Many are appropriately outraged by Altman’s comments here implying that raising a human child is akin to “training” an AI model.
This is part of a broader pattern where AI industry leaders use language that collapses the boundary between human and machine.
🧵/
📉 Soziale Ungleichheit ist ökonomisch riskant: Studie der #TUBerlin & @pik-potsdam.bsky.social zeigt, dass bereits wenige sehr wohlhabende Haushalte durch ihr Spar- & Konsumverhalten eine Rezession auslösen können.
➡️ Je größer die Kluft zwischen Arm & Reich, desto instabiler wird die #Wirtschaft.
Warte ja noch drauf, dass er mal bei Wrint auftaucht!
I'd like to see some coordination in the open source community to respond to this. The outcome could shape significant funding and strategic moves.
Off the top of my head, there's a few things that I'd like to see in a new European open source strategy. 🧵
2025 was the year of teaching git to designers
Floskeln, die ich 2026 nicht mehr in der Presse lesen will:
"[Handlung x] spaltet die Gesellschaft."
"[Person y] spricht endlich Klartext."
"[irgendwas Belangloses] löst Empörungswelle aus; hier von uns cherrypickte Reaktionen auf Social Media, die diese Behauptung belegen sollen."
Congrats! Looking forward to see a copy in my mailbox.
Great HCI still happens but it’s usually tugged away in (the few industrial, and academic) research labs. For reasons mentioned below, they have a crazy tough time to get their foot in the door with product orgs. Even if the ”novel” thing was already well explored in academia.
I really hate it when scientists keep saying that “we need to rebuild trust in science,” because it implies that scientists are to blame for the mistrust rather than the millions of dollars of dark money that have funded political attacks on science in order to advance a far right agenda.
Wenn du sie nicht wiederfindest, kenne da eine Apotheke. Schreib ne DM, falls du die Adresse brauchst.
Die Kürzung der Ticketsteuer fürs Fliegen führt zu Mindereinnahmen von 350.000.000 Euro im Bundeshaushalt.
Das ist genau der Betrag, um den die Bundesförderung für das Deutschlandticket gekürzt worden ist.
Prioritäten. So wichtig.
you can't make a technical contribution in HCI by just connecting an LLM to something. we need to know how it works, beyond a user study with a few people. what are the ways that it fails? how fundamental are these problems, what does it reveal about the potential of this kind of use, etc.
"Der Publizist Michel Friedman warnt davor, dass Deutschland in zehn Jahren keine Demokratie mehr sein könnte. »Ich bin erstaunt, wie viele Menschen ihr Leben normal weiterleben (...)".
www.spiegel.de/panorama/leu...
Because of the rise of pundits and influencers and the relative extinction of professional critics and public-facing experts, we have utterly lost the distinction between matters of opinion and matters of fact. (Yeah it's not a total dichotomy)
Karpathy: "English is the hottest new programming language" Meanwhile me: "What xml structure should I use to get better llm eval results?"
Good write up by Nick Heer at @pxlnv.bsky.social:
> We should also not lose sight of the increasing hostility of the U.S. government making these moves to reassert its dominance in [tech]. We can stop getting steamrolled if we want to, but we *really* need to want to.
pxlnv.com/blog/us-incr...
Have only read the abstract and eval metrics (will read the rest once I have some time) so slightly unrelated but curious: Have you by any chance experimented with trying to correlate "suspense" through embeddings as well? Like in @wattenberger.com's work? wattenberger.com/thoughts/yay...
Product managers tell me that you don't need UX when you can just "build measure learn" your way to product market fit.
But if your sense-response mechanism is entirely built around trying new shit, and the new shit fails - what are you going to do? Continuing to tweak it only upsets users more.
Every rich person is going to tell *you* how great AI teaching is while sending *their* kids to the kind of schooling the Ancient Greeks would recognize. I just wish everyone would think about why that is.
Is anyone using anything else but Arq for (cloud / s3 / glacier) backups on macOS these days?
Do you write TypeScript? Enjoy Automerge? Love building both tools and community? You might be a great fit for our ✨new✨ Automerge TypeScript Maintainer role!
www.inkandswitch.com/jobs/automer...
(Remote role, 🇬🇧 UK-based preferred but not required)
Ich erinnere gerne nochmal an diesen Text von @georgloewisch.bsky.social über Jens #Spahn.
Es steht alles drin.
Die Bude voll mit rechtsradikalen Abgeordneten, die Identitäre und Neonazis als Mitarbeiter beschäftigen - aber die Bundestagspräsidentin macht Jagd auf Fahnen, die gesellschaftlichen Zusammenhalt, Vielfalt und Offenheit symbolisieren. www.tagesspiegel.de/gesellschaft...
chatbots für therapiezwecke ist nicht nur deswegen gefährlich, weil es einem quatsch rät (das passiert auch), sondern weil eine llm nicht "nein" sagen kann. ich mein, sie kann schon das token "nein" generieren, aber ein wirkliches "nein" muss letztendlich notwendig materiell sein. ein widerstand.
Damn, how did I completely miss this... Hope this works out well enough for you to make another batch in the future. Super interesting work!
300.000 Tote bisher, die Mehrheit davon Kinder, dank Musks „Kürzungen“, Zahl steigend. Aber „Selbstheilungskräfte“, alles klar. www.nytimes.com/2025/05/30/o...
Figma (Sites) creating inaccessible, non-semantic div soup? I’m not surprised at all. That’s basically the main point of the talk I just gave at #btconf. Figma is for painting beautiful pictures of websites. Not for making websites.
Rolf Mützenich und Stegner verteidigen das skandalöse Treffen. „Jeder Kontakt“ zu Russland sei sinnvoll. Das ist m.E. keine Naivität mehr, das ist das alte deutsche koloniale Denken, dass Moskau und Berlin gemeinsam Weltpolitik machen. Gerade jetzt brandgefährlich für die Ukraine und Europa.
1/2
Our colleagues over at @msftresearch.bsky.social have tried to dissect the metacognitive demands a little more. Being aware of and considering them in the first place helps formulating good prompts imho. www.microsoft.com/en-us/resear...
