A PhD student at Stanford noticed her classmates were asking Al to write their breakup texts. So she ran a study. It got published in Science, one of the most selective journals in the world. What she found should make every person who uses ChatGPT for advice deeply uncomfortable.

‘While some cling to the promise of an AI “revolution,” the cost of adoption is proving a stubborn bottleneck. These developments also suggest that the economics of replacing human labor with AI may be more complicated than some early forecasts originally implied.’ fortune.com/2026/05/22/m...

Other people have noticed this phenomenon too but danah boyd put the best slogan to it: social media is now parasocial media. Social media hasn't been social for a long while—how long depending on the platform, depending on the milieu. It's not folks sharing equally, but broadcasting and receiving.

Here are some flagrant examples, although the appendix is full of them of them too — basically experts and journalists alike seem to repeat this mantra that we do not understand how contemporary AI works. See appendix at the end here for more: doi.org/10.5281/zeno... 2/

🧵 We tend to think the internet is permanent, but VANISHING CULTURE—the new 📚 from Internet Archive—starts with a simple reality: our digital record is far more fragile than we think 🕳️ 📖 Read for free: archive.org/details/vani... 🛒 Purchase in print: www.betterworldbooks.com/product/deta... ⬇️

NEW: A new pre-print study offers the first causal evidence that outsourcing reasoning tasks to AI can rapidly erode users' independent performance AND their will to persist despite difficulty. "People’s persistence drops... they’re also not willing to try without AI.” futurism.com/artificial-i...

🚨New preprint and our results are rather concerning.. We find the "boiling frog" equivalent of AI use. Using large-scale RCTs, we provide *casual* evidence that AI assistance reduces persistence and hurts independent performance. And these effects emerge after just 10–15 minutes of AI use! 1/

Hey, I wrote a thing about AI in astrophysics ergosphere.blog/posts/the-ma...

The. Rise. Of. Cognitive. Surrender. Study finds that people who use GenAI chatbots rely on them 80% of the time, and develop almost no capacity to recognize when a chatbot is feeding them faulty information. papers.ssrn.com/sol3/papers....

"A new study out of Cornell Universit found workers most excited and impressed by corporate speak may be the least equipped to make effective, practical business decisions, and it can leave companies with dysfunctional leaders."

Simultaneously: - Microsoft bragging that 90% of the code its devs ship is AI generated - Microsoft demanding that its devs burn 10x fewer tokens because it's getting too expensive - Every week brings new and exciting emergency security patches for Windows 11

“I was disappointed to realize that from the waist up I could, on some days, look like I work for ICE.” www.nplusonemag.com/issue-52/ess...

Criminally funny, but also I'm crying inside Norway comes for the #TechBros www.forbrukerradet.no/breakingfree vimeo.com/1168468796?f...

I keep reading posts about people who wrote applications entirely with "zero coding." They say they don't read the code and instead refine the application using increasingly precise prompts. I have some questions: * Is the application nontrivial? 1/7

Built a palette extractor that thinks like a curator, not a pixel counter. – Physics-based color repulsion – Auto-detects muted vs vibrant bias – Adaptive spacing to hit target count – OKLCh color space (perceptually uniform) Early Access: please break it! chipper-otter-d6c600.netlify.app

The reason CEOs think genAI is so capable of doing everyone's jobs is because literally the only job genAI has been able to do from the get go, and will possibly ever be able to completely do, is a CEOs job.

it's not just that I find it chilling to imagine Sam Altman wearing my dead mom's face in order to sell me ads it's that the way we ACTUALLY keep our lost loved ones alive requires a myriad of human skills these assholes perpetually devalue, like storytelling and deep emotional connection

@ivanmorgillo.bsky.social and @sasasekulic.bsky.social's talk was beautiful, a wonderful chat with the audience about the use of AI for developers, friend of for? 💚

Kudos to @benkadel for making the most interesting presentation on admittedly not very interesting topic of regex in #kotlin and #Android #droidconIT2025

I'm somewhat disappointed I couldn't find any of the newer @scalzi.com books at the Dussman bookstore in Berlin, but still I managed to get something new to me - Starter villain! Also, this shelf is very different to how it looked like ten years ago, and only a few authors were in there 20 years ago

I am (self-evidently) very interested in language. And also in technology. From that perspective, the rise of LLMs has created many new things for me to think about. One is: why is it that certain very small clusters of words are *clearly written by an LLM*? What is the quality of that writing?

you don’t understand bro, if we didn’t lie, no one would buy our product 😩

I keep reading about things like detailed up-front Sprint plans, about "failed" Sprints, about people panicking when a Sprint doesn't go according to plan, etc. There is no agility in that thinking, whatsoever. 1/7

So, last Sunday my tuxedo cat, Tom, thought to check if he knows how to fly. He does not. He was in the window of my second floor apartment, got spooked by a curtain and jumped down. 🤦All in all it could've been worse - a broken elbow and a perforated palate. Now for two months of meowing. 😅 #cats

Begging someone to go into the Google offices and just spend every day yelling at random managers "remember when your products worked? Remember when they worked so well your company name became a verb? Have you considered doing that instead of shoveling this AI dogshit all day????"

What do you mean it's not #caturday today?!

I think it's now possible to make a poli-sci course that equips one for modern political analysis better than most classic theory and has a syllabus sourced entirely from random internet posts. Text 1. Wilhoit's Law, born as part of a 2018 blog comment crookedtimber.org/2018/03/21/l...

There are lots of avenues to challenge any study like this, especially when the sample size is so low (n=16). But it is 1000% more transparent in its methodology than ANY marketing claim about increased developer efficiency with AI tools. 1/2

looks into the camera

AI agents get office tasks wrong around 70% of the time, and a lot of them aren't AI at all

Often, when people are baffled by why change can't happen, they don't consider incentive systems. Incentive systems take many forms. Here are a couple: * Individual bonuses. These create a competitive culture within the team, and that typically leads to hoarding skills and information. 1/10

On vibe coding:

“Why GUIs are built at least 2.5 times Or Why I Think Lean Software Development is Wrong” patricia.no/2025/05/30/w...

*Career artisanal programmer eloquently beefing about AI copilots. He probably got fired this week deplet.ing/the-copilot-...

Just saw this interview where the question was “how far do you think we are from having AI operate as a junior engineer 24/7?” I cannot stress enough that I would pay a reasonable fee to 𝘯𝘰𝘵 have an unsupervised junior engineer operating 24/7 in our codebase.

Newsletter: We are in the era of the Business Idiot, where middle management has seized power, breeding out true meritocracy and value-creation in favor of symbolic growth and superficial intelligence. This piece ties together everything I've ever written. www.wheresyoured.at/the-era-of-t...

Physics is learned by solving physics problems. A guy who "learns" physics by talking to chatgpt for hours a day, can spend an entire year never once solving a single problem, but thinking that he is learning physics. A freshman on their very first day, has learned more physics than he has.

Spent too much time on this

"Individual developers report productivity gains using these tools. And at the same time, data shows that the more teams use them, the bigger the negative impact on team outcomes like delivery throughput and release stability. How can both be true?" codemanship.wordpress.com/2025/05/11/i...

Let's get ready to ruuuuumbleeeee!!! #caturday

Did AI kill your job? If so, I want to hear about it. The response to my story on the AI jobs crisis has been overwhelming. I heard from so many people impacted by AI that it made me want to collect and organize these stories to better understand the phenomenon. Share yours at: [email protected]

Mish and Tom are happy in back home and don't want me to travel again anytime soon #caturday

Here's a crazy idea: I've decided to give away 300 (300!) signed books on Bluesky today. Who knows if this will work but let's see, shall we? Be in the first 300 to reply to this & you can have one.(Limiting the offer to UK but happy to send overseas if people cover the postage.) More info below...

When the only service that works is your uptime page...

Really liked this @aaronrosspowell.com piece as an antidote to the idea, valorized by a lot of the internet left in the 2010s, that being an asshole is good praxis. That idea failed miserably. It's time to rediscover being decent. www.aaronrosspowell.com/p/the-necess...

NEW POST To work effectively with agentic coding assistants, Birgitta Böckeler found she needs to intervene, correct and steer all the time. She describes examples of these interventions indicating the skills we need to correct the tools' missteps martinfowler.com/articles/exp...

Just got a lovely, long email from a former Commodore retailer about my Tramiel piece. About how it helped him finally understand why Tramiel was the way he was in their interactions, over 50yrs ago "I finished reading the last paragraph of your article and just sat quiet for a moment or two."

The most important thing I can tell everyone that I've learned from being a YouTuber who explains how stuff works is that each and every time I thought "this is common knowledge and doesn't need to be explained" I was wrong.

"The Impact of Gen AI on Human Learning: a research summary A literature review of the most recent & important peer-reviewed studies " drphilippahardman.substack.com/p/the-impact...