Nice to see more dancing on the grave of Sal Khan’s blighted “vision” for AI in education. (I, ahem, did call my shot two years ago.)

“All of them are now pronouncing with equal confidence that the future is AI. But we weren’t consulted on the metaverse, either, and as it turned out that mattered a great deal more to the companies than to us. There are limits to how far big tech moguls can impose their vision of the future.”

Powerful thread on what's been a very foreboding week in the US and how we can continue to fight back. Inviting people to participate in civic resistance is key -- activism is contagious, and we need pro-democracy efforts to "break contain" by this November.

“The academic hiring system is simply not designed for candidates like me. It privileges a kind of performative intellectualism…This is not how science works anymore. Science works by prompting, iterating, and deploying.” God I wish I’d written this, it’s perfect.

The “Age of AI” is gonna open up so many opportunities for new businesses that are explicitly pro-human and anti-weird shit like this.

Many interesting thoughts here on AI in education, including this one: “One of the biggest surprises for me was the vocal opposition from a group of students around any AI usage at all—something that was almost entirely absent last year.” I expect this will keep growing.

This essay present a highly accessible overview of what we know, and don’t know, about artificial neural networks. A refreshingly sane, skeptical, and insightful take.

European countries were about a year ahead of the US on banning smartphones in school. Just an observation.

The three players in this photo: 1. Nigerian heritage, grew up in England, and is only a citizen thanks to birthright citizenship. 2. Born + raised in Holland, has a dad who immigrated to the US from Suriname. 3. Said he wanted to “make people who look like me proud” while playing on Juneteenth

I'm not really a "gamer," but I delight in evaluating generative AI by making it play human games. This week I enjoyed watching Agentic ChatGPT struggle mightily with the classic computer game Zork. This one is for all my fellow Gen Xers who remember being eaten by grues.

“In natural moments of doubt or uncertainty, students found more comfort in AI, lured into a falsehood and then ensnared and imprisoned in self-doubt. Instead of Aadya growing stronger, AI grew on what she fed it: her ideas, prompts, work ethic, engagement, and ultimately, her confidence.”

Call me crazy but I think deploying a tool of cognitive automation within a culture of algorithmic social influence is not good for education.

This is an excellent thread on the nature of what LLMs are doing and, nestled neatly within, an important observation about the limitations of human communication (and the context we are always inferring). Come for the broomsticks, stay for the genies.

This, but for all of education.

Dartmouth professor shares extended thoughts from his students on how much they loathe AI. A thread to make your blood boil.

Anthropic: We have invented a technology so dangerous to mankind we must move quickly to curtail it from destroying humanity. Also Anthropic: alright alright alright

This is so deeply shameful. I live in Texas and have many Muslim friends, the irony is that most are culturally conservative (and even support Trump to this day). But nothing matters when paranoia and white supremacy become the baseline position of the ruling party in this state.

Inspired by this research, I unleashed OpenAI’s agentic mode upon classic Zork. On the one hand, it’s impressive that it could play this sim at all! On the other, after figuring out basic gameplay, it got stuck looping in circles and randomly dropping stuff, like a clumsy lobotomized hamster.

Just a remarkable paragraph this morning from Ben Thompson of Stratechery (an unabashed AI enthusiast, mind you). I’ve written previously that OpenAI reminds me of Enron. But Anthropic? The only analogue that comes to mind is Gattaca.

The people who genuinely and without any artifice, believe this today are part of a long lineage in AI of mis-predicting the future. Some genuinely believed this would happen ***75 years ago***. To understand what’s happening it’s worth revisiting that history.

“The earth is the answer to every question. The earth is a face of an exulted lover; they watch it sleep and wake and become lost in its habits. The earth is a mother waiting for her children to return, full of stories and rapture and longing.” - Samantha Harvey, Orbital

“Americans hate data centers. They really, really hate them…The public seems to agree that data centers are giant, ugly, noisy, smelly altars to industrial-scale hostile architecture. In our virulently partisan country, this constitutes a rare show of consensus.” - @tressiemcphd.bsky.social

“Help”

O(M)G

Anthropic has been “unnecessarily antagonistic to open weight models [and] open research. [The company] has a specific view of AI, but such a powerful technology will never have its final equilibrium be one of singular control by a private company.” - @natolambert.bsky.social (Only if we act!)

It’s a pretty weird week when a guy who wants to be a trillionaire is simultaneously getting headlines for the economy-shattering IPO expected Friday and the racist pogroms he is accused of stoking — but these stories are viewed as separate and in no way connected or prescient of the future.

When these people insist the mere presence of people unlike them is an attack on them, their country, or their rights, this is exactly the outcome they are seeking, and disappointment among far right Trumpists is largely that he failed to produce events like this here. www.yahoo.com/news/us/arti...

Horrifying escalation of radical right-wing terrorism in Belfast, fueled in part by Musk and his algorithmic hate machine. I watched this playing out in real time last night as the mob walked through the streets, kicking on doors, lighting fires. It is pure fascism and it is a worldwide movement.

You gotta pinch to zoom to read it, but Apple’s new Siri will regurgitate texts from your friends before suggesting you…bring a Sauvignon blanc or rosé…to this zesty summer potluck. Behold, the future!

“AI is here to stay” — well maybe so, but even the companies making these tools aren’t sure what the product-market fit is supposed to be. Remember prompt engineering? Like, just last year?

What is punk? Punk is that ineffable human quality of raging through art. Which is something AI will never do. Inspiration this week from @brianphillips.lol @tressiemcphd.bsky.social @ali-alkhatib.com @gregpak.net @eryk.bsky.social, the Pope, some Cali 8th graders, and others.

No idea why “cyborg feedback” on writing would be an improvement to human alone; it’s certainly not true empirically, and questionable theoretically. These tools largely churn out bland, surface-level commentary in high volumes.

Nice level-headed thread here of what we are talking about when we talk about next-token prediction. Things aren’t the same as they were in 2024, so some updating of priors is necessary. At the same time…LLMs are still stochastic-y and parrot-y in important ways.

More and more we are seeing critical commentary on the potential harms of AI in education. Glad to see @edtechprof.bsky.social cited at length in this.

Hell yeah it is. And I'm looking for examples of punk resistance to AI, whether music or art or defaced subway advertisements or whatever, and punk means whatever you want it to mean. Drop a link in the replies to this message if you've got something for me.

This is such an elegant testament to our essential humanity. Cognitive scientists call this “theory of mind,” our capacity to imagine the thoughts of others, and at root it’s simply our ability to combine our curiosity and inquisitiveness about the thoughts of a fellow human.

“I won’t be a slave to AI AI is only imitation It’s not the voice of my generation Artificial intelligence can’t know the human experience” (By Knights of Molina, eighth graders; 4.6 million IG views) youtube.com/shorts/S_ncz...

Turnitin pivoting to the bold marketing strategy of “actually, it’s the humans teachers who cognitively offload their students!” Good luck with that.

Anthropic is the worst, not because they can actually deliver on this nonsense, but because they’ve willfully embraced the rhetoric of AI as quasi-divinity while trying to consolidate their own corporate power. It’s truly disgusting.

OpenAI is running a massive randomized control trial on 20,000 Estonian high school students. They apparently are not thrilled.

Agonizing over Tilly Norwood’s pronouns actually lends to the ruse that she is some discrete entity that exists in real life. It’s perfectly fine to use she/her pronouns for your interview with Tilly Norwood just as would be fine to use he/him pronouns in an interview with Toucan Sam.

“For now, roughly half of the university system’s ChatGPT licenses have been activated, suggesting that at the very least C.S.U. paid for tens of thousands of logins that have gone entirely unused.” This is the scandal in plain sight. Why is CSU paying for a product that is free?

A new report on AI in education from Australia refreshingly dispenses with the “but it holds so much promise!” verbiage and instead goes straight to cataloguing its many harms. More of this please! (Link to follow)

Yesterday my friend @eryk.bsky.social shared this poetic observation about the ineffability of being human. Who's up for some Sunday morning musings as to how it might be have been interpreted by pragmatist philosopher Richard Rorty re his project to "de-divinize" humanity? Just me? Thread! 1/n

Fantastic, @bcmerchant.bsky.social takes on the bullshit propaganda machine that is Anthropic, saving me the trouble (and doing it better than I ever could). They are the worst. www.bloodinthemachine.com/p/how-anthro...

The fun thing about this sort of mistake is that it’s understandable if you have a basic mental model of how LLMs work. Not conscious beings. Just token samplers capable of generating — wait for it — hallucinatory output.

“Roughly 84% of students said they used ChatGPT. About a quarter of them said they used the version provided by CSU and the vast majority said they used the free version.” Why then is CSU paying OpenAI millions? It’s never made sense to me.

Huh, so weird that ‘worrying’ about AI is gendered.

You know how Groundhog’s Day is a really funny comedy but also a profound meditation on living a purposeful life? This essay is the Groundhog’s Day of essays about tech and AI. Tremendous from start to finish.