.
Replying to
In fact, it would be way easier to implement than the "ban social media for kids+teenagers" stuff considered currently.
You just need to either get buy-in or force social media providers, VPN providers and ISPs to implement that.
Replying to
Replying to
@zkat keep in mind that they probably wouldn't be able to assert copyright over most of it anyways because it was dog-fooded so much (i.e. was written mostly by itself by now, which explains its horrible performance)
Replying to
Replying to
@davidbcohen @fullfathomfive @FreakyFwoof It is also an area where people who just came into it from a coding angle have not enough experience in how to get stuff right, and getting experience "naturally" is very labor-intense, and there aren't as much good teaching resources about it as there are about "just writing software".
Replying to
@condret@shitposter.world @theorytoe@ak.kyaruc.moe @hllizi@hespere.de > it is rather easy to write memory safe C
that's debatable. it is way harder to write memory safe C than write memory safe code in most other languages.
But I agree, it isn't what breaks the language.
if UB doesn't fuck you over.
That large parts of the language are UB, and that compiler authors are implementing that in the sense of "the optimizer is allowed to assume that never happens,and if it does, anything is possible" instead of somthing impl-def.breaks it
Replying to
@ireneista@adhd.irenes.space @lcamtuf@infosec.exchange I think in this case it would've been possible to properly rewrite hem one by one, but the metric appears to be more a quantity-over-quality "lets get this out of the door quickly", i.e. the incentives weren't placed right.
Replying to
@ireneista@adhd.irenes.space This problem probably could've been avoided by involving the original authors as much as possible (which afaik isn't what happened, but I also haven't verified this claim deep enough to be certain).
Replying to
@marmarta@chaos.social It is possible, but the easiest way afaik is to build a wrapper in C++ around the Qt parts, which exports a C interface that can be used from Rust. (stuff like the cxx crate (unfortunately by dtolnay) make some stuff easier)
The Vague is calling
Replying to
@jonny *insert "infinite stairs are unacceptable!" meme here*
https://www.youtube.com/watch?v=BWBf2ZKY2xc
Lemongrab Infinite stairs - The Mountain 9
Replying to
@2something@transfem.social @Lana@beige.party association fallacy / guilt by association
is perhaps more fitting.
Replying to
actually, if the weird bans itself can't be prevented, this approach would be imo one of the best possible implementations.
Maybe I'm a bit biased. Germans love their overlay networks. Just look at Gematik, DN42, etc. (this is perhaps the most hilarious list I've ever made)
Replying to
@virtuafox One could additionally, for the "internet access VPN device" mentioned in my thread, implement it such that the license is "plugged in" by being saved on a smart card. Then you can even just distribute the device very cheaply, and make "obtaining the necessary smart card" the hard part.
Replying to
@chrisstoecker@mastodon.social "eine" hört sich nicht nach einer guten Idee an, "so viele wie praktisch möglich und nötig" wäre besser.
Replying to
@natty@astolfo.social @cwebber@social.coop Java2K (wait, that's more like stochastic interpreter)
Replying to
@foone@digipres.club @domi@donotsta.re I read a book about this when I was like ≈15, so I also knew about this.
Replying to
@jdw this feels very connected to independence systems (and perhaps matroids, but not necessarily)
where can I find good books about homology & cohomology theories?
the hackening
how does one translate the term "cohomology" (from mathematics / algebraic topology) into Greek? That "homology" corresponds to "ομολογία" is clear, but idk how to arrive at the "dual" term...
Replying to
also, the (technically optional) introduction of random chance into LLM outputs (including intermediates, etc.) can be seen as intentionally addictive design. It might be a good idea to outlaw that practice, or at least force it to be opt-in.
ctx=https://stunlaw.blogspot.com/2025/06/prompt-anxiety.html
Replying to
@julesh@mathstodon.xyz @mc@mathstodon.xyz Yeah, the application of throwing it on time series data are quite obvious.
Although stocks data is problematic due to having the unit root property ( https://en.wikipedia.org/wiki/Unit_root ) after one has thrown some normalization on the data to get rid of the inflation trend and recurring trends like quartal periodic stuff (e.g. earning reports).
Replying to
@tante@tldr.nettime.org I believe one shouldn't ground one's analysis just in such a single question tho, because due to being basically an utilitarian formulation, it's vulnerable to its classic pitfalls like the https://en.wikipedia.org/wiki/Utility_monster