Mastodon Feed: Posts

Boosted by pixxl:
glyph ("Glyph") wrote:

@mitsuhiko @miguelgrinberg

> for the most part I have seen that eventually people cave and start using these tools at least for themselves in one form or another or they have to interface with people who do

I think you are wildly underestimate the depths of bloody-minded determination, both from me personally and from the rest of the community. It's easier to resist than you imagine and we are _much_ angrier than you think.

Boosted by pixxl:
glyph ("Glyph") wrote:

@mitsuhiko @miguelgrinberg

> the speed by which security issues are found by LLMs

everything about the slopsec apocalypse that I've experienced so far is that we dumped $300MM into security research and got about $50MM of value out, because it all used LLMs. It's just a super inefficient way to find bugs, but thanks to investor insanity, the resources are free if you do it this way. $50MM is a hell of a lot of security research! We can't ignore it! But that doesn't mean this is the future.

Boosted by pixxl:
miguelgrinberg@mstdn.social ("Miguel Grinberg") wrote:

@mitsuhiko @glyph

> I'm genuinely curious if you think this policy will last much longer.

I'm scratching my head on this take, Armin. Speaking about my own policy banning LLMs in my projects, what mysterious force do you think is going to make me change it? Why is it so hard for boosters like yourself to accept a "no"?

You don't like my policy, fine, look for projects that are more to your liking, but don't come finger-wagging, as it shows a lack of respect for the choices others make.

Boosted by rmrenner ("The Old Gay Gristle Fest"):
KaraLG84@dragonscave.space ("Kara Goldfinch") wrote:

One of the first text adventures I tried playing was a DOS game called "Legends" by Ed Roman. I found it in the "extras" folder on a Kurzweil 1000 demo CD in 1999.
The only place I could find it online was buried in an ISO of a newer Kurzweil 1000 CD on the Internet Archive, so I uploaded the game to the IF Archive.
For some odd reason I can only seem to get it to run in my Windows XP virtual machine's DOS Promt. It Throws a Runtime error in DOS Box.
Anyway If you want to try one of the games that got me into interactive fiction/text adventures, have at it.
https://www.ifarchive.org/if-archive/games/pc/legend.zip
#InteractiveFiction

Boosted by ireneista@irenes.space ("Irenes (many)"):
Unixbigot@aus.social ("Kit Bashir") wrote:

"It's a robot, get it!"

"My pronouns are she/her, thank you"

"I don't care what you call yourself pervert!"

"Oh, have you been using the internet. Please let me explain"

"Explain what, vikker?"

"The three laws of privacy. A robot may not record video of individuals not wearing a privacy preferences beacon. A robot must delete recordings within one hour of a privacy revocation. A robot may not share video without a digital signature from all pictured parties."

"Whatever, WRECK IT, boys!"

"You may notice I said nothing about not harming humans..."

"Ooof, you hit me when I wasn't looking!"

"You want to watch the next one coming?"

#Tootfic #MicroFiction #PowerOnStoryToot

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

If this is real, i am absolutely vibrating with anticipation. A Nathan fielder documentary about Elizabeth Holmes is one of the most "this is for me" things I can imagine
https://www.worldofreel.com/blog/2026/9/6/surprise-telluride-screening-is-holmes

Edit: in case it wasn't clear this is a rumor, don't want to be spreading that misinformation

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

Am I annoying?

Boosted by bcantrill ("Bryan Cantrill"):
mikeolson ("Mike Olson") wrote:

I learned about https://www.pangram.com/ from @ahl and @bcantrill on the Oxide and Friends podcast while doing some yardwork today. Max Spero, Pangram CEO, joined the two of them to talk about detecting AI-generated text and images.

The story he told on the pod was good enough that I signed up for Pangram to try it out. Remarkably good -- confirms my strong beliefs on a bunch of online text.

I much prefer to follow and read real humans with real thoughts. Useful new tool in the kit.

andrewnez ("Andrew Nesbitt") wrote:

Powered by https://github.com/git-pkgs/licenses

andrewnez ("Andrew Nesbitt") wrote:

A little weekend experiment: https://github.com/git-pkgs/git-spdx

Boosted by pzmyers@freethought.online ("pzmyers 🕷"):
darth_hideout@mas.to ("Darth Hideout 🏳️‍🌈") wrote:

#Capitalism #Guillotines #Memes

ireneista@irenes.space ("Irenes (many)") wrote:

@glyph oh well written, very solid!

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

A few photos I shot in 2008 that I'm unreasonably fond of. The first is a photo of the shop window for the used book store in Hverfisgata. Second is of the original Kaffistofa in Hverfisgata run by Samhjálp, a free kitchen for the needy. Third is of a mannequin in a shop front.

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

Here's a 20 year old blast from the past. Took this one during Sigurrós's public concert in Iceland in 2006

Boosted by dysfun@treehouse.systems ("gaytabase"):
tinker@infosec.exchange ("Tinker ☀️") wrote:

Folks are asking if ClimateGPT is funding Internet Archive... and asking if that is why IA posted about using it on their Mastodon feed as a sort of ad.

Y'all.

It's the other way around.

Internet Archive actively supports Erasmus.AI (the makers of ClimateGPT).

Heck, the founder of Internet Archive - Brewster Kahle - is on the *board* of Erasmus.AI

(Also, if you go back through Internet Archive's blog posts... they've been actively supporting and fostering AI for a long while now.)

Quoted Post: https://mastodon.archive.org/@internetarchive/117219501245722161

Erasmus.AI Board: https://erasmus.ai/#board

Info on Internet Archive support on Erasmus.io: https://www.erasmus.io/

Blogposts about AI on IA: https://blog.archive.org/tag/ai/

andrewnez ("Andrew Nesbitt") wrote:

Dogs

Boosted by glyph ("Glyph"):
tef wrote:

> we have invented a new way of programming efficiently

oh, so like a high level language, where there's no loss of precision and higher semantic weight on terms, so you write less and get more

> no, we've found a way to generate a million lines of code for a modest fee, but it feels like you're doing less work

couldn't we, y'know not have to write a million lines of code to get a basic app up and running

> that's too hard :)

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

looking around for things to watch and i have a lot of food programmes i never got around to.

seems after a certain amount of poverty-induced poor eating, i don't want to see "this is what you could have won".

Boosted by pixxl:
kevlin ("Kevlin Henney") wrote:

I've noticed some folk in tech (people who are in a position to know better) ascribing sentience and even consciousness to LLMs.

I don't mean simple anthropomorphising, which is a near inevitability (we, as humans, project and see the world through ourselves), I mean actually believing that the hard problem of consciousness has been solved by fronting linear algebra at scale with a chatbot interface.

This is magical thinking. It's known as "the hard problem of consciousness" for a reason.

glyph ("Glyph") wrote:

I can't just publish one of these threads because the Actual Good version of this argument needs sourcing and sequencing and so much more _work_ but auaugghghhhhh

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

“Bad Vibes Coding – Spectre Collie”

https://spectrecollie.com/2026/09/05/bad-vibes-coding/

> My take on generative AI has pretty rapidly gone from a tentative, “it sucks and I hate it, but after the bubble pops, there might be a few very specific use cases for it, like with programming,” to a vehemently zero-tolerance stance. It’s all garbage, burn it down.

Boosted by pixxl:
p@hj.9fs.net wrote:

#neverslop

dedicated to @ozzelot@mstdn.social

#ai #claude #chatgpt #vibecoding #programming #linux #kitty_odysseus

glyph ("Glyph") wrote:

RE: https://mastodon.social/@glyph/117225718993260555

this blog post is just _leaking_ out of me in every conversation that I have these days

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

breaking bed

Boosted by cstanhope@social.coop ("Your weary 'net denizen"):
cstanhope@social.coop ("Your weary 'net denizen") wrote:

Logging off like...

Attachments:

• gifv: Short clip of Spock smashing the everlovin' crap out of a computer terminal with his bare fist.

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

Because of its historic popularity as a murder weapon, thallium has gained notoriety as "the poisoner's poison" and "inheritance powder"

"inheritance powder"? really? 😅

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

@ireneista full story here if you're adequately interested https://www.youtube.com/watch?v=mqSuuGlJod8

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

@ireneista no, it ploughed into the back of another train and derailed

ireneista@irenes.space ("Irenes (many)") wrote:

@dysfun if the train was taking on water and sinking, we wouldn't personally object to the usage though it's a unusual

........ was that what it was doing? because we kind of want to know the story ....

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

huh, someone just said "capsized" of a train.

is that even valid?