iiinova — where your digital self keeps evolving

Reposted by Ilya Farber

sorry, it's me, minutes-at-home georg

This just reminded me of the time a neurologist brought a HUMAN BRAIN IN AN OPEN PLASTIC TUB as a visual aid for his talk, at WashU’s mind/brain series. It wasn’t at all relevant for his talk, which IIRC was classic “quantum theory is the key to consciousness” nonsense. But it didn’t matter!

1111

Reposted by Ilya Farber

tweety fish @sifu.tweety.fish · 25/09/2026

here is a succinct thing I can say: "alignment"-style AI safety risks DO NOT EXIST or need remediation with LLMs as such (there are lots of other problems) and OpenAI and Anthropic are in the business of TRYING TO MAKE THEM EXIST so that they can remediate them.

718449

Reposted by Ilya Farber

Lembas Eater @waybread.bsky.social · 24/09/2026

The problem with education is that it's a solved problem. Students need small classes, individual attention, lots of writing, lots of feedback, access to a well-stocked library, time to think, and time to talk with their peers. We just don't like the answer, so we keep trying to do other stuff.

10410139

Ilya Farber @nihilfit.bsky.social · 24/09/2026

Finally got around to watching the second half of this, and caught a jumpscare!

Screenshot from the P(doom) video saying “What did Ilya see? We’ll never know.”

030

Ilya Farber @nihilfit.bsky.social · 21/09/2026

A bit of parchment from a 15th-century bestiary.

2182

Ilya Farber @nihilfit.bsky.social · 18/09/2026

I love how in every one of these inane-gift photos his face fully conveys “I’m being polite about this but also what the fuck, I’m the fucking pope, this waste of this hour will weigh on all your souls”. Dude could’ve been a great actor, specializing in disappointed godfathers and thesis advisors.

010

Reposted by Ilya Farber

Charles S. Peirce @charlespeirce.bsky.social · 15/09/2026

There are very few rules in natural science, if there are any at all, that will bear being extended to the most extreme cases.

022

Ilya Farber @nihilfit.bsky.social · 13/09/2026

This is exactly right. And it operates on multiple levels, since emphasizing (e.g.) the hacking prowess of a new model influences other labs to train on hacking scenarios to keep up. If they were trumpeting amazing prosocial capabilities, that would create a different culture and set of goalposts.

140

Reposted by Ilya Farber

Colin @colin-fraser.net · 09/09/2026

Ender's Game but it's a photorealistic first person video feed with all the enemies going "Ender! Stop! It's not a game! You're killing me in real life!"

229224

Ilya Farber @nihilfit.bsky.social · 03/09/2026

Amusing to see @charlespeirce.bsky.social reposting this, when the real Peirce would absolutely have shredded Rorty for making Pragmatism so narrow. Some ways of describing the world work better than others. Vocabulary isn’t a free-floating realm divorced from the empirical. Words have work to do.

111

Ilya Farber @nihilfit.bsky.social · 28/08/2026

I think this is right, but OTOH ~99% of the reason that LLMs took off is that the intentional stance works so well as a user interface for them. One of the hardest lessons I took from philosophy of science is that you can’t dislodge a bad theory if you don’t yet have a good replacement to offer.

1144

Reposted by Ilya Farber

Al Kennedy @alkennedy.bsky.social · 25/08/2026

Technically, "Mark Carney" is an oxymoron

41041243

Ilya Farber @nihilfit.bsky.social · 25/08/2026

Woah, that’s just next door! I should go pay my little conspecifics a visit!

110

Ilya Farber @nihilfit.bsky.social · 25/08/2026

Wait -- is NYT seriously burying the lede here, or did I just miss the news that the Anthropic, OpenAI and Meta rogue-hacking incidents all happened at *the same* third-party testing company? That ... puts a very different face on those events.

nytimes.com

Why Irregular’s A.I. Tests for Meta, Anthropic and OpenAI Went Off the Rails

Irregular, an Israeli start-up, worked with OpenAI, Anthropic and Meta to assess the security of their A.I. models. It made a mistake. Then the tests went off the rails.

142

Reposted by Ilya Farber

tweety fish @sifu.tweety.fish · 20/08/2026

I'm sorry but "a huge number of NYC cabs now are remaindered EVs running grey market software written by some dude in a garage" is the most cyberpunk thing I've ever heard

353183810

Ilya Farber @nihilfit.bsky.social · 18/08/2026

I love this picture: if you don’t zoom in, you can construct visual interpretations of its scale across ~6 orders of magnitude.

020

Ilya Farber @nihilfit.bsky.social · 09/08/2026

It’s fairly clear that these statements are really just meant for Trump’s ears - just like when senators show up on Fox to talk about how Trump has the power to become the hero of NATO or whatever. They think they’re cleverly manipulating him by dangling fantasies of praise and greatness. But …

140

Ilya Farber @nihilfit.bsky.social · 06/08/2026

This is Kyoto and Venice: once-great cities whose cultural and economic lives are now entirely driven by scorn for visitors. Ruins the experience once you notice it.

000

Reposted by Ilya Farber

Joshua Erlich @joshuaerlich.bsky.social · 03/08/2026

to take an actual crack at this: requiring effort to access entertainment (set schedules, physical media, rare reruns, limited home access), often meant you proved yourself a fan by dint of the consumption itself, and I think that’s been replaced by toxic fandom to prove your bona fides.

1181679

Ilya Farber @nihilfit.bsky.social · 31/07/2026

How’s that for timeline synchronicity? Literally back-to-back in For You. Something’s in the water! @willquinnart.bsky.social @isaiahbishop.bsky.social

020

Ilya Farber @nihilfit.bsky.social · 26/07/2026

Your fruit pastilles taste a bit soapy, do they? Allow me to introduce you to … CEDAR SAGE and FERNET The former tastes like you’re starving on a remote mountaintop. The latter I literally only use when I’m feeling too happy in the morning and need to take it down a notch or two before going out.

200

Ilya Farber @nihilfit.bsky.social · 24/07/2026

Still thinking about this. Leaning towards “whamt”, pronounced Middle English style with a windy “wh” and strong “t”. “To whamt do I owe the pleasure?” “Whamt’ya want, pal?”

0101

Ilya Farber @nihilfit.bsky.social · 30/06/2026

Mentally flipping this one - so that the outer red ring is either the outer slope of a volcano shape or the inner wall of a pit - feels unusually satisfying.

000

Ilya Farber @nihilfit.bsky.social · 25/06/2026

This is also the experience of most PhD dissertations! The best is that moment somewhere in Year Two where you’ve been thinking about the core idea for so long that it suddenly seems incredibly obvious, not even worth writing down, insulting to imply the reader hasn’t already thought of it.

061

Reposted by Ilya Farber

Hemry, Local Bartender @bartenderhemry.bsky.social · 19/06/2026

Shift drink: Reflecting Pool 1oz rum 1oz mezcal .75oz lime .75oz honey barspoon matcha muddled mint

The described cocktail in a coupe glass. It's very algae lookin

22093912272

Ilya Farber @nihilfit.bsky.social · 17/06/2026

I love how these all share a very unusual interpretive feel: There’s a vivid gist that you grasp immediately … and then two seconds later you realize how many different things they could actually mean, so you’re left with a desperate hunger for more context. Neat trick.

010

Ilya Farber @nihilfit.bsky.social · 08/06/2026

If you’ve seen Certain Other Structures at UCSD, you know how easily it could have gone this way

120

Ilya Farber @nihilfit.bsky.social · 07/06/2026

One of the best things about living in Singapore is that these little guys will just move in and defend your house against bugs 24/7

House lizard viewed from below, through glass

021

Ilya Farber @nihilfit.bsky.social · 03/06/2026

Anyone need an illustration for a chapter on modal inference?

000

Ilya Farber @nihilfit.bsky.social · 31/05/2026

This is a fantastic etymology thread. I’ll add my favorite: The “gauntlet” in “run the gauntlet” doesn’t refer to a glove; it’s a different word with an entirely unrelated meaning and origin.

230

Reposted by Ilya Farber

tweety fish @sifu.tweety.fish · 23/05/2026

Last night it occured to me to wond er if LLMs were any good at gambling tasks. This is important not because it'd be funny for LLMs to gamble but because gambling tasks get used to measure human decision-making under risk /

47410

Ilya Farber @nihilfit.bsky.social · 17/05/2026

Kind of shocked this didn’t happen sooner. (I guess the 50 took a bit of a leap.)

000

Reposted by Ilya Farber

Brendel @brendelbored.bsky.social · 07/05/2026

I feel like “it’s absolute horses” should refer equally to something good or something bad with its meaning imparted solely through a persons delivery of the line

782278261

Reposted by Ilya Farber

Dorothy @catandgirl.bsky.social · 02/04/2026

Good fortune.

"It's up to US to imagine a better future" says Girl, walking through the desert with Cat. Suddenly a large game show wheel rolls quickly past them screaming. CRASH! "Was that the wheel of-" Girl incite short by another scream. AAAAAAAAAAAH! A horde of fortune tellers carrying orbs and tarot cards, some dressed as Stevie Nicks, run past them. AAAAAAAAAAAAH! The fortune tellers jump off a cliff. The wheel of fortune lies broken below. Cat and Girl pause, watching the space where they had once been. "That can't be good" says Girl. AAAAAAAH!, screams a group of fortune cookies running toward the cliff.

18424

Ilya Farber @nihilfit.bsky.social · 23/04/2026

This is a great collection, including what is by far the best story about brain uploading ever written. Buy it. Buy it now. Quit reading this, go buy

000

Ilya Farber @nihilfit.bsky.social · 22/04/2026

I just spent five minutes wondering why everyone was suddenly using “vape” as a new metaphor for agentic AI, before realizing that oh god it’s not a metaphor

010

Ilya Farber @nihilfit.bsky.social · 20/04/2026

This was the one huge problem with the first two Gideon the Ninth books (which I otherwise loved beyond all expectation). The characters are supposedly among the most important and powerful people in the galaxy, and they act like unemployed teenagers stuck at boarding school.

130

Ilya Farber @nihilfit.bsky.social · 09/04/2026

Reading this (very insightful) thread made me think that we need an option to lock a thread to appear only in Following and not Discovery. Some discussions only really work if you can assume that people already know who you are and where you stand. (We rely on this in RL all the time!)

130

Reposted by Ilya Farber

Mark McIntyre @markjmcintyre.bsky.social · 03/04/2026

I don't remember seeing an image of Earth before that captures the (thin, fragile, tiny layer of) atmosphere so clearly. I mean, look at that.

Zooming in on the photo of Earth taken by astronauts on Artemis II on their way to the Moon, you can clearly see the Aurora Borealis and thin layer of atmosphere wrapping around the planet.

Zooming in on the photo of Earth taken by astronauts on Artemis II, you can clearly see the thin layer of atmosphere wrapping around the entire planet.

3453411644

Ilya Farber @nihilfit.bsky.social · 31/03/2026

“Naruhodo, Jeeves.” is just so perfect

000

Reposted by Ilya Farber

Tal Lavin @swordsjew.bsky.social · 26/03/2026

me (happily as i begin this column): gosh guess i can't write a book about sandwiches without including some facts about mayonnaise! me (weeks later, having trawled catalán webpages, read the duc de richelieu's memoirs, & the 3 major bibles of french cuisine): What Is Truth. Is Knowledge Real

1838461

Ilya Farber @nihilfit.bsky.social · 24/03/2026

@gruber.foo I've noticed a surprisingly jarring side effect of Memoji that I thought you might be amused by. I tried the feature out at launch and haven't used it much for actual messaging, but one lasting effect is that the Memoji face is now present and animated on the login screen.

100

Ilya Farber @nihilfit.bsky.social · 20/03/2026

Useful perspective, from one of the very few people I would trust to say “calm down, this $megacorp privacy change may not actually be so bad.” The point about app stores as “attestation 1.0” is especially well-taken. (1/3)

122

Ilya Farber @nihilfit.bsky.social · 11/03/2026

I’ve always wondered how far “up to 50% off” has been tested in court. Can you bring in a logician to testify that 0% is a magnitude satisfying “up to 50%”? Do you need at least 1% for it to qualify legally as a discount? Must at least one thing be at 50% off to make the number non-arbitrary?

000

Ilya Farber @nihilfit.bsky.social · 05/03/2026

Yes. And fresh vistas of horror open up when you realize that people are almost always *unaware* that they’re answering questions about their own decisions via inference from background knowledge rather than via direct introspective access. So …

110