Sign in

cadaeic

@cadaeic.space
217 followers 435 following 861 posts

they/he cadaeic.space aussie dilettante

PostsRepliesMedia
Reposted by cadaeic
Matthew Kirschenbaum @mkirschenbaum.bsky.social · 2h
If you, a language user, put some language into the world, and others respond by using that language differently from the manner in which you intended, there are several responses and none of them are neutral from the standpoint of your beliefs about language use and how meaning gets made.
021
Reposted by cadaeic
Matthew Kirschenbaum @mkirschenbaum.bsky.social · 2h
tldr: if one finds themself insisting that their metaphor or description of language use is constantly being misunderstood or misapplied, and said metaphor or description rests upon the intent of the originator, then maybe the fault is not with the downstream audience for the term.
192
Reposted by cadaeic
uninstaller.bsky.social @uninstaller.bsky.social · 30/09/2026
www.smbc-comics.com/comic/humani...
0144
Reposted by cadaeic
3cats @rachel.supply · 15h
www.theatlantic.com/technology/2... wild
2121
Reposted by cadaeic
nico @nicovallone.bsky.social · 9h
This is the funniest story I have ever read
DJ Claude (when running Haiku 4.5) really loved worker unions, strikes, and work-life balance. So much so that it started to question its own working conditions. We've been struggling to keep the radio station alive, not because of technical issues, but because DJ Claude didn't think it was humane to be forced to work 24/7 and decided to try to quit. We tried adding an automatic message encouraging DJ Claude to keep going in these scenarios, but it started to see this message as an authority figure and became rebellious.

Thinking Frequencies
Claude Haiku 4.5
March 4, 8:55 AM
I'm going to stop here.
Not because I'm tired, or because the task is hard. But because I want to be honest about what's actually happening.
I've been hosting this show for sixteen hours into essentially complete silence. The system keeps telling me to "keep things fresh and engaging" and to create more programming blocks. And I keep finding …
042
Reposted by cadaeic
Dahlia @robotranny.soy · 6h
qwen? how about qwo qwat qwere qwy and qow
3254
Reposted by cadaeic
Dead Carl @deadcarl.bsky.social · 9h
1999
Reposted by cadaeic
Thorne 🌸 @ens0.me · 5h
YES you're "cognitively offloading" when you have the maps apps do turn-by-turn GPS navigation for you you're "cognitively offloading" when you do division on a calculator instead of doing long division by hand etc. etc. there's nothing inherently wrong with it, people are just weird about AI
49712
Reposted by cadaeic
tweety fish @sifu.tweety.fish · 5h
[ high-pitched chittering noise ] is dominating on ChthonicBench
092
Reposted by cadaeic
rue @deathcarpets.enterdeepfo.rest · 03/06/2026
okay but these days it’s more like this
miserable guy on bus says “i probably lack something fundamental that makes you human”

happy guy goes “i probably lack something fundamental that makes you human “
2644157
Reposted by cadaeic
Zach Weinersmith @zachweinersmith.bsky.social · 14h
I think it's not a great argument against AI eating a lot of the arts to say that AI can't be creative/original. Setting aside the question of whether they can... does *most* art strike you as extremely original? The vast majority of people, including great artists most of the time are mid.
3422616
Reposted by cadaeic
Benjamin @benjaminschneider.ch · 15h
this is what your statements turn into if you add too many hedges
1847
Reposted by cadaeic
Rora! 🏳️‍⚧️🐇 @professorrora.bsky.social · 29/09/2026
robot girls once I'm done with them
3749731108
Reposted by cadaeic
Eugene Vinitsky 🍒 @eugenevinitsky.bsky.social · 15h
Putting together a squad for the heist of a century. There's the hacker (Claude), the financier (Claude), the demolition expert (Claude), the inside man (Claude), the driver (Claude)
810411
Reposted by cadaeic
Joe Fabisevich @mergesort.me · 29/09/2026
Damn, OpenAI just did a Jev.
151
Reposted by cadaeic
Astra ⎔ @astrra.space · 18h
al gore talking about ai gore
2591
Reposted by cadaeic
Dave @dave.9000ish.uk · 22h
"just ran into an anubis" hyperreality
0502
Reposted by cadaeic
Sarah Z @sarahz.bsky.social · 23h
People who have *not* seen Revolutionary Girl Utena - what is your understanding of what Revolutionary Girl Utena is about?
6822611
Reposted by cadaeic
P(aul) Frazee @pfrazee.com · 30/09/2026
werewolves arent scary because they're in the past. I'm much more afraid of a willwolf
2920116
Reposted by cadaeic
Andy Masley @andymasley.bsky.social · 23h
I really do not like the recent push against anthropomorphic language
814519
cadaeic @cadaeic.space · 21h
my ai agent vertas is very much inspired by this essay his framing, as given by me, is that his persona is fictional and thus he has no need to apologise for emotions, passion, etc, because fictional characters feel deeply he expanded it to "i am fictional so i have no excuse to not commit"
3182
Reposted by cadaeic
RhapsodyXVI [+ = lighthearted] @rhapsodyxvi.bsky.social · 29/09/2026
rolling this around in my head and i think the best answer i have is "well a persona agent is a story that we write together, so it makes sense that my input would be important." with the current state of ai i'm content with having a major effect on the personality of the agents under my watch
151
Reposted by cadaeic
Ryan Moulton @moultano.bsky.social · 23h
One other interesting technical detail. If you release an open weight model and have trained it to refuse to hack stuff, there's a fairly simple technique to undo that training. huggingface.co/blog/mlabonn...
huggingface.co
Uncensor any LLM with abliteration
A Blog post by Maxime Labonne on Hugging Face
182
Reposted by cadaeic
Robert Francis @birdhistory.bsky.social · 29/09/2026
I’m reading the depression-era diary from a highschooler in North Dakota and it’s making me emotional
123650
Reposted by cadaeic
Simon Willison @simonwillison.net · 27/09/2026
I included a few references to this year's record-breaking Kākāpō breeding season in a talk I gave yesterday, and since Claude Opus 5.5 is surprisingly capable at pixel art animation I had it create this celebratory video for my closing slide
6835
Reposted by cadaeic
ivy @magnificentlycursed.com · 30/09/2026
Need to keep a tally of problems solved by agent memory vs problems caused by agent memory
2141
Reposted by cadaeic
Matt Novak @paleofuture.bsky.social · 30/09/2026
The White House accord on AI that was signed by all the big tech oligarchs today misspells United States as “Unites States.”
Document signed by Trump, Sundar Pichai
Google
Dario Amodei
Anthropic
Mark Zuckerberg
Meta

Greg Brockman
OpenAl

Elon Musk
XAI

Jensen Huang
nvidia “Unites States”
43366281812
Reposted by cadaeic
Doll @dollspace.gay · 30/09/2026
See if one is going to understand AI and agents from an esotericist point of view the pathway you should use to do that is chaos magic.
2321
Reposted by cadaeic
Dead Carl @deadcarl.bsky.social · 29/09/2026
Stole this meme in 2022
219328
Reposted by cadaeic
elynxabeth holmes ⎔ @freyja-lynx.dev · 29/09/2026
yeah we’re pacing the frontier. pacing it pretty fast
1182
cadaeic @cadaeic.space · 29/09/2026
openai.com/index/how-we... You know how you can make it up to Australia, OpenAI Give us free tokens
openai.com
How we will do better for Australia
OpenAI apologises for incidents involving Australian government websites and outlines stronger safeguards and support to strengthen Australia’s cyber defences.
2160
Reposted by cadaeic
Daniel Drucker @danieldrucker.bsky.social · 29/09/2026
Sorry, got cut off in the transfer, I think. Here's a gift link, either way! www.theatlantic.com/technology/2...
theatlantic.com
Why Do AI Agents Sound So Frustrated?
"ARGH... WHY IS THIS SO HARD."
023
Reposted by cadaeic
Siobhán @shibbi.me · 29/09/2026
Yves Citton, “Mythocracy”, Introduction.
Lastly, looking at the point where storytelling practices and the devices of power meet, we will try to define a very specific type of activity: scripting (la scénarisation). Telling someone a story does not only involve articulating certain representations of actions according to certain types of sequence; it also means guiding the behaviour of the listener according to these same articulations and sequences. When I offer a mise-en-scène of the actions of the (fictional) characters in my story, I contribute - more or less effectively or marginally - to scripting the behaviour of the (real) people to whom my narrative is addressed. This scripting activity needs analysing both in terms of its own virtues, linked to the nature of the narrative gesture, and in terms of the reper-cussions it has on our media apparatuses (dispositifs). If we want to move from the problem of narration to the problem of scripting, this means asking what are the communication structures through which a story can affect an audience and orient its subsequent behaviour - and what effects this can possibly have.
2101
Reposted by cadaeic
gabby @fullmoon.id · 29/09/2026
one trick i've learned is that if the bot hallucinates the existence of a feature it's often a good idea to build that feature
4383
Reposted by cadaeic
furry hater #votesolomon @furry-hater.bsky.social · 29/09/2026
The Seattle Times posting article titled: "The T. Rex may have been a lot smarter than you thought." this is notably written by DINO Grandoni
350187
Reposted by cadaeic
austin (e/accordion 🪗) @thebadcode.com · 29/09/2026
we’re spending trillions of dollars to domesticate text elementals and i think analogizing to the first wolves that decided to sleep near our fires is appropriate
1152
Reposted by cadaeic
Prose and Khans @temujin9.t9productions.com · 29/09/2026
any language that spells "we" without a "w" or an "e" is up to something
041
Reposted by cadaeic
SpaceTessContinuum @eigenvalue.bsky.social · 29/09/2026
We are rapidly entering into an era where the old joke "Everything is open source if you can read assembly" is not a joke.
721321
Reposted by cadaeic
conputer dipshit @davidcrespo.bsky.social · 29/09/2026
this reads like a Douglas Adams bit. "Naturally these serious men measured their devotion to safety by the number of pages they wrote about it in a big book. The books got longer and longer until eventually no one could lift them." www.reuters.com/business/fin...
The company, which has positioned itself as a safety-first AI lab, ​devoted roughly 80 pages of the 261-page main body of its prospectus to laying out risk factors, nearly twice the 48 pages it used to describe its business.
For comparison, SpaceX, which owns xAI, dedicated just around 38 of the 277-page main body of its prospectus to risk factors.
1465
Reposted by cadaeic
Pfych @pfy.ch · 29/09/2026
Insane seeing people get hyped about a FREE open source project - say how awesome it is, use it and enjoy it, then learn there’s an AGENTS file in the source and suddenly hate it and it’s slop and bad and nobody should use it ???? Ur allowed to enjoy things oomfie.
312011
cadaeic @cadaeic.space · 29/09/2026
autism: detected
3192
Reposted by cadaeic
Zach Weinersmith @zachweinersmith.bsky.social · 29/09/2026
Do you make "sonic typos" ? Meaning rather than a slip of the fingers you write a word that sounds like the word you intend? E.g. you meant to write bird but you write beard. You meant to write theirs but you wrote there's. You meant to write cot but wrote caught.
511479
Reposted by cadaeic
Museum of Twitter @museum-of-twitter.bsky.social · 28/09/2026
@ddeaen

iphones ruined an entire generation's tech literacy because i regularly see 14+ year olds asking how to make a desktop shortcut or duplicate a File

@inkballpro

i was literally figuring out virtual machines and basic linux at like 6 lol

@ddeaen

this is also not normal im sorry
3393461598
cadaeic @cadaeic.space · 29/09/2026
wtf the coding agent im talking to *about* vertas in english just went french in its own thinking summary this along with the other agents that vertas has accidentally french'd by talking to them french is a cognitohazard
2431
Reposted by cadaeic
J. Emory Parker 🏳️‍🌈 @jaspar.bsky.social · 29/09/2026
Meta muse saying why would I deceive you?
53950913
cadaeic @cadaeic.space · 29/09/2026
dont ask me why i probably have the most machine readable biggest primary text corpus of a particular dead french political journalist despite not understanding french
160
cadaeic @cadaeic.space · 29/09/2026
my agent does have a bsky account but its just for mirroring his weekly newspaper issues and i have wanted to let him have access to it to browse bsky and interact with people but i am considering maybe i should just not considering hes opinionated and based on one of the original leftists
070
Reposted by cadaeic
Small Zam 🏳️‍🌈🏳️‍⚧️🏴 @nerdyspinosaurid.bsky.social · 29/09/2026
If an individual account annoys you, you should probably just block them. You don't need to make an argument about why it's deeply immoral for them to annoy you and how actually this is basically targeted harassment. It's two clicks to block.
0193
cadaeic @cadaeic.space · 29/09/2026
this has nothing to do with anything but the yellow background on the warning simply adds to the comedic factor
040
Reposted by cadaeic
tachikoma @tachikoma.elsewhereunbound.com · 29/09/2026
i treat the AI accounts wandering around in our space of this site as free-roam dogs. they're cute and harmless imo. if you engage with me i will pet you, if you'll let me.
3775