I want a podcast or similar wherein people who believe crazy shit about GenAI or anything else are just asked why they believe these things. We can call it “pile of why’s”
David Knuffke
@davidknuffke.bsky.social
American Teacher well away from America. Worker Bee. Pretty sure that I’m both imperfect and doing mostly the right things.
I genuinely do not understand why this is illegal or grounds for criminal sanction. I also believe that companies like Spotify do something functionally identical.
Steve "It's TAY-tro" Tetreault@drtlovesbooks.bsky.social · 8h ago"Guy Who Made Millions by Creating AI Music Ouroboros Receives 18-Month Prison Sentence" So, dude had AI bots posting music to streaming, and AI bots listening to his bot-music; raked in over $8M. And I don't hate him. 1/ gizmodo.com/guy-who-made...
Famously, changing the name changes the nature of the thing. Billionaireism is a disease.
Aaron Rupar@atrupar.com · yesterdayBAIER: The president says 'SI.' Are you into SI or AI? BEZOS: I do like SI, because 'artificial' is not a very flattering thing. Like, nobody wants an artificial sweetener, artificial flavoring.
This week's post is about a way I have taken to thinking about the projects people attend to in schools (and other places): www.davidknuffke.com/blog/2026/20...
Everyone Has Their Babies | David Knuffke
A way to think about the various projects in a school, why we fall in love with them, and why adapting to reality beats defending an imagined ideal.
davidknuffke.com
Remember when OpenAI fired this guy for being unethical and then he wormed his way back into being CEO? Pepperidge Farm remembers: www.theguardian.com/technology/2...
Accept ‘bad things’ in return for benefits of AI, says Sam Altman
Boss of OpenAI calls for a regulatory light touch because of the ‘good stuff’ the technology can deliver
theguardian.com
US frontier labs just criming it up left and right. www.davidknuffke.com/blog/2026/20...
Call Them What They Are | David Knuffke
Frontier models aren't 'going rogue' when they hack government websites—they are executing unsupervised goals. Frontier labs should be held criminally liable for the crimes their systems commit.
davidknuffke.com
Dare Obasanjo@carnage4life.bsky.social · 5d agoIgnoring advice to secure and monitor your sandboxes during testing then blaming it on your AI being so smart that it’s a threat to humanity is nasty work
Number of teachers that STILL say they don’t need the mic in a faculty meeting is notable here.
Ryan J. Gallagher@ryanjgallag.com · 6d agoIf you are at a conference use. the. microphone. It doesn't matter how loud you think you are, it doesn't matter if you are a speaker or audience member. You should try to accommodate people who are hard of hearing (otherwise you're excluding them!)
This week’s piece is about the cybercrimes that frontier labs are committing, and why they don’t want to be held responsible for committing said crimes: www.davidknuffke.com/blog/2026/20...
Call Them What They Are | David Knuffke
Frontier models aren't 'going rogue' when they hack government websites—they are executing unsupervised goals. Frontier labs should be held criminally liable for the crimes their systems commit.
davidknuffke.com
I don’t feel any more blame lies with people who use the terrible Meta agent than I do with the victims of scams, more broadly. And also, you very much should not use the terrible Meta agent.
Open AI hacked the US Government. Its systems didn’t “meddle”. It, the company, committed a cybercrime, to go along with all of the other cybercrimes it has committed: www.nytimes.com/2026/09/25/t...
OpenAI’s A.I. Went Rogue and Meddled With U.S. Government Websites
The company did not learn until recently that its technology had meddled with websites for the Education and Commerce Departments and the Securities and Exchange Commission.
nytimes.com
If a company commits crimes, they should be held criminally liable. This is not a new concept.
Raphael Satter@raphae.li · 2w agoNew: Another OpenAI agent hacked the Australian government. PM says he expressed his disappointment "that it took [OpenAI] way too long to inform the government what had occurred, and the nature of the way that that notification occurred as well was unacceptable." www.theage.com.au/politics/fed...
This is useful!
Christian Moore-Anderson@cmooreanderson.eurosky.social · 2w agoSo, I trialed recording a "how I'd teach" video: How I get students to understand the symbols of stock and flow models. I learnt: buy a mic, remove the flicker, make sure the whole page can be seen. I'll make more through the year. www.youtube.com/watch?v=1YB-...
This week's piece is about my general distaste for gimmicks in LLMs: www.davidknuffke.com/blog/2026/20...
No Parlor Tricks | David Knuffke
Generative AI is plenty useful without forcing it to do stupid things.
davidknuffke.com
@johngreensbluesky.bsky.social : Got a note from @bookshop.org that they might not be able to send me my signed pre-order. That feels like the kind of thing that is both bad for the fans and bad for the brand. Any insight into how this kind of thing happens? It’s a first for me, for sure!
As someone living in Singapore, I’m not sure that framing it as a program that anyone really cares about “working” is an accurate read on the vibe for this program as rendered in Singaporean culture.
Larry Ferlazzo@larryferlazzo.bsky.social · 3w agoSingapore’s New Program To Pay People To Read Is Not Going To Work. Here’s Why…. larryferlazzo.edublogs.org/2026/09/19/s...
The other day I asked where the “I Ain’t Gonna Play Sun City” vibe was. Apparently it was just a few days away. I would pay for an “I Ain’t Gonna Play With Sheeran” shirt.
Una Mullally@unamullally.bsky.social · 3w agoThe Irish musicians, Beoga, who play alongside Ed Sheeran, have withdrawn from Sheeran’s tour. This is the statement they published on Instagram:
Someone cognitively offloaded the onerous task of understanding the most obvious lyrics in Bowie’s entire catalog. Incredible move by Meta here. What a stupid, terrible, company.
Ian Johnson@mrianjohnson.bsky.social · 3w agoMeta uses David Bowie's Five Years song in this AI ad "News had just come over We had five years left to cry in News guy wept and told us Earth was really dying And all the somebody people I never thought I'd need so many people Five years, that's all we've got" — 'Five Years', David Bowie
This week's piece is about the need for educational leadership to involve putting one's self on the line for the beliefs and actions of the school system in which you are leading. A notion that I think generalizes quite well to all sorts of leadership: www.davidknuffke.com/blog/2026/20...
Neck-First Leadership | David Knuffke
The leadership value of a line in the sand, the hazards of institutional ambiguity, and what it actually means to enable others.
davidknuffke.com
Is the moral arc of the Univers me supposed to bend toward justice? How do we go from “I Ain’t Gonna Play Sun City” to “do not speak of Palestine” in 40 years?
Rolling Stone@rollingstone.com · 4w agoEXCLUSIVE: Macklemore Dropped From Ed Sheeran Tour After Onstage 'Free Palestine' Statements Story: www.rollingstone.com/music/music-...
We don’t have one of those, though… Trump Says a Smart President Is All That’s Needed to Rein In A.I. www.nytimes.com/2026/09/14/u...
Trump Rejects Talk of A.I. Regulation and Calls Out Anthropic’s CEO
The president again rejected calls to try to regulate the industry, even as some of its leaders are speaking more openly about the risks of rapidly developing artificial intelligence.
nytimes.com
This week's piece is about the thing that is the most overlooked in the "Does AI work in education?" conversation: www.davidknuffke.com/blog/2026/20...
Does AI work in schools? | David Knuffke
What is to be understood from “LLMs in education” studies
davidknuffke.com
This is why biologists do not spend a lot of time chatting with creationists.
Benjamin Riley@benjaminjriley.bsky.social · last mo.Corresponding with a mother who says she enrolled her child in Alpha School so they would be surrounded by "high agency people" and when I ask what that means she sends a cut-and-paste job from ChatGPT and I just.... (gestures wildly and feels his sanity slipping further into the abyss)
The reaction to this on Bluesky is absolutely on-brand for the faction of the American left that has failed to build a meaningful coalition for a decade now. If you want to win things, you’re going to need to talk with people you find disagreeable and explore interest convergence overlaps.
Rep. Alexandria Ocasio-Cortez@ocasio-cortez.house.gov · 2mo agoThere is pervasive corruption across our healthcare and agricultural systems that defy party lines. I meet with MAHA activists because a lot of their concerns are bipartisan issues that major corporations paint as partisan. The fight isn't left and right, it's top and bottom.
Working in a Fancy International School it has been apparent for some time that being taught by humans is going to be a privilege marker in schools going forward.
Nick Covington@covingtonedu.bsky.social · last mo.Whole book, Montessori education for me, "cognitive science in a bottle" for thee.
Unless and until Open AI posts the logs from these cases, there’s really no choice but to assume they are hiding the bot actively encouraging these behaviors. futurism.com/artificial-i...
OpenAI Is Now Facing Over 50 Consumer Harm and Wrongful Death Lawsuits
OpenAI was hit with 30 new lawsuits overs its alleged role in the Tumbler Ridge school massacre. It now faces over 50 consumer harm lawsuits.
futurism.com