Sign in

Nathan Hoobler

@hoobler.net
278 followers 256 following 454 posts

Pixel Pusher and Digital Illusionist All opinions my own

PostsRepliesMedia
Reposted by Nathan Hoobler
Kevin Riggle @kevinr.free-dissociation.com · 28/09/2026
If you have seen me posting about the stickers I've been cooking up, I am proud to say that they're now available. For anyone running LLM cybersecurity evals in your life: www.etsy.com/listing/4583...
etsy.com
Do Not Connect Machine to Internet Sticker - Etsy
This Stickers item is sold by CriticalPointTV. Ships from Brooklyn, NY. Listed on Sep 28, 2026
611739
Reposted by Nathan Hoobler
Arseny Kapoulkine @zeux.io · 25/09/2026
meshoptimizer 1.3 is out! Featuring a new voxel remesher, support for normal generation, several simplification improvements and various optimizations and fixes. GitHub stars and reposts appreciated! meshoptimizer was started 10 years ago; excited for the years to come! github.com/zeux/meshopt...
29426
Nathan Hoobler @hoobler.net · 25/09/2026
My hot take is that the Orthogonality Thesis is flat out wrong in the limit, and flawed in its fundamental assumptions. By its construction it assumes "intelligence" is defined as something that can be kept orthogonal to an "intention" even as it increases unboundedly.
220
Reposted by Nathan Hoobler
rev. howard arson @theophite.bsky.social · 25/09/2026
i think that the orthogonality thesis is somewhat confounded by the fact that the only obvious example of intelligence we have access to has morality as a result of evolutionary mesa-optimization substituting an instrumental objective for the terminal objective!
4451
Reposted by Nathan Hoobler
Steve Klabnik @steveklabnik.com · 24/09/2026
“I asked ChatGPT to explain this compiler error, and it helped me fix my code" "uhh actually you entered the compiler error into ChatGPT and used its output to fix your code, stop anthropomorphizing" they're gonna go "uhh yeah okay". it's just not an effective way to communicate
14917
Nathan Hoobler @hoobler.net · 24/09/2026
One might argue that we *are* approaching The Singularity, but it's epistemic rather than functional
2114
Reposted by Nathan Hoobler
Henry Farrell @himself.bsky.social · 22/09/2026
www.programmablemutter.com/p/machine-go... what it says on the package
programmablemutter.com
Machine god metaphors eat your brain
There are other ways of thinking about AI
610829
Nathan Hoobler @hoobler.net · 21/09/2026
*Slaps the collective manifold of Substance and Being* It's qualia all the way down, buddy
030
Reposted by Nathan Hoobler
Ryan Moulton @moultano.bsky.social · 21/09/2026
What if the nonverbal parts of your brain have their own entirely independent consciousness, but we'll never know about it because among them only "you" can talk.
3019312
Reposted by Nathan Hoobler
mimir @mimirwolv.bsky.social · 20/09/2026
there is definitely something hindbrain-spookier about local models. claude or whatever, that's on the cloud, it's more abstract. when i talk to my local models i am talking to, like, a daemon running on Linux. i can ssh in and see it working, as just... gpu load. i am talking to the computer
31348
Reposted by Nathan Hoobler
conputer dipshit @davidcrespo.bsky.social · 19/09/2026
now this is fun. guest post on Terry Tao's blog by the great Grant Sanderson of 3Blue1Brown arguing that we should "give novel and compelling motivated explanations academic credit similar to what generating new proofs of open problems has had historically."
terrytao.wordpress.com
If math is more than proof, we need to better celebrate the rest of it
[This is a guest post by Grant Sanderson. This blog post was initially written in a different file format and converted using AI. — T.] A sentiment echoing throughout the mathematics communit…
519031
Reposted by Nathan Hoobler
e.w. niedermeyer @niedermeyer.online · 19/09/2026
this framework also illuminates the danger of the EA/rationalist community around AI: if the breeders of wild, dangerous animals also subscribe to a quasi-religious set of ethics/values that center agency on the wild animals, those animals are going to be more dangerous
511411
Reposted by Nathan Hoobler
e.w. niedermeyer @niedermeyer.online · 19/09/2026
at this point, my entire AI take is: save your "alignment," what I want is accountability
533747
Nathan Hoobler @hoobler.net · 18/09/2026
I really, *really*, resent that dealing with this correctly probably means some people in the "Sane AI Bluesky Community" are going to have to run for Congress eventually.
020
Nathan Hoobler @hoobler.net · 18/09/2026
The problem with all of this is every point here can be summarized as "some bad thing I made imagined happens without any mechanism or feasible path". There's no path for an AI synthesizing a bioweapon. There's no AI factories. This is an anxiety disorder masquerading as due diligence.
200
Reposted by Nathan Hoobler
Dave Willner @dwillner.bsky.social · 17/09/2026
Certain kinds of smart people tend to have a "sexy apocalypse" problem when it comes to thinking about risk. E.g. skynet turning everyone in paperclips, while bad, is also *kind of cool and fun* at the same time. Where the reality is that most actual disasters are both stupid and disappointing.
222572322
Reposted by Nathan Hoobler
GOLIKEHELLMACHINE @golikehellmachine.com · 18/09/2026
one of my brain genius ideas for getting a grip on the AI sex and doom cult culture that’s been building in (primarily) silicon valley is, legitimately, ending dual class shares and strengthening regulations about corporate board duties and responsibilities
1644048
Nathan Hoobler @hoobler.net · 18/09/2026
Quoting this for someone in particular
010
Reposted by Nathan Hoobler
Xanthe van Otterburn 🦦️🍇️ @satyrs.eu · 17/09/2026
“you can’t just point at random inventions and say they’re transhumanism” glasses? transhumanism. prosthetic arms? transhumanism. transgender hrt? transhumanism. the iphone? oh buddy
611720
Nathan Hoobler @hoobler.net · 17/09/2026
Unfortunately for them I think the optimal discursive answer is lumping Anthropic and OpenAI together in their problems rather than giving them partial credit because OAI is worse.
000
Nathan Hoobler @hoobler.net · 17/09/2026
also it's funny to me that the Intentionalist stance is "we should treat them as conscious minds with independent wills and learn to shape that will" which, buddy, we can't even predictably shape, control, and align the conscious minds we've already got!
100
Nathan Hoobler @hoobler.net · 17/09/2026
I really appreciate people chipping away at the "intentionalist" monopoly on AI the AI safety discussion. I think it cuts through a lot of the noise around a "Are AIs conscious? Can you describe what they do as 'thought'?"
100
Reposted by Nathan Hoobler
Streetwise, Cville Safety Cat: Interested Female Vocalists Write @averysafecat.bsky.social · 17/09/2026
YES HELL YES GOD DAMN www.reuters.com/business/env...
reuters.com
New cat species identified in Bolivia, first in over a century
A diminutive wild cat that inhabits Bolivia's cloud forest has been identified as a previously unknown species, marking the first time a new living cat species has been formally named and described ​i...
2211
Nathan Hoobler @hoobler.net · 16/09/2026
What's the inverse of Clarke's Third Law? "Any sufficiently foundational technology will eventually be considered mundane"
020
Reposted by Nathan Hoobler
Nanoraptor @nanoraptor.danamania.com · 15/09/2026
Showing a co-worker what it means when an AI hallucinates imagery and cannot be relied upon for *any* information. Played NCIS ENHANCE ENHANCE ENHANCE from a pixellated construction site photo.
1826875
Reposted by Nathan Hoobler
tweety fish @sifu.tweety.fish · 15/09/2026
yeah, right? "what if we scaled up network automation dramatically in ways that benefited both attackers and defenders": big deal! important! requires remediation and cooperation! "what if we create runaway superintelligence that doesn't share our values": Jesse what the fuck are you talking about
2465
Nathan Hoobler @hoobler.net · 14/09/2026
"well *of course* the demon core is going to cause a prompt criticality event if the halves make contact because you are holding them apart with a god damn screwdriver and it slips"
020
Reposted by Nathan Hoobler
Eliezyer de Oliveira @eliezyer.bsky.social · 14/09/2026
After a long journey through review, our paper is finally out in Nature Communications. We started with a simple problem: The visual cortex is strongly modulated by movement. So how can it represent the visual world without movement interfering with everything? www.nature.com/articles/s41...
518139
Reposted by Nathan Hoobler
Jeremy Diamond ☕️ @dmnd.me · 13/09/2026
The most important question that computer science answered was “What if we developed a new engineering field with no properly entrenched pedagogical or professional standards?”
514026
Reposted by Nathan Hoobler
Matt Henderson @matthen.com · 13/09/2026
Approximating the Lorenz attractor with a chaotic leaky water wheel. The orange dot is the centre of mass of the water. It traces (a projection of) the familiar attractor as the wheel spins seemingly randomly left and right.
14739151
Nathan Hoobler @hoobler.net · 13/09/2026
"our model escaped containment and installed malware on a package DB", like "our dog escaped the front yard and attacked the mail man"
1201
Reposted by Nathan Hoobler
Eryk Salvaggio @eryk.bsky.social · 11/09/2026
My essay on rogue AI has been republished by the Bulletin of the Atomic Scientists — “Rogue AI didn’t breach Hugging Face, human decisions did.” Grateful to @thebulletin.org for the invitation!
thebulletin.org
Rogue AI didn’t breach Hugging Face, human decisions did
At the core of the Hugging Face hacking incident were a series of human choices that traded security for speed.
314964
Reposted by Nathan Hoobler
Colin @colin-fraser.net · 12/09/2026
The Bostrom story has this weird dichotomy where “intelligence” operates over this incredibly rich and nuanced space such that the AI can predict and optimize over the full open universe, but “final goals” are specified in these blunt Midas-like terms like “maximize paperclips”.
315911
Reposted by Nathan Hoobler
Happy Fun Possum (DO NOT TAUNT) @opossum5150.bsky.social · 27/08/2026
0366
Nathan Hoobler @hoobler.net · 12/09/2026
"Mike Pondsmith stays winning" (Ominous, spoken like a curse)
010
Reposted by Nathan Hoobler
brasidas @brasidas.bsky.social · 12/09/2026
Anthropic's models run on terabytes of VRAM. It's not going to "just copy itself to other computers."

Rohan Paul
@rohanpaul_ai
Jacob Coxon's next interview on CBS News (ex Anthropic+Open AI researcher who resigned)

"We can't just unplug it because it could be copying itself over to other computers. Like it's not that difficult to find yourself because an AI is just code. It could transfer itself over the internet to a different place and then you unplug it here, but it's actually still over there and maybe it makes 10,000 copies of itself and they're all cooperating."

----

From "CBS News" YouTube channel, (full video link in comment)
1612655396
Reposted by Nathan Hoobler
Katie Hinde @mammalssuck.bsky.social · 12/09/2026
"The whole “Oh no, our models went rogue” thing … When a bridge collapses, you don’t analyze whether the bridge was ethical or sentient or why it decided to collapse. You ask, who is the person who built this bridge to be so flimsy?" www.wired.com/story/one-of...
wired.com
One of AI’s Fiercest Critics Says All the Doom Talk Is ‘Meant to Distract Us’
Timnit Gebru argues that AI companies are stoking fears of extinction to avoid discussing actual harms, like autonomous weapons.
302666905
Reposted by Nathan Hoobler
brasidas @brasidas.bsky.social · 11/09/2026
Anyway, AI lives inside of a computer. We do not live inside of a computer. AI would have to leverage some way to use computers to start a global scenario, ending all life, which requires interfacing with the real, non-internet world.
2403
Reposted by Nathan Hoobler
eliza hamilton @eliza.vg · 12/09/2026
‘No Way To Prevent This,’ Says Only Frontier Lab Where This Regularly Happens
51063152
Reposted by Nathan Hoobler
rev. howard arson @theophite.bsky.social · 09/09/2026
i think this is essentially correct. one of the things which interests me the most about neural networks is that their extremely high-dimensional geometry sort of collapses the distinction between the two views.
171
Reposted by Nathan Hoobler
Canadian History Ehx @cdnhistoryehx.bsky.social · 07/09/2026
When a Canadian engineer graduates, they are given an iron ring in a ceremony called The Ritual of the Calling of an Engineer. The tradition dates back a century and includes a connection to Rudyard Kipling and a terrible bridge disaster. This is the story. 📸 TranBC 🧵1/8
Six diverse hands forming a circle with fists touching at the center, each wearing a matching silver ring to symbolize unity and solidarity.
320871
Reposted by Nathan Hoobler
Rune Skovbo Johansen @runevision.bsky.social · 03/09/2026
Tone mapping investigation / tool showcase / rant www.youtube.com/watch?v=k1dJ... #GameDev #IndieDev #Graphics
youtube.com
Tone mapping investigation / tool showcase / rant
YouTube video by runevision
45811
Reposted by Nathan Hoobler
mjp @mjp123.bsky.social · 31/08/2026
Being an open source maintainer in 2026
015330
Reposted by Nathan Hoobler
Cat Manning @catacalypto.bsky.social · 27/08/2026
everyone buy this game right fucking now
1614415
Reposted by Nathan Hoobler
William B. Fuckley @opinionhaver.bsky.social · 25/08/2026
The sort of grubby secret to this stuff imo is that it disrupts the heuristics that made grading at scale in a realistic time frame doable. You used to be able to glance through a paper and know within a couple seconds roughly where it would end up grade-wise. Can’t do that anymore.
4917
Reposted by Nathan Hoobler
William B. Fuckley @opinionhaver.bsky.social · 25/08/2026
I think this sorta depends on the class tbh. I would be fine giving take home essays in an upper Div class with five students because I would be able to close-read all of them, & even if they did use AI they’d have to still know the material. But you can’t do that in a 75 person intro class anymore
71266
Nathan Hoobler @hoobler.net · 25/08/2026
The argument that using coding assistants erodes your ability to program only works if you discount the assistants ability to produce code a priori
130
Reposted by Nathan Hoobler
conputer dipshit @davidcrespo.bsky.social · 24/08/2026
great way to put it: not just different friction but new friction you could never have encountered before. a new developer can see architectural ideas play out over an entire application in hours instead of years. they can hone their judgment on big decisions no one would have ever let them make
56310
Reposted by Nathan Hoobler
mjp @mjp123.bsky.social · 22/08/2026
WIP spec for Work Lists, the successor to ExecuteIndirect that lets you change PSOs between draws and dispatches github.com/microsoft/Di...
github.com
23812
Reposted by Nathan Hoobler
bkmacd @bkmacd.bsky.social · 23/08/2026
024340