Mastodon Feed: Posts

Boosted by baldur@toot.cafe ("Baldur Bjarnason"):
jenniferplusplus@hachyderm.io ("Jenniferplusplus") wrote:

There's one very important thing I would like everyone to try to remember this week, and it is that AI companies are full of shit

Only rarely do their claims actually bear scrutiny, and those are only the mildest of claims they make.

So, anthropic is claiming that their new, secret, unreleased model is hyper competent at finding computer security vulnerabilities and they're *too scared* to release it into the wild.

Except all the AI companies have been making the same hypercompetence claims about literally every avenue of knowledge work for 3+ years, and it's literally never true. So please keep in mind the highly likely possibility that this is mostly or entirely bullshit marketing meant to distract you from the absolute garbage fire that is the code base of the poster child application for "agentically" developed software

You may now resume doom scrolling. Thank you

Boosted by baldur@toot.cafe ("Baldur Bjarnason"):
androcat@toot.cat ("Androcat") wrote:

@jenniferplusplus Literally seconds ago I wrote elsewhere: "first rule of LLMs: If someone from an LLM company says their model can do x, it can't do x, but it includes some thoughts and prayers to please do x."

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

Worth your time:

https://leehanchung.github.io/blogs/2026/04/05/the-ai-great-leap-forward/

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

RE: https://mamot.fr/@thibaultamartin/116368015104267206

This. It's worth asking deeply what lets it happen, given that people _have_ been beaten down and aren't used to using their creativity and ambition to do things.

There is so damn much untapped human potential.

Boosted by aredridel@kolektiva.social ("Mx. Aria Stewart"):
thibaultamartin@mamot.fr ("Thib") wrote:

“What people face too often is being ground down by systems, institutions, and unjust leaders who insist on creating roles where people are forced to do dehumanizing, isolated, meaningless work, while not being given the agency to make smart and empowered decisions about how the work gets done”

“All you have to do is provide a worthy mission and get the fuck out of the way.”

@anildash hits the nail on the head, again.

https://www.anildash.com/2026/04/06/people-love-to-work-hard/

Boosted by slightlyoff@toot.cafe ("Alex Russell"):
css@front-end.social ("CSS by T. Afif :verified:") wrote:

My blob generator is ready to give you border-only blob shapes.

https://css-generators.com/blob/

Update your Chrome and give it a try!

#CSS #HTML

Attachments:

• gifv: A screen record of my blob shape generator

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

using dnschecker.org to see DNS has propagated around the entire world, except my computer - now playing hunt the cache

Boosted by baldur@toot.cafe ("Baldur Bjarnason"):
inthehands@hachyderm.io ("Paul Cantrell") wrote:

RE: https://wandering.shop/@susankayequinn/116374937786561286

I see this with my college students too: while some use it to cheat and a few are fully AI-pilled, at least as many treat it with utter contempt. It’s dividing the up-and-coming generation just like it’s dividing software developers.

Boosted by cstanhope@social.coop ("Your weary 'net denizen"):
kristapsdz@bsd.network ("Mr Taps") wrote:

Does computer #history interest you, or maybe you're just curious where well-known and well-used tools come from? I've just updated the History of #Unix #Manpages, https://manpages.bsd.lv/history.html, with the content you didn't know you wanted til this very moment. Learn about how the "man" program came to be, and just why are manpages styled like that? It includes snippets from Cynthia "Cindy" Livingston, who wrote the manpage language "mdoc"; John Eaton, who wrote the first GPL man tool; Doug McIlroy, who helped to divide manpages into sections; and more. Did you know that serving manpages online was part of one of the original http daemons? Or that an xman existed before X11R6, in X10? Enjoy!

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

It's not AI, it's real

Mastodon wrote:

In a new blog post, @imanijoy provides a deeper dive into the design process for Collections, and explains what we’ve included (and things that are left out) for the first release. We’ll be enabling this feature on mastodon.social next week, and rolling it out more widely in Mastodon 4.6 very soon. We’re excited to hear what you think!

https://blog.joinmastodon.org/2026/04/designing-collections/

Mastodon wrote:

The initial release will be a starting point, while we learn from your feedback. The early focus is on creating Collections, with search and additional discovery options to come later. There are a few choices we’ve made (on size, following, etc) that we expect to revisit.

Mastodon wrote:

We’re ready to show you a new feature coming in Mastodon 4.6 - Collections. These are a way for people on Mastodon to curate and share bundles of accounts that they’d recommend to others, to help find connections and grow their network more quickly.

Boosted by aredridel@kolektiva.social ("Mx. Aria Stewart"):
whitequark@treehouse.systems ("✧✦Catherine✦✧") wrote:

monthly routine ^_^

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

@jwz Extremely reasonable!

Boosted by adele@social.pollux.casa ("Adële 🐁"):
HadasWeiss ("Hadas Weiss") wrote:

the mother is smiling

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

so as it turns out, i am in fact not an amateur and i had everything dialled in pretty fucking well for productivity for a good decade or two

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

I did not find this amusing.

https://freethoughtblogs.com/pharyngula/2026/04/09/not-funny/

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

i will say that the main difficulty with continuing to not smoke is that my productivity has taken a huge hit.

this is where you're all supposed to tell me you love me even if i'm completely unproductive.

Boosted by slightlyoff@toot.cafe ("Alex Russell"):
zackwhittaker ("Zack Whittaker") wrote:

Since North Korea has been in the news thanks to two massive hacks just days apart, I'm re-sharing my long-read primer on the ongoing and major threat from North Korean hackers, what they do, how they do it, and why they're incredibly successful — so much so that they could even be your co-worker.

https://this.weekinsecurity.com/thousands-of-north-koreans-have-secretly-infiltrated-us-and-european-companies-as-remote-it-workers/

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

I've been watching a technical meltdown about to happen for a few weeks, and the communication about it was very abstract, and _nobody_ but a few caremad nerds got any real conversation going. And nothing was done. Turns out: caremad nerds talking about abstract problems doesn't do shit.

But then the problem became less theoretical, and actual measurements of the problem came out, pointing to specific impacts and places and situations. Now people are taking action!

I wonder if there's a lesson in here somewhere. :P

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

Looking for some good, clean C++ fun? [1] Here's a vectorised HTML parser based on the research @lemire's incredible work from simdjson [2]:

https://source.chromium.org/chromium/chromium/src/+/main:third%5Fparty/blink/renderer/core/html/parser/html%5Fdocument%5Fparser%5Ffastpath.cc

[1]: yes, I understand. That's the joke.
[2]: https://arxiv.org/pdf/1902.08318

Boosted by aredridel@kolektiva.social ("Mx. Aria Stewart"):
filippo@abyssdomain.expert ("Filippo Valsorda") wrote:

Two papers came out last week that suggest classical asymmetric cryptography might indeed be broken by quantum computers in just a few years.

That means we need to ship post-quantum crypto now, with the tools we have: ML-KEM and ML-DSA. I didn't think PQ auth was so urgent until recently.

https://words.filippo.io/crqc-timeline/

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

Every morning I open Ivory and see a bunch of notifications ... vanish before I can read them because people have deleted the posts.

I really dislike this, both the behavior, and the way the app enables it.

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

Doing research into the performance and memory overhead of the most popular "CSS-in-JS" approaches, and my question for the authors of these systems is always the same: are you proud of this?

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

Tech journalists that do not understand prohibitions on shipping software as power relations, or do not understand technology enough to connect API access to profits and predation, are not useful as narrators of the present moment:

https://infrequently.org/2026/04/the-web-is-an-antitrust-wedge/

Boosted by dysfun@treehouse.systems ("gaytabase"):
acarsdrama@live.acarsdrama.com ("ACARS Drama") wrote:

Air to Ground Message:

CERTIFICATE WILL BE SENT TO YOU TO PROUDLY DISPLAY AT YOUR DESK.

Area: Washington DC, USA
Type: Embraer 175 (Enhanced Wing)
A: #a5bc77a5724
F: #f6e484d3f4a

#acars #vdlm2

Boosted by baldur@toot.cafe ("Baldur Bjarnason"):
dingemansemark@scholar.social ("Mark Dingemanse") wrote:

I was supposed to finish this last week but then the #Claude Code leak happened, promptly giving me an excellent opening example (h/t @jonny for their digital archaeology work that drew my attention to the magic prompting techniques)

(I think it is likely btw that #Anthropic shifted the #Mythos announce forward to this week to bury the leak & its security implications)

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

currently playing "where the fuck's the beef" with claude mythos amongst all the proclamations of THIS IS IT. the openbsd "zero day" does not AIUI in fact appear to be one, for example - just a non-exploitable bug. what about these much hyped claims checks out?

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

when a doctor makes their own cheese, call that medi feta