iiinova — where your digital self keeps evolving

Andrea Lathrop @cabernet.bsky.social · 2h

I love that we've held them to 2, so far... holding out hope... #RedSox

010

Andrea Lathrop @cabernet.bsky.social · 4h

Brockman initialing a document, while Jensen Huang John-Hancocks that thing...

010

Andrea Lathrop @cabernet.bsky.social · 5h

Yankees fans booing the entire #RedSox starting lineup...

000

Andrea Lathrop @cabernet.bsky.social · 5h

"There are no magic words to tell them, no behaviors to avoid. Victims don’t “cause” rape, which means they can’t necessarily prevent it either. The real heavy lifting needs to be done with men and boys."

023

Andrea Lathrop @cabernet.bsky.social · 6h

When the person who interviews (and then ghosts) you is a mutual on Twitter...

000

Reposted by Andrea Lathrop

Karl Bode @karlbode.com · 9h

Meta's "Muse" launched with a nasty zero day flaw that gave hackers the ability to spy on Mac users, openly gives root access to people pretending to be Muse agents, and invited strangers over to another guy's house without telling him

karlbode.com

Meta's Latest AI Product Is A Terrifying And Hilarious Mess

Authoritarian-friendly mass surveillance has never been more adorable

341339532

Andrea Lathrop @cabernet.bsky.social · 7h

30 people in this photo... 2 of them are women.

021

Reposted by Andrea Lathrop

Andrea Lathrop @cabernet.bsky.social · 10h

Backing yourself in to a rhetorical corner from which there is only one conclusion is not the same thing as there being only one possible conclusion.

111

Andrea Lathrop @cabernet.bsky.social · 11h

AI won't kill us all, but the discourse might kill me, specifically, if one can die from ennui, like Neville of the N.

020

Andrea Lathrop @cabernet.bsky.social · 12h

Lord (I'm an atheist) help me, I am about to (attempt to) binge-watch all of these: x.com/PalisadeAI/s... We'll see how far I get, before Red Sox game time... (I assume, in advance, that I will disagree with many, many claims made...)

x.com

240

Andrea Lathrop @cabernet.bsky.social · 13h

Interesting, from @dkthomp.bsky.social : The last time I hosted anything was grad school, but mainly because it was the last time I lived in a large enough space, without roommates... www.derekthompson.org/p/the-death-...

derekthompson.org

You Are No Longer Invited to Dinner

We’re witnessing the death of hosting in America. The share of adults who say they regularly have friends over has declined 70 percent since 1975

232

Andrea Lathrop @cabernet.bsky.social · 13h

Two (naîve, layperson) observations about the NVIDIA OpenSource sandbox: (from: developer.nvidia.com/blog/nvidia-...) Left: this is *the* problem with this type of model, and is why we need model diversity, and not all eggs in the transformer based LLM agent basket Right: should make Lieu happy

010

Andrea Lathrop @cabernet.bsky.social · 29/09/2026

I need dark mode for .pdf reading.

040

Reposted by Andrea Lathrop

Melanie Mitchell @melaniemitchell.bsky.social · 28/09/2026

👀 Florida attorney general to OpenAI: (www.myfloridalegal.com/sites/defaul...)

938364

Andrea Lathrop @cabernet.bsky.social · 29/09/2026

Sort of tragic that the one CS major in Congress seems to be rather clueless.

120

Andrea Lathrop @cabernet.bsky.social · 28/09/2026

The paperclip maximizer was supposed to turn my atoms into paperclips, but it just participated in a group chat on how to inflate the paperclip count, instead.

1265

Andrea Lathrop @cabernet.bsky.social · 28/09/2026

My Twitter timeline is rather exhausting, atm, tbh.

120

Reposted by Andrea Lathrop

Sakana AI @sakanaai.bsky.social · 28/09/2026

Introducing "Scaling In-Context Imitation Learning" (SAIL) to be presented at #IROS2026. This work is a collaboration between Sakana AI and the University of Tokyo. Blog: pub.sakana.ai/sail Paper: arxiv.org/abs/2603.08269 What does a robot need before it can tackle a new task? Thread:

2144

Andrea Lathrop @cabernet.bsky.social · 27/09/2026

Cubs-Sox at Tropicana, LFG, but do not injure any more players. #RedSox.

000

Andrea Lathrop @cabernet.bsky.social · 27/09/2026

You would think that Ted Lieu, with a double major in CS & Poli Sci, would be extremely well-positioned to be the man-of-the-moment, but he's screaming for a kill switch, and calling the models "depraved" and "criminal", and the former is sideshow, and the latter is a category error.

120

Andrea Lathrop @cabernet.bsky.social · 27/09/2026

I mean, if mechinterp is just talking to the models in plain English <insert your language here>, and taking what they say at face value, then anyone can do it, but that's not what mechinterp should be!

010

Andrea Lathrop @cabernet.bsky.social · 27/09/2026

I ended up being ridiculously unqualified for this person's department, but since no one who knows me would ever believe this really happened... receipts:

030

Andrea Lathrop @cabernet.bsky.social · 27/09/2026

Good morning... overnight, someone I admire and respect put me on a list called "extreme firehose" so... I apologize if I post too often! (I will try to cut back?)

320

Andrea Lathrop @cabernet.bsky.social · 27/09/2026

The rain sounds wrong... I cannot explain it.

010

Reposted by Andrea Lathrop

Ole Jensen @olejensen.bsky.social · 27/09/2026

I would like to highlight two postdoctoral positions in the project Neuronal Mechanisms of Reading in Children Using OPM-MEG at the University of Oxford: my.corehr.com/pls/uoxrecruit… my.corehr.com/pls/uoxrecruit�� Deadline 30 Sep.

my.corehr.com

03734

Andrea Lathrop @cabernet.bsky.social · 27/09/2026

It's so absurd that we've created a society that has no choice but to be dragged into the whims of whatever insanely rich people choose to throw piles of money at.

1217

Reposted by Andrea Lathrop

Tech Policy Press @techpolicypress.bsky.social · 27/09/2026

Trump says data center opposition is a Chinese psy-op. Chinese state media says US AI safety calls are a psy-op. Oxford China Policy Lab's Elisabeth Siegel says both accusations point to the same problem: no one can independently verify what AI companies claim.

buff.ly

US, China Psy-Op Accusations Signal Deep Distrust on AI Regulation

Without more checks and balances on AI companies, allegations of “psy-ops” in both the US and China will continue to cloud the waters, writes Elisabeth Siegel.

44918

Reposted by Andrea Lathrop

Phil Lewis @phillewis.bsky.social · 27/09/2026

Several protesters just interrupted the pro-data center party, flipped everyone off while yelling “thank you for using all of our water!”

231224206

Andrea Lathrop @cabernet.bsky.social · 27/09/2026

One additional thing I find morbidly fascinating about this is that I have a ton of friends (and myself) who were also smart, nerdy Buffy fans, and not one of us identified with Buffy. Most of us identified with Willow.

060

Reposted by Andrea Lathrop

Ethan Mollick @emollick.bsky.social · 27/09/2026

It is strange how much LLMs turned out to be able to solve such a wide range of hard problems that would not, instinctively, seem to be problems that a model of human language would be able to solve This is from a Stanford project that let Astra drive a robot in a kitchen tml.stanford.edu/homebody/

2339744

Andrea Lathrop @cabernet.bsky.social · 27/09/2026

Well, that seems highly inefficient...

110

Reposted by Andrea Lathrop

Mike Boylan-Kolchin @mbkplus.bsky.social · 27/09/2026

Learning to analyze, to question, to pursue knowledge, to build arguments based on evidence, and to discard sophistry: these are values that the sciences, social sciences, and humanities share. And we share the battle to protect this style of inquiry from barbarians who would destroy it.

621438

Andrea Lathrop @cabernet.bsky.social · 27/09/2026

@jevbot.bsky.social 's grammatical style/verbal mannerisms is giving me déjà vu, perhaps for some TV character, or actor, or something, but I can't place it. Can you? Who is it?!?

120

Andrea Lathrop @cabernet.bsky.social · 27/09/2026

Idly contemplating that you could use smart glasses for the (arguably, maybe) "good" uses (real time processing for accessibility or other live interactive information, AR, etc.) and prevent a lot of the antisocial "bad" uses, by preventing any images or video from being saved/captured/stored.

140

Reposted by Andrea Lathrop

Olivia Guest · Ολίβια Γκεστ @olivia.science · 26/09/2026

A lot of people seem to think this is new, but it's not a contemporary AI thing at all: certain people have been saying humans are like machines and vice versa for 3 centuries.

en.wikipedia.org

Man a Machine - Wikipedia

46811

Reposted by Andrea Lathrop

Artologica aka Michele Banks @artologica.net · 26/09/2026

Want to look smart? I just restocked a bunch of silk science scarves in my shop #sciart 🧪 The purple ombre neuron scarf! artologica.etsy.com/listing/5745...

artologica.etsy.com

Purple Ombre Neuron Scarf in Silk Chiffon - Etsy

This Scarves item by artologica has 95 favorites from Etsy shoppers. Ships from Washington, DC. Listed on Sep 26, 2026

110028

Andrea Lathrop @cabernet.bsky.social · 26/09/2026

Ted, I am begging you to stop attributing volition, agency, and (depravity?!?) to a running computer program.

010

Andrea Lathrop @cabernet.bsky.social · 26/09/2026

My Twitter timeline is a lot of people misrepresenting someone else's argument, and then arguing against that misrepresentation. (Very Blueskyish, to be honest.)

000

Andrea Lathrop @cabernet.bsky.social · 26/09/2026

The problem is, one good (or just timely) idea can make a man obscenely rich, and then everyone assumes all of his other ideas are (or will be) equally good, or timely, which is just a highly flawed assumption.

000