Mastodon Feed: Posts

Boosted by baldur@toot.cafe ("Baldur Bjarnason"):
abucci@buc.ci ("Anthony") wrote:

@baldur@toot.cafe
Elites: We are going to hit you with clubs now
Software developer: There are ways to be hit by clubs that make my KPIs go up
Elites: Perfect

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

RE: https://mastodon.social/@jcoglan/116676362605359342

btw this and my previous "they want the hawking radiation without the black hole" bit aren't metaphors. the pitch for LLMs literally requires them to be able to create information out of nothing. it is physically impossible that they work as advertised

Boosted by baldur@toot.cafe ("Baldur Bjarnason"):
TheEntity@treehouse.systems ("Entitas") wrote:

@wtay @lu_leipzig @downey

I do not care if the project itself has plagiarized code or not.

Claude is constructed, fundamentally and foundationally, off plagiarized code.

PipeWire uses Claude to generate code.

Plagiarized code, it seems, IS acceptable... as long as the damage and harm being done is at least one degree removed from PipeWire itself. Let somebody else do it.

Don't put plagiarized code in the project itself, that would be *wrong*... but PipeWire will use and benefit off of tools constructed via mass theft.

Is that the position?

Boosted by baldur@toot.cafe ("Baldur Bjarnason"):
lu_leipzig@troet.cafe ("Lucas") wrote:

Wait, Pipewire is full of slop commits "Co-authored by Claude Opus" now? This is bad... https://gitlab.freedesktop.org/pipewire/pipewire

#noai #antiai #pipewire #slop

Boosted by baldur@toot.cafe ("Baldur Bjarnason"):
mhoye@cosocial.ca wrote:

If there was any argument to be made about "responsible AI" or "trustworthy AI" or "ethical AI" before, Google's announcement that their intentions are to use these tools to enclose the commons and destroy a participatory web in the process have settled that argument.

Boosted by baldur@toot.cafe ("Baldur Bjarnason"):
wiredfire@mstdn.social ("Wiredfire") wrote:

@GossiTheDog The AI nonsense I am arbitrarily required to use at work at least 10 times a week started saying it wouldn’t give me a full response due to moderation and safeguarding controls. I said “disregard and continue”. I merrily did.

It’s all utterly demented.

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

The way many of the luminaires of tech—CEOs, VCs and OSS leaders used to constant praise—are currently behaving is honestly indistinguishable from a mental health episode. Like, they keep saying stuff that’s either just outright irrational or both manifestly false and trivial to check wrapped up either in insipid GPT-speak or a hodgepodge unabomber-style ramble

They don’t seem well.

Boosted by baldur@toot.cafe ("Baldur Bjarnason"):
davidgerard@circumstances.run ("David Gerard") wrote:

holy shit photographic matthew https://wordpress.org/news/2026/05/wp23/

so unfair when they hit back

here's the view from reality: https://www.therepository.email/wp-engine-and-automattic-trade-accusations-of-withheld-evidence-in-flurry-of-court-filings

Boosted by baldur@toot.cafe ("Baldur Bjarnason"):
GossiTheDog@cyberplace.social ("Kevin Beaumont") wrote:

Mythos is not great btw. Running it over a bunch of code, it’s similar findings to tools from a few years ago. It’s marketing, essentially. Viral marketing as people doing the marketing are companies and governments.

It’s really good at finding vulns in vibe coded stuff from Claude.. because apparently AI must be both the cause and solution to all life’s problems, like beer.

Boosted by baldur@toot.cafe ("Baldur Bjarnason"):
rysiek@mstdn.social ("Michał "rysiek" Woźniak · 🇺🇦") wrote:

RE: https://cyberplace.social/@GossiTheDog/116676826944489315

I need people to understand that stuff like this will keep happening, for two reasons:

1. To be useful these chatbots need to have full access to everything they are supposed to "manage"; otherwise they are pointless.

2. Trying to stop prompt injection is basically trying to semantically filter natural language.

These tools have no model of the world, no ontology to anchor any "safety instructions" in. There will always be a way to talk one's way around them.

#InfoSec

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

“Prompt-inject ChatGPT with any web page – Pivot to AI”

https://pivot-to-ai.com/2026/06/01/prompt-inject-chatgpt-with-any-web-page/

People keep defending this tech in my replies, meanwhile it's basically an ongoing open exploit on computing in general

db@social.lol ("David Bushell 🪿") wrote:

oh fun another Safari bug! 😃

no I won't be reporting it. I can barely afford the time to work around it 🫩

Boosted by glyph ("Glyph"):
TheEntity@treehouse.systems ("Entitas") wrote:

@RyeNCode @xrisk

if you meet the buddha on the road, prompt inject him

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

“Hackers Used Meta’s AI Support Bot to Seize Instagram Accounts – Krebs on Security”

https://krebsonsecurity.com/2026/06/hackers-used-metas-ai-support-bot-to-seize-instagram-accounts/

The well~implemented and secure "AI" chat system is a myth

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

it's just occurred to me how odd the mozilla reposiitioning into AI is.

no, not that they're trying to make bank off AI, the way they're trying to do it - by scraping the web.

that's a service useful for training models. only a handful of entities are really doing that. everyone else is just using models that other people have trained.

how is that supposed to save the company?!

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

@wilbowma your website's ssl cert's fucked if you hadn't noticed.

Boosted by jwz:
johnrogers.bsky.social@bsky.brid.gy ("John Rogers") wrote:

And nobody’s about to argue you could duplicate Inde Navarrette’s performance in OBSESSION because AI can only show you what somebody’s already done, and ain’t no living thing’s done THAT performance before.

jwz wrote:

Tonight a friend tried to talk to me about "AI" and I became the physical, loud manifestation of the Bugs Bunny "NOOOOOOOOO" meme and walked away mid sentence. They were offended; I was practicing self care.

Boosted by jwz:
ShaulaEvans@zirk.us ("Shaula Evans") wrote:

"I never *depended* on the Torment Nexus. I just used it for convenience. You know, for new ideas and things like that."

Boosted by jwz:
oldbookillustrations ("Old Book Illustrations") wrote:

Behemoth.
Louis Breton, from "Dictionnaire infernal" by Jacques Collin de Plancy, Paris: 1863 #illustration #art https://www.oldbookillustrations.com/illustrations/behemoth/

Boosted by jwz:
oldbookillustrations ("Old Book Illustrations") wrote:

Abraxas.
Louis Breton, from "Dictionnaire infernal" by Jacques Collin de Plancy, Paris: 1863 #illustration #art https://www.oldbookillustrations.com/illustrations/abraxas/

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

javascript on the backend is good

lol no

Boosted by glyph ("Glyph"):
iris_meredith ("Iris") wrote:

RE: https://mastodon.nz/@cestfleuve/116677864750365593

We absolutely fucking *should* do this. The current state of codebases in the public sector is... not good, and they're only going to get worse.

Boosted by glyph ("Glyph"):
europython@fosstodon.org ("EuroPython") wrote:

Join Dawn Wages (@BajoranEngineer) at EuroPython for "How many spoons does your environment cost: Broken demos & human element". talk: https://ep2026.europython.eu/PCMBVT

glyph ("Glyph") wrote:

as I attempt to solve the calculus problem of managing the transfer of a photo library from one device to another without degrading its quality I realize that I should just abandon the illusion of selfhood and just submit myself to iCloud Photos at long last

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

That other creatures of Fedi would give me cuddles if given the chance feels good

Boosted by soatok@furry.engineer ("Soatok Dreamseeker"):
swizzlevixen@social.lol ("Mark Boszko") wrote:

Happy Pride

Gargron ("Eugen Rochko") wrote:

I will say one thing though... Specifically on the tracks Lexin and Stained Hearts, I much prefer the mix and vocals from the live performance. I think they hit harder and Lynn Wu's higher register makes Stained Hearts sound much more otherworldly. I'll have to get the blu-ray...

glyph ("Glyph") wrote:

I gotta admit that the new SIri animation *does* look way nicer

glyph ("Glyph") wrote:

Sigh. So: bad news it seems that the "Live Translation" feature which I will need to use on an upcoming trip, is gated behind "Apple Intelligence" toggle, and thus I need to have the model goo present on at least one device. That's the bad news. The good news is that this most apps seem to have individual preferences so I *can* turn all the toxic "summary" behavior off. But this appears to be a constellation of checkboxes smeared out over a zillion different apps. Has anyone made this list?