Mastodon Feed: Posts

Boosted by jonny@neuromatch.social ("jonny (nonvenomous)"):
thetransmitter ("The Transmitter") wrote:

To catch illicit AI use in peer review, the organizers of the NeurIPS conference have deliberately concealed instructions for large language models in papers sent out for peer review—a practice that some reviewers say erodes trust.

By Dalmeet Singh Chawla

https://www.thetransmitter.org/publishing/scientists-decry-conferences-use-of-hidden-prompts-to-snare-ai-peer-reviews/?utm%5Fsource=mastodon&utm%5Fmedium=org-social&utm%5Fcampaign=20260701-news-publishing-scientists-decry-conferences-use-hidden-prompts-to-snare-ai-peer-reviews

Boosted by soatok@furry.engineer ("Soatok Dreamseeker"):
IkodoMoonstrife@meow.social ("Ikodo 🔜 Furality!") wrote:

Come with me, and you'll be, in a world of pure imagination #furry #VRChat #FenFenFriday

Attachments:

• gifv: An animated GIF of a winged Fennix Ikodo singing and beckoning you to him.

Boosted by soatok@furry.engineer ("Soatok Dreamseeker"):
q@glauca.space ("210561 ✨ [Q]") wrote:

@jkmcnk @christopherkunz @soatok that’s an annoyingly good summary, yes

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

And the code is already 4x faster! So our graph processing lib will have p much negligible overhead in the worst case, and allow slick graph optimized execution in the normal case. I think this thing might actually be good and useful

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

Another reason not to use the pervert glasses

https://www.bbc.com/news/articles/c3wy317d71jo

fribbledom ("muesli") wrote:

Physical media survived scratched CDs, red rings, and Blu-ray lasers, but It might not survive Sony's corporate strategy.

Looks like the PS5 will be my last ever PlayStation. Call me old-fashioned, but I like owning the games I buy.

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

After a couple years of waiting for a project where it makes sense, I'm ~ upskilling ~ by learning to drop my python code down into rust in a non-crappy way by properly learning the language idioms and I gotta say "why didn't I do this before now." It was hard when I was writing old-style python and most of the bad patterns there like lots of indeterminate types and hidden coercions and whatnot, but since ive changed to new-style typed python pretty much everything translates. The borrow checker isn't even hard with a few hundred guess/checks and having compiler feedback in the IDE. Unwrapping stuff and matching stuff also only took a few examples. I love the Into/From coercion stuff.

The python/rust boundary is still something I don't know the optimal way to work thru, but I picked this project to learn because the interface I need to make is the size of a pinhole, with only one well-structured type going in and out.

I still am baby obviously and like all programming languages it will take years to get the subtleties and intuition built up, but ive only been seriously working on this for a week and I am already p much up and going. Trying to do this before I need to look for work again so I can demonstrate I can do the thing everyone wants, high performantz code wrapped in nice and friendly python, and also all the stuff I want to do writing low level raw dog p2p networking routines will require it bc python concurrency can't really do it.

Try 2 stay alive and have a job, and also looking forward to finally doing some fucked up p2p work that doesn't rely on riding on the back of other projects that are all mysteriously dead or cranky.

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

@david_chisnall I mean... they made a staggeringly bad call on EternalBlue, which led to WannaCry, so the third option seems plausible. However, I think reviewing code and missing a bug in an obscure feature and missing that it's enabled by default is also not implausible.

Boosted by soatok@furry.engineer ("Soatok Dreamseeker"):
gsuberland@chaos.social ("Graham Sutherland / Polynomial") wrote:

@soatok I particularly enjoyed @q 's post in that thread. perfectly put.

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

i know, i'm starting to sound like a stuck record, but since i've gotten hooked on air accident documentaries, i've learned what actually trying to do a good job looks like outside of the software industry and damnit if i don't see things we could do better.

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

of course they expect a bit more than phoning it in if you want to put a plane in the air.

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

my god, i'm reading phrases like "normalization of deviation" and "alarm fatigue". it's like somebody's discovered 'human factors', as practiced in aviation for decades, apply to software too.

Boosted by dysfun@treehouse.systems ("gaytabase"):
glyph ("Glyph") wrote:

@ecn @davidgerard this is logically true of course and I have been saying it for years now but everyone also needs to know that it is now empirically validated https://flipboard.com/@thenextweb/ecosystems-aj8kk6atz/-/a-KvDNdak0T5CTTNZhqtllzg%3Aa%3A47933081-%2F0

Boosted by glyph ("Glyph"):
ecn ("ian (cool)") wrote:

So when somebody tries to defend their use of these tools because a human is still reviewing it remember that they're lying - to you and to themselves.

They know there's problems with these tools, and are trying to convince themselves that somehow it's fine and you should stop calling out for their poor choices.

Boosted by glyph ("Glyph"):
ecn ("ian (cool)") wrote:

I can attest with the utmost confidence that every developer or manager or whoever that claims "all code written by AI is reviewed by a human" is a boldfaced lie.

AT BEST they have quickly skimmed it, but more likely they've not looked at it at all.

Actually reading code THAT YOU DID NOT WRITE YOURSELF takes quite a bit of time, and the one thing that people who use AI tooling care most about is just how fast it makes things.

They will not spend the time needed to actually read it.

Boosted by jonny@neuromatch.social ("jonny (nonvenomous)"):
niksirbi@neuromatch.social ("Niko Sirmpilatze") wrote:

RE: https://mastodon.online/@neuroinformatics/116851125455601659

I'll be in Barcelona for #FENS2026 next week, presenting a poster about https://movement.neuroinformatics.dev/, and also hanging around our team's other posters (details in quoted post).

Drop by if you want to chat about any of these projects or just to say hi!

Boosted by jonny@neuromatch.social ("jonny (nonvenomous)"):
elduvelle@neuromatch.social ("El Duvelle") wrote:

What's better for a researcher working to publish a project but also applying for fellowships and with their current contract ending in 1 year?

1. Make a paper with a focused claim that might not go in "high impact" journal because it's too focused (and perhaps work on other findings from the same data later) but would be ready faster
2. Make a paper with several claims which might have a bigger chance of "high impact" but will take several additional months to be finished?

#Neuroscience #Research #AcademicChatter

Boosted by glyph ("Glyph"):
atax1a@infosec.exchange ("mx alex tax1a - 2020 (6)") wrote:

weird how all of this "ai assisted" dogshit keeps having wild-ass basic-bitch reliability and security bugs

Boosted by soatok@furry.engineer ("Soatok Dreamseeker"):
markstos@urbanists.social ("Mark Stosberg") wrote:

I'm a Google Workspace admin, and a team member sent me a screenshot showing that a message they wanted to send was being blocked because it violated an organizational policy-- which was news to me-- I hadn't put in place to block messages based on their content.

Turns out Google Workspace now warns by default when the message content mentions gender identity.

#google #privacy #politics

Boosted by soatok@furry.engineer ("Soatok Dreamseeker"):
silver-sebastian.bsky.social@bsky.brid.gy ("Sebastian Silverfox") wrote:

Keeping it realistic at the realistic fursuiting panel. #fursuit #anthrocon
📸: @crucifox.bsky.social@bsky.brid.gy

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

Computers were nice; I miss them

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

no o o o o o

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

Just finished setting up Immich (Pic unrelated)
#fursuitfriday

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

Sometimes it can be very windy outside, even in summer. That's why I always wear a scarf so I'm prepared and protected :blobfoxcomfy:

📍: CSD Augsburg 2026
📸: @kianga

#Fursuit #Furry #FursuitFriday

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

Nah, let him cook (literally)

RE: https://www.threads.com/@markshrime/post/DaVc-tjCg5V

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

She wants to be your friend.

https://freethoughtblogs.com/pharyngula/2026/07/03/goldenrod-crab-spiders-want-to-love-you/

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

you won't survive software development of the 80s

my dude, i'm not surviving software development of the 20s

Boosted by soatok@furry.engineer ("Soatok Dreamseeker"):
Em0nM4stodon@infosec.exchange ("Em :official_verified:") wrote:

I will never understand why so many activists still use generative images for their content, and rely on big tech for communications and organizations when there are now so many practical, more secure, and much more ethical alternatives.

Your principles aren't just to be stated, they must also be embodied and integrated in your work.

We must show by example what can be done with ethical tech and ethical practices.

#Activism #NoAI #NoBigTech

fromjason ("fromjason.xyz ❤️ 💻 ✍️ 🥐 🇵🇷") wrote:

A Puerto Rican, an African, and Jewish person walk into a bar. 😍

Boosted by dysfun@treehouse.systems ("gaytabase"):
yourfutureex@metalhead.club ("Your Future Ex") wrote: