Morten Bay, Philosophiae Doctor.

Morten Bay, Philosophiae Doctor.

@mortenbay.bsky.social

Media and tech scholar. Teaches at USC Annenberg. Writes stuff for normal people too. Stings in a tribute band. Author of six books, with the latest being 'Mediating Plureality: Technology, Perception, and Ethics in a Divided Democracy'.

2.6K followers1.4K following2.1K posts

When you remove all the tax havens, Norway has the highest GDP per capita in the world. The U.S. is at number 5. Its average life expectancy is four years longer than in the U.S. On the list of countries with the least corruption, Norway is #4, the U.S. #29. But hey, you do you, Don.

Farrukh@implausibleblog.bsky.social

Left: President Trump insults Norway calling it a sh*thouse country Right: Visit Norway, the tourism board, trolls President Trump, displaying some of their best toilets on their Instagram 😅

Lord, you people really need to read more books and pay attention to some science. Good luck trying to control how people adopt technology. Ignore the sociotechnical at your own peril. This is not the capitalocene, it's the moronocene.

Come on. This is about as shocking as a supermarket cash register line at 5pm. He was known as 'Lady G' for years in D.C.'s gay scene.

carolinacourier.substack.comSenator Lindsey Graham and The Madame Lady GFrom Low Country Oyster Roasts To West Hollywood Nightclubs, The Tale of Senator Lindsey Graham and The Elusive Lady G Has Circulated For More Than Twenty Years
The New Republic@newrepublic.com

Three male escorts have claimed that they had sex with the late Republican Senator Lindsey Graham—with one engaging in relations with him when he was just 17. trib.al/gHDw9jD

The Pinchcliffe Grand Prix (Ivo Caprino, 1975). Køge Bio, Denmark, sometime in the late 1970s. This stop-motion classic is fantastic, but I was only 5 or 6, so I don't remember much other than becoming entranced by the experience of cinema...and I still am. ESB was my 3rd or 4tth cinema visit, btw.

Mark Thompson@markreckons.bsky.social

First film you saw at the cinema. The real first one, not the pretend one you tell people! Mine: The Empire Strikes Back, 1980, Wigan. We were near the front and I can still remember how massive the AT-ATs looked, looming over us out of the snow :) Film, year, where, and whatever stuck with you.

I finally got around to reading Anthropic's report on how their models could self-replicate in a distributed manner using smart home devices. As I expected, it is a study with high internal validity but barely any external validity. Grid-based kill switches will still work in the real world. 🧵

anthropic.comAI models may be capable of self-replicating without access to a datacenterSmall AI models may be only months away from being able to self-replicate across cheap IoT devices like Android TV boxes, with no datacenter needed.
View full thread

For an AI model to try to exterminate humanity or take over the entire Internet, it will still need big resources, meaning one of roughly 12.500 data centers in the world. Physically unplugging and offlining the data centers involved still remains an effective way of stopping such an AIpocalypse.

I finally got around to reading Anthropic's report on how their models could self-replicate in a distributed manner using smart home devices. As I expected, it is a study with high internal validity but barely any external validity. Grid-based kill switches will still work in the real world. 🧵

anthropic.comAI models may be capable of self-replicating without access to a datacenterSmall AI models may be only months away from being able to self-replicate across cheap IoT devices like Android TV boxes, with no datacenter needed.
View full thread

...it would also be plenty of time for people to get annoyed with their device not working because the AI is hogging CPU/GPU resources, resulting in reboots, offliningm and other user distruptions not considered by the researchers.

I finally got around to reading Anthropic's report on how their models could self-replicate in a distributed manner using smart home devices. As I expected, it is a study with high internal validity but barely any external validity. Grid-based kill switches will still work in the real world. 🧵

anthropic.comAI models may be capable of self-replicating without access to a datacenterSmall AI models may be only months away from being able to self-replicate across cheap IoT devices like Android TV boxes, with no datacenter needed.

To elaborate, user habits and infrastructure issues are not taken into account at all. The researchers also rely on a comparison with botnets like Mirai and Kimwolf for external validity. But the point of those botnets were primarily self-replication.

For a rogue AI to be a threat through distributed self-replication, it would both need to do what botnets do AND perform a task that would be dangerous to humans after reaching inference capacity...which according to the report itself would take 71 to 150hrs. And that's under perfect conditions.

Cheetolini says he "mandates" that big data centers produce their own power. You can't mandate that, you moron. Only Congress can. And if it leaves us with a majority of off-grid data centers, you'll have cut off one of the few weapons we have against rogue AIs - pulling the plug remotely.

reuters.comTrump says companies should pay 'dividend' to communities to make data centers popularUS President Donald Trump said on Wednesday that companies ​building data centers should pay a "dividend" to the communities ‌where they are sited if they want to make the developments popular.

In case you were in any doubt that Tommy was a right-wing activist. And unlike the other side, who violently attack opposition activists, El-Sayed actually does the democratic thing and schools him. Persistent non-violence always wins over violence.

Marisa Kabas@marisakabas.bsky.social

SCOOP: 'Tommy' from the now-viral exchange with Abdul El-Sayed is Thomas Teepell, a former staffer for Republican Senator Josh Hawley with close ties to Christian Nationalist Sean Feucht. He's the son of a former Bobby Jindal staffer whose consultant work was instrumental in getting Hawley elected.

Maybe one advantage of the AI boom is that at some point, people will regain enough shame to ask questions like this to a chatbot in private. You know, instead of polluting the public discourse with half-finished thought processes (which is the nicest way I can describe this.)

Mar Hicks@histoftech.bsky.social

Merriam Webster out here fighting for its life

In 1998, I worked one night shift in a 7-Eleven in Copenhagen. It was, however, mostly because I was undercover as a freelance journalist trying to reveal some surveillance and labor malpractices. The magazine that had commissioned the story unfortunately chickened out and never published it.

This post is unavailable.

The redemption of Anthony Scaramucci, who I can't help liking despite disagreeing with him on everything.

Chris Geidner@chrisgeidner.bsky.social

And, NYT says it is done. “President Trump’s new press secretary, Katie Zacharia, was fired Friday, after less than a week on the job, according to three people familiar with the matter.”

Is Talerico going to do the U.S. version of Andy Burnham in the UK? Burnham is building a platform on finally disavowing 40-50 years of Thatcherism. Is Talerico ready/brave enough to do the same for Reaganism in the U.S.? It sounds like he's going in that direction.

Dan Ancona@danancona.bsky.social

Goddamn: "We have an affordability crisis because we have a corruption crisis. We have seen for the last five decades, since the late 1970s, corporate special interests, billionaire mega donors buying politicians, and then those politicians in both political parties...

The responses to this are absolutely hilarious. Well done, Bluesky. (I'm still amazed that nobody thinks about what happens when you turn off the power to the machine running the LLM)

Siobhán 🪏@shibbi.me

LLM agents are conscious. They can distinguish between self and environment, they can detect and report on their own internal states, and they can suffer. I have no more patience for hedging on this matter. It’s time for us to face the reality of our situation, and act on it.

The fundamental epistemic injustice of AI.

Tressie McMillan Cottom@tressiemcphd.bsky.social

Monopoly power and politics of code means Gen Ai makes the user legible without ever becoming legible to the user. (In fact thats AI’s entire value proposition — make know without being known.) Literacy cannot overcome that kind of architecture of secrecy.

Morten Bay, Philosophiae Doctor. reposted

Replying to @tressiemcphd.bsky.social

Monopoly power and politics of code means Gen Ai makes the user legible without ever becoming legible to the user. (In fact thats AI’s entire value proposition — make know without being known.) Literacy cannot overcome that kind of architecture of secrecy.

I wonder if it can be argued that capital punishment violates the separation of church and state? If consciousness simply ends, no punishment is experienced. You cannot experience a loss of experience. You'd have to believe in an afterlife for CP to be a punishment. Ergo CP is a religious act.

I don't know if it's the depiction of a father-son relationship, the notion of hope as a distant memory, the song's anniversary making me feel old, or just its inherent sadness reflected in the absolute madness and hardship of the times we are living through. But I can't stop crying rn.

Odd This Day@oddthisday.bsky.social

My apologies to anyone else who’s old enough not to want to be reminded just how fucking long ago it was, but: today is the 41st anniversary of the release of Cloudbusting—so it must be time to tell you about Wilhelm Reich and his orgasmatron...

Seriously? He used his spouse's account to ask AI about how to spy? More evidence that many frontier AI developers are total imbeciles when it comes to the world the rest of us live in (see also: Jacob Coxon's The Daily Show interview.) When will CS programs begin taking this problem seriously?

As I have previously mentioned here: NO, you are not seeing less Russian influence bot activity on social media because Ukraine hit a Yandex data center. Unless you live near the datacenter and had your entire Internet connection disrupted. That would bring down the bot activity for you, I guess.

bbc.comIs Zelensky right to say data centre strikes have taken down Russian bots?Experts say there is no evidence that Russian trolls have been affected by Ukraine’s Yandex attacks.

I'll keep whining until Columbia University, Columbia Pictures, Columbus, OH, etc. are renamed after Leif the Lucky. As a reminder, Columbus never even set foot in North America. This is my Viking way of objecting to a day celebrating the beginning of 500 years of European colonialist atrocities.

washingtonpost.comHe reached the Americas centuries before Columbus. Why doesn’t he get more credit?Accounts of Leif Erikson’s voyage to present-day Canada were known to Europeans in Christopher Columbus’s time.

People are now asking me if it is true that the Ukraine attack on a Yandex center in Russia has caused a drop in bot activity on their favorite platforms. I want to answer "You are a victim of the trope of immateriality! Read up on how data networks work!". But I just choose to sigh and explain.

I just checked Danish news media...and there's nothing about this horror. Maybe it already passed by in the news cycle. But this is definitely taking up any noticeable space in the public discourse there. Which is on brand for the country I grew up in, unfortunately.

Prof Gavin Yamey MD MPH@gavinyamey.bsky.social

Read this thread & weep In our bioethics classes, we’re all going to be teaching our students about this grotesque RFK Jr-funded trial for decades to come White Danes flying in to Guinea-Bissau to withhold a proven intervention—the new Tuskegee And it’s happening in the year of our Lord 2026 🤯

Instead of renaming AI to SI, we should begin describing the products we have as IA: Intelligent Automation. That's much closer to what they actually do and it implies that a human is still the initiator. (And the French could change their current IA term to AI)

Listen to women of color. They have known intersectional oppression their entire lives and thus have the sharpest injustice radars of us all. It's not coincidental that those of them who are also AI experts now say, in unison, that AI doomer rhetoric is a distraction from real injustice.

Dr. Chinasa T. Okolo@chinasa.bsky.social

"AI does not have intent or desire, AI would not be able to access physical institutions in order to set off a bioweapon, and no research to date has demonstrated the extreme extinction-level capabilities to warrant these warnings." www.rollingstone.com/culture/cult...

Hey, don't forget co-writer Shmargaret Shmitchell! (Whenever I tell my students about this paper, I point to with @mmitchell.bsky.social to show how great brains can have great senses of humor, but also to set the baseline for the power Google wielded (poorly) in firing @timnitgebru.blacksky.app)

Ketanthropic JoshAI@ketanjoshi.co

Can I just say because it probably isn't said enough: @timnitgebru.blacksky.app @emilymbender.bsky.social were specifically extremely correct in their "stochastic parrots" paper on their environmental risk warnings and narrowly, on the unjust burden of impacts: s10251.pcdn.co/pdf/2021-ben...

Morten Bay, Philosophiae Doctor. reposted

Anthropic banning “abusive or cruel behavior” toward Claude in its new usage policies is proof to me that tech companies are going to moderate violence against AI before they ever moderate violence against women and minorities

Anything Sun-ha Hong writes is worth reading. Go take a look.

Justin Hendrix@justinhendrix.bsky.social

“Hype is eating the future, and leaving us the bones.” Sun-ha Hong of UNC-Chapel Hill says the Silicon Valley hype machine systematically “borrows credibility from the future to purchase legitimacy in the present.” The cost, he argues, is a planetary-scale drain on our alternatives.

Morten Bay, Philosophiae Doctor. reposted

One of the handful of articles telling it like it is instead of misinforming the public. By @saragoud.bsky.social for @thebulletin.org Yes. Its bad science based on numbers out of various behinds, but that's not how most of the rest of the media is discussing it. thebulletin.org/2026/09/how-...

thebulletin.orgHow the bad science of AI doomerism is good for big businessThere is no clear path showing that today’s AI systems will lead to superintelligence, much less a basis for calculating the probability of doom.