Mastodon Feed: Posts

Boosted by soatok@furry.engineer ("Soatok Dreamseeker"):
emilvolk@tiggi.es ("EV, your Witch Wolf ΘΔ🧙🪄🐺🔞") wrote:

@soatok more seriously: I’ve been consistently surprised by how many people are reasonable when you communicate with them one-to-one.

And this even goes for technical advice. At least here, the attitude of ‘RTFM n00b’ isn’t really sufficient anymore (and with the way the modern internet information is, not really that helpful either).

I believe it’ll ultimately be the thing that helps steer people away from LLMs (and LLM-dependency!), assuming they haven’t managed to create an atomized toxic hellscape of a society before then…

(In other words, thank you! 💜)

Boosted by isagalaev ("Ivan Sagalaev :flag_wbw:"):
janamarie@mystical.garden ("Jana") wrote:

went on another walk with my IR camera from my previous thread. Even managed to catch a few rays of sunlight in between the otherwise grey week.

It continuous to be a lot of fun and I slowly get the hang of focusing correctly! There are some image artifacts and out-of-focus regions at times, but that is okay with this 2011 camera.. :D

The resulting photos also just keep blowing my mind, this is sooo good! Also there are some interesting observations, the concrete in the third photo does not reflect much, if any IR, while the foliage does. This is also something you can feel in the summer, while streets are incredibly hot, trees and grass usually are pretty chill as they do not absorb a lot of IR!

#photography #infrared

Boosted by soatok@furry.engineer ("Soatok Dreamseeker"):
ana@starlite.rodeo ("ana «model a7m2» :neobot_cat_headphones_pink:🎶") wrote:

why did they call them "ai datacenters" when they could have called them "slopping malls"

soatok@furry.engineer ("Soatok Dreamseeker") wrote:

I'm going to answer an obvious question that might not immediately materialize in everyone's minds:

Well if social status doesn't matter, what does?

Connection.

Community cannot exist without connection.

Learning can happen without it, but only from experience. Even an extremely weak tie (e.g., reading something that someone else wrote) is still a connection. After all, how many people have favorite authors?

Connection does not imply hierarchies. A decentralized mesh of nodes might form a graph that can be studied, but there is no axis to measure quality on.

Sure, some connections might be closer than others.

Yeah, some people might have more or fewer than others.

But those connections, and the people you share them with? That's what you should care about.

The gay furrry with a blog that talks about applied cryptography is but one among ~8 billion.

Boosted by soatok@furry.engineer ("Soatok Dreamseeker"):
0xabad1dea@infosec.exchange ("abadidea") wrote:

RE: https://hachyderm.io/@ekuber/116585853572401757

One of the virtues of big open source projects is that little happens behind closed doors. However, since 99% of what they discuss just isn’t interesting to downstream end users, they forget that downstream end users *can* read this stuff and *can* post very angry comments on a discussion that they feel invited and entitled to, *as users* who are perceiving their role here as *delivering accountability.*

Github PRs are a nightmare edge case for attracting unplanned attention, because even most people who got past the first paragraph of the post about policy everyone was mad about never actually saw *the policy itself* due to github’s unhinged UI. Mastodon’s overall success rate for walking away with good, clear information about the policy itself was… low. Very low.

On the one hand, I think that anticipating the nature of the issue would attract public attention, and putting effort into framing it in a way more ready for public non-contributor consumption, would have reduced the misunderstandings and blowup once it escaped containment. On the other… nothing can fix the fact that a rando who’s been linked to a comment on a github PR in an active project is going to be lost and confused.

So, contributors: “How does this come across to someone who’s not intimately familiar with our process and involved in our ongoing discussions?” Might save you a lot of notifications one day.

And, non-contributors: that big project that definitely existed more than five seconds before you found a hot link to a hot post? Spend more than five seconds figuring out what’s going on before you deliver accountability. And if you don’t have the time or energy, that’s fine! But then don’t push the angry comment missile strike button for something you’re not willing to follow up on.

soatok@furry.engineer ("Soatok Dreamseeker") wrote:

Let me put it another way:

I hate hierarchies. Social status is a hierarchy.

(Social status insecurity is also a big part of the right-wing pipeline that drives people towards fascist violence.)

We don't need hierarchies!

If you admire anything about me or anything I've done, just use that as evidence to tell yourself, "If Soatok can do it, there's no reason that I can't also succeed" when you're pursuing your own passions.

Do not ever tell yourself that anyone else is better than anyone else. That's so narrow-minded.

soatok@furry.engineer ("Soatok Dreamseeker") wrote:

The funniest thing about the Internet (but mostly the Fediverse these days, in my experience) is how often you'll cross paths with someone who's done something meaningful and impressive and could cash in their nerd cred to pwn you in an argument but instead they're just like, "I have a funny story about that," and then omit the part where they did crazy/cool shit.

soatok@furry.engineer ("Soatok Dreamseeker") wrote:

I'm not going to win a Turing Award or Levchin prize for any of this, because it's not really groundbreaking or innovative in any meaningful way.

But it is useful.

soatok@furry.engineer ("Soatok Dreamseeker") wrote:

To put things in perspective: The most interesting thing I've done in the past few years is boring, unglamorous infrastructure design that is only fully comprehensible to someone with deep cryptography expertise, but comes in a simple enough package to vaguely understand and hopefully work with: https://publickey.directory

This isn't the kind of thing that's going to get 10 million views on TikTok tomorrow.

At best, when the Fediverse goes all in on end-to-end encryption, it'll solve one of the hard problems that most people who decide to bolt encryption onto their messaging app inevitably punt because of its difficulty.

But the thing that most people will actually give a shit about is "zomg mastodon has end-to-end encryption now!!!" and maybe have one less reason to avoid it.

And the Fedi folks can still ship E2EE without key transparency.

It would be a terrible idea, but not one without precedent or forthcoming disaster.

soatok@furry.engineer ("Soatok Dreamseeker") wrote:

It's really funny to get messages from strangers who sound like they're nervous like they're messaging some big shot celebrity or whatever.

Like, yeah, I get it. Anxiety's a bitch. But I'm just a guy with weird hobbies; a nerd who likes to make sense of the world and share his notes.

But when it comes to social status, I'm also kind of a loser?

I'm definitely too lame to get a boyfriend, and the local queer/furry scene treats me like a ghost whenever I try to organize an activity (even if I'm offering to pay for everyone's way).

And unless you're very interested in the same technical topics I write about frequently, well, my fursona is a forgettable blue canid, like the tends of thousands other blue canids. I'm basically background noise.

So it gives me a lot of proverbial whiplash to be treated that way >.>

glyph ("Glyph") wrote:

no gods no kings no masters no LLM contributions please

glyph ("Glyph") wrote:

Anyway, when I visited the Codspeed booth today and mentioned an issue with intermittent variance in our benchmark results, the CEO was present and immediately walked me through the history of Twisted's build results, explained why they might look the way they did, and also gave me a personal update from the engineer working on the occasional variance issue. Kind of an awe-inspiring level of support.

glyph ("Glyph") wrote:

(This was back in March and there has been no intrusion of further junk into our otherwise excellent codspeed experience, which clears a bar for user respect that many other service providers are crashing into right now.)

glyph ("Glyph") wrote:

Honorable mention to https://codspeed.io who, while they do offer some "agentic" stuff, they:

1. also have a robust offering without any AI, and
2. when the "AI" button appeared one day and freaked me out a little bit, I sent an email, and within hours the CEO replied "I understand your concern, and it makes sense to be able to disable all AI features on CodSpeed.". Less than a day later, they followed up with "You can now disable [AI] in the Capabilities tab of your organization settings"

Boosted by cstanhope@social.coop ("Your weary 'net denizen"):
SabrosiaVit@epicure.social ("Glen L Creasy") wrote:

Found this cute wee frog when I was looking for fossils along a creek near our village - so tiny!!

#Nature #Frogs #TinyAnimals

aredridel@kolektiva.social ("Mx. Aria Stewart") wrote:

All y'all have the average ethical sense of “we should build more housing" while opposing development of housing, y'know that, right? Like, that's the baseline quality of ethical position.

Boosted by glyph ("Glyph"):
ubernostrum@infosec.exchange ("James Bennett") wrote:

My approach to Python packaging infrastructure is mostly "look up whatever @hynek is doing and copy that", and I'm glad to have that vindicated by Andrew Nesbitt's #PyConUS GitHub Actions talk recommending Hynek's 'stamina' package as one with a setup worth studying and learning from.

Everyone, do as Hynek says!

(also, go sponsor him so he can keep doing good things)

Boosted by glyph ("Glyph"):
jalefkowit@vmst.io ("Jason Lefkowitz") wrote:

ME [ killing you and taking all your stuff ]: Ethical issues aside, this is gonna be a big win for me

Boosted by cwebber@social.coop ("Christine Lemmer-Webber"):
kevingranade@mastodon.gamedev.place ("Kevin Granade") wrote:

@richpuchalsky @cwebber I just checked and the pb and pbj sandwich recipes are autolearn at cooking skill 0.

Boosted by kevinevans@hachyderm.io ("Kevin"):
freya@highenergymagic.net ("Ra (Freyja) (it/its)𒀭𒈹𒍠𒊩") wrote:

hey so my lovely pretty amazing daughter @kebokyo wants to connect with #Seattle #Cascadia trans folks! anyone out there want to drop in and start up a conversation? it's got questions, yall got answers I'm pretty sure. if \you're not from there, please boost!

pzmyers@freethought.online ("pzmyers 🕷") wrote:

Who is more smug? Me or Andrew Klavan?

https://freethoughtblogs.com/pharyngula/2026/05/16/am-i-smug/

Boosted by cwebber@social.coop ("Christine Lemmer-Webber"):
richpuchalsky ("Rich Puchalsky ⩜⃝") wrote:

@cwebber

I will never play CDDA because the tutorial showed how I needed a recipe in order to put peanut butter in between two saltines

cwebber@social.coop ("Christine Lemmer-Webber") wrote:

Normal person: "This is absurd, I just want to go fight some zombies. I am going to go outside and start fighting them."

CDDA fan: "I wouldn't do that if I were you"

Normal person: "That's the whole point of the game! I am going to go do it"

*one minute passes*

Normal person: "WTF I died immediately"

CDDA fan: "Well what did you expect to happen"

cwebber@social.coop ("Christine Lemmer-Webber") wrote:

Cataclysm Dark Days Ahead Fan: "Okay so this is my favorite game, it's about fighting zombies, do you want to watch"

Normal person: "Sure"

CDDA fan: *3 hour long lecture about how to optimize putting items in each of the pockets of your cargo pants, how to stay hydrated without getting sick, how to get comfortable enough to get a good night's sleep, some tips and tricks of how not to get scurvy or depression*

Normal person: "Wow that was quite a tutorial. Okay, now let's start having some fun"

CDDA fan: *blinking* "But we are having fun"

Boosted by adele@social.pollux.casa ("Adële 🐁!"):
moon@m.g3l.org ("Moon") wrote:

New Moon (0%)

brib@bribstodon.xyz ("brib :neofox_floof:​ :Nonbinary:") wrote:

11.3 million lines in Gram. Eeeek

brib@bribstodon.xyz ("brib :neofox_floof:​ :Nonbinary:") wrote:

I run a script to count the lines of code in the last non-slop version of Zulip, it runs for a couple of minutes, then spits out 1.6 million lines. Eek.

I run a script to count the lines of code in Gram. It has run for 10 minutes. It is still running.

@krig I admire your dedication in maintaining this... what is your secret?

Boosted by slightlyoff@toot.cafe ("Alex Russell"):
adamshostack@infosec.exchange ("Adam Shostack :donor: :rebelverified:") wrote:

https://kevinpatel.xyz/posts/no-way-to-prevent-this/ 🔥

slightlyoff@toot.cafe ("Alex Russell") wrote:

Ukraine is truly getting into the Eurovision spirit:

https://newsie.social/@Tendar/116586451513435369

Boosted by slightlyoff@toot.cafe ("Alex Russell"):
danderson@hachyderm.io ("Dave Anderson") wrote:

Lilypond is a remarkable piece of software.

Music notation typesetting ("engraving" is apparently the term of art) is a remarkably deep and complex problem, especially if you're trying to represent all the different notation systems and subtleties that people have come up with, spanning a millennium or more.

Lilypond's documentation has an essay on why it's such a hard problem, and I highly recommend reading it: https://lilypond.org/doc/v2.26/Documentation/essay/music-engraving

It's an ode to the art of sheet music engraving, and motivates why lilypond's output is very deliberately not perfectly regular, but instead tries to work with the flow of the music and shape itself to convey it clearly.

The essay also has a passage that captures the aesthetic of one of my favorite classes of software:

> We have written LilyPond as an experiment of how to condense the art of music engraving into a computer program. Thanks to all that hard work, the program can now be used to perform useful tasks.

As I'm forced to deal more and more with software written with an extractive mindset, I find this statement very moving: the starting point was trying to understand and celebrate a beautiful craft. The production of useful output is an incidental side effect of the pursuit of understanding.

I'm sure there's some poetic license there and the authors did sort of set out with the goal of producing useful output. But I also wish that more software endeavors approached their problem domain with this sort of humility and curiosity. Less "we can use computers to fix X", more "can we find a way for computers to do justice to X?"