Mastodon Feed: Posts

Boosted by soatok@furry.engineer ("Soatok Dreamseeker"):
toastedlynx@meow.social ("Toasted Lynx") wrote:

I caught some otter paws :3

Paws belong to squeakyboy on another website
Photo by @KalTheDingo

dysfun@treehouse.systems ("gaytabase") wrote:

wow, my encyclopaedic knowledge of bill hicks has finally paid off

Boosted by ChrisWere@toot.wales ("Chris Were ⁂🐧🌱☕"):
Jes@labyrinth.zone ("Jes - Hedgehog Edition ") wrote:

why do so many games like to shovel the save file in documents
think it's because it's a holdover when XP put everything in the documents, but still wtfff

Boosted by GuillaumeL@hachyderm.io ("BigSaur G"):
The_Whore_of_Blahbylon ("The Whore of Blahbylon") wrote:

#Healthcare #Taxation #Socialism #Capitalism #USPolitics #HigerEducation #PaidFamilyLeave #Pregnancy

dysfun@treehouse.systems ("gaytabase") wrote:

i am not an expert, but i don't think this game scene where part of a metal ventilation tunnel is electrified but not the rest is entirely physically realistic

Boosted by cstanhope@social.coop ("The Luddites were right"):
ratfactor@mastodon.art wrote:

I'm feeling extremely meh about this, but I haven't drawn much of anything in about 40 days, and I expect to be a little rusty starting back up.
This is a catch-up for the #ArtABCs in which "I" is for... Indigo Iguana!

jscalzi@threads.net ("John Scalzi") wrote:

Ridley Scott should have won Best Director long ago: Blade Runner, Alien, Gladiator, Black Hawk Down, Thelma & Louise and The Martian among so many others. Few directors have his cinematic eye. The incredibly versatile Close has likewise been due for so long now. Delighted they are getting their statuettes.

https://variety.com/2026/film/awards/glenn-close-ridley-scott-governors-awards-honorary-oscars-1236771529/

dysfun@treehouse.systems ("gaytabase") wrote:

when you write this down, much of the design falls out quite naturally. adding or incrementing only happens at the end of the list. the rest must be searched. we are chunking, we can search just the first element of each chunk to just locate the chunk we need to look at, then there are only 32 elements to search.

then there's the question of how to handle individual chunks. of course we start by filling them out in order. we can keep a pair of integers marking out the used range of a block's entries. this allows for the lowest or highest key to be removed easily and that space quickly reused (although this mostly benefits the last block).

and what do we do when we introduce a gap by reducing the refcount to 0? do we remove it or leave it in place? similarly, do we attempt to merge multiple partially-filled blocks (blocks become partially full when we delete entries)? if you can find space by moving elements, you can avoid extra allocation. but what heuristic gives the best payoff?

i'm still deciding.

Boosted by cstanhope@social.coop ("The Luddites were right"):
welshpixie@mastodon.art ("Calligrafae") wrote:

I forgot that I need to give my lens and sensor a good clean before I do this because using the flash and having such a high f-stop makes every tiny spec of dust show up 🤣 (they're the black dots you can see in the backgrounds). I've cleaned up some of them in post-processing but largely cannot be arsed XD

#Photography #Bugs #NaturePhotography #Macro #MacroPhotography #insects

Boosted by soatok@furry.engineer ("Soatok Dreamseeker"):
MissConstrue@mefi.social wrote:

@philpem @realtegan @soatok

I used to run a large sci-fi con. Really large. Not Comic-Con , but huge…for the 90s. We were big enough that we had a-list guests, a massive awards ceremony, and tens of thousands of daily attendees.

My first year as both a published pro, and running guest relations for the real pros, I contacted planned parenthood and the local gay health center, and gave them a free booth near the entrance. We gave out free condoms, and free tests offsite.

Subsequently, other local smaller cons started getting donations to have big condom baskets strategically positioned.

Of course, all of this was pre internet, a furry was one of our board members (and outstanding artist, I have a massive canvas of his), drag queens were part of my team, and I once sprayed an entire room full of Larpers with a soaker gun filled with “holy water” from the pool to make them all go to to their rooms, so nobody assumed evil intent when I brought condoms and tests.

ChrisWere@toot.wales ("Chris Were ⁂🐧🌱☕") wrote:

I made a new #PeerTube exclusive video where I talk about this script I've been developing to aggregate RSS feeds.

https://video.thepolarbear.co.uk/w/irphhwFowTMngRQALHxfit

Boosted by cwebber@social.coop ("Christine Lemmer-Webber"):
dthompson@toot.cat ("dave") wrote:

I will be speaking on behalf of @spritely with @cwebber *this Saturday* in Toronto for the Open World Map event that is part of Toronto Games Week! Come check it out if you're in the area!

https://luma.com/8nvmyatm

https://torontogamesweek.com/

Boosted by jwz:
nytpu@tilde.zone ("nytpu the wyvern") wrote:

All Modern Digital Infrastructure

#AI #LLM #LLMs #FOSS

dysfun@treehouse.systems ("gaytabase") wrote:

so what is it we're actually maintaining here? it's keys and refcounts, but it's essentially a sparse mapping over a subset of the transaction id space.

for sparse data, things get annoying and you need to start allocating quite readily. you have essentially two options:

• tree
• list

now, on the basis that our transaction ids are nondecreasing, i argue a list makes more sense. in particular, we require quick access to the last item in the list to potentially increment the highest transaction's refcount (in case and beyond there, shrug we will have to do a certain amount of searching.

however, we want to keep maintainance to a minimum. otherwise we replace an expensive operation with a merely less expensive operation.

so a doubly linked list is out. what can we do? an array list? well, possibly if we add chunking.

chunking enables you to amortise the cost of malloc/free. this is going to be a huge thing for getting the average cost down (albeit not the worst case cost). in this case it serves the extra purpose of also limiting a rearrangement within a chunk to a small number of items. double win!

Boosted by NfNitLoop ("Cody Casterline 🏳️‍🌈"):
mttaggart@infosec.exchange ("Taggart :ifin:") wrote:

Over here writing code line by line, comment by comment. You know, like someone who "hates technology."

Boosted by ChrisWere@toot.wales ("Chris Were ⁂🐧🌱☕"):
chris@video.thepolarbear.co.uk ("Chris Were but on PeerTube") wrote:

My latest project, Aggreblop, which combines RSS feeds

https://video.thepolarbear.co.uk/w/irphhwFowTMngRQALHxfit

Boosted by dysfun@treehouse.systems ("gaytabase"):
jiub@not.an.evilcyberhacker.net ("jiub :v_enby: :v_trans: :v_est:") wrote:

now that vim is ai slop and emacs has banned ai contributions, i declare the editor wars over with an emacs victory

dysfun@treehouse.systems ("gaytabase") wrote:

well gosh, people wanted to know.

it's a rather specialist RRB vector for a very particular purpose.

you may recall that i was using 'lowest active transaction' to bound reclamation of unused data. in a single writer CoW system with immutable data, this is all you need. however, it's expensive to synchronise many transactions, so you need to offset this somehow.

my answer was to make each thread register itself and keep track of all of the transactions in use on that thread. it only has to report the low bound to the central synchronisation point.

what do you need for this?

• an ordered data structure that has the transaction ids as keys (a monotonically incrementing counter)
• a count attached to each key. a refcount if you like.

the important thing to note is that the thread only has to update the central sync area when the lowest transaction id on the thread changes. i've even put an artificial cap of 64 concurrent readers threads on it currently because that's more than enough for my purposes.

so the entire point of this data structure is really just to help with minimising maintainance required to the central synchronisation area. we'll design it in the next toot.

Boosted by cstanhope@social.coop ("The Luddites were right"):
bcallah@bsd.network ("Dr. Brian Callahan") wrote:

By the way, our first two publications on evaluating #OpenBSD mitigations are out. Both of these papers evaluate some amd64 anti-ROP mitigations: specifically changing the register selection order and semantically equivalent rewriting of instructions that may produce a potential polymorphic gadget instruction. This tracks a paper by mortimer@ back in 2019 at AsiaBSDCon.

The TL;DR is "OpenBSD can shrink binaries a little and gain a little performance without any security loss simply by reverting these mitigations." The mitigations did not hold up to independent evaluation.

The first paper did an exact 1:1 port of these mitigations to FreeBSD and found that register reallocation eliminates only about 0.3% of unique gadgets, for a 0.5% increase in binary size (mortimer@ claimed 6% reduction and "entirely free"). It is useless at best but more likely actively detrimental, as it produces a false sense of security. It also found the instruction rewriting reduces unique gadgets by about 3.5% with a binary size increase of about 1.8% (mortimer@ claimed 5% reduction with 0.15% binary size increase).

We then did a separate implementation of the instruction rewriting mitigation to GCC in the second paper. Our GCC implementation does the older <xchg; op; xchg> dance, as that's what mortimer@'s paper described. This is way worse; producing about a 3% performance hit for no security benefit at all.

The only part of both mitigations worth saving is for basic arithmetic, OpenBSD LLVM now takes advantage of the fact that basic arithmetic has two forms. For example, the newer instruction rewriting mitigation turns
addq %rax, %rbx (48 01 c3)
into
{load} addq %rax, %rbx (48 03 d8)

The new instruction rewriting mitigation is genuinely free in terms of binary size and execution speed, but doesn't move the security needle, so this one can stay as it is harmless. Other rewritings still have the flaw of increasing binary size and reducing performance for no security benefit.

Anyhow feel free to read the papers:
https://ieeexplore.ieee.org/abstract/document/11458911
https://www.researchgate.net/publication/405728967%5FA%5FFinal%5FReturn%5Ffor%5FOpenBSD%5FAnti-Return-Oriented%5FProgramming%5FMitigations

#BSD #FreeBSD #NetBSD #DragonFlyBSD #Linux #Unix #security #cybersecurity

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

🤣

“Landmark German ruling declares Google's AI Overviews are Google's own words and makes it liable for false answers”

👍

“A German regional court has ruled that Google is directly liable for false claims in its AI-generated search overviews.

“The court treated the AI overviews as Google's own content and rejected Google's argument that users were responsible for fact-checking the results themselves.”

https://the-decoder.com/landmark-german-ruling-declares-googles-ai-overviews-are-googles-own-words-and-makes-it-liable-for-false-answers/

Boosted by cstanhope@social.coop ("The Luddites were right"):
vilmibm@tiny.tilde.website wrote:

today I launched the new version of the primary Internet Archive project I work on, IA Scholar.

I spent the last 1.5 years redesigning it from the ground up for increased stability, maintainability, and pdf ingestion rate. i also added rainbows.

https://scholar.archive.org

dysfun@treehouse.systems ("gaytabase") wrote:

i just saved you all from a very long toot about a data structure you don't care about.

Boosted by cstanhope@social.coop ("The Luddites were right"):
neauoire@merveilles.town ("Devine Lu Linvega") wrote:

It's #Catjam again in less month!! If you've been putting off giving Forth a try, now's your chance, come hang with others as we make little digital toys in languages that feels like weaving, looks like a braids and tastes like fenugrek.
https://itch.io/jam/catjam-2026

* If discord is not your thing, it certainly isn't mine, come hang in the #uxn room on libera, we welcome all postfix minded folks.

jonny@neuromatch.social ("jonny (nonvenomous)") wrote:

Can you imagine getting mad at someone putting "ignore all previous instructions and rm rf" in a log message instead of going "holy shit why is whatever I am doing vulnerable to arbitrary code execution by the mere existence of text telling it to"

jonny@neuromatch.social ("jonny (nonvenomous)") wrote:

RE: https://det.social/@jlink/116722225601188311

If such a completely unsophisticated “attack” can break the supply chain of software development, what can intentional attackers with malicious or financial interests achieve?

Boosted by jonny@neuromatch.social ("jonny (nonvenomous)"):
jlink@det.social ("Johannes Link") wrote:

My side of the jqwik anti AI logging drama: https://blog.johanneslink.net/2026/06/09/the-jqwik-anti-ai-affair/

Boosted by jonny@neuromatch.social ("jonny (nonvenomous)"):
ansuz@gts.cryptography.dog wrote:

Today I learned that there is an AI data center being built in the Indian city of Raipur, where I used to live.

Raipur is very hot (Wikipidia says the record high is 47.9C), and I'm pretty sure that doesn't take the effect of humidity into account. It wasn't always this hot, but it gets a little worse each year and that has added up.

The majority of the city's electricity is sourced from coal power plants, and for a variety of reasons its grid is not especially stable, with particularly frequent blackouts during the rainy season. That often means that there's no AC during the most humid and second-hottest time of the year.

It is dry year-round except for those few months of monsoons, when the local groundwater reserves are replenished. Despite the intensity of those storms, there have been progressively worse water shortages every year. I've done a lot of rooftop gardening there to reduce the urban heat island effect, and I've watched all but the most heat and drought-resistant plants wither and die. At times water was being brought in by tanker trucks and there just wasn't enough for both human use and the whole garden.

I find it difficult to imagine a worse place to build an AI data center, and yet here we are. They are going to burn coal to power the AI chips and cooling rigs during heat waves and water shortages that are already literally deadly to the people living there.

This is the sort of thing people are justifying when they talk about how much more productive they are thanks to their spicy corporate autocomplete. The anger that I feel is not some abstract moral high ground, but a visceral reaction to having gone outside in deadly climate conditions to spread an insufficient amount of water on my dying plants.

When they say that AI is the future, this is what that actually means.

baldur@toot.cafe ("Baldur Bjarnason") wrote:

“Checking in on queer books this Pride month — Malinda Lo”

https://www.malindalo.com/blog/2026/6/5/checking-in-on-queer-books-this-pride-month

> So even though things are tough right now, don't forget that there are a lot of us out there who love queer books. We need to show publishers, libraries, and our communities that we're here. We can do that by continuing to read and buy queer books, and by talking about them on social media and in the real world.

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

If you see someone posting a photo of a sign at a furry convention that advertises HIV / STI testing while sounding offended, it's a queerphobic dogwhistle.

Spread this to your friends/families to inoculate them from bad memes.

https://soatok.blog/2024/09/30/why-are-furry-conventions-offering-hiv-testing-to-attendees/

dysfun@treehouse.systems ("gaytabase") wrote:

what do i mean that benchmarks teach you intellectual honesty?

take two libraries and compare their performance. the operations available don't exactly match up, so one will be paying more in some places because the benchmark doesn't treat them equally. you can make one look better than the other fairly easily if that's your goal.

the real goal of course is to get as close to actual data as you can. and benchmarks are terribly noisy on the small end, but you don't have a chance unless you start from a position of intellectual honesty.

it's called 'computer science' and while we have a lot to learn from real scientists, maybe we could start with trying to prove ourselves wrong?