This researcher created a fictional illness, and fake studies funded by the Professor Sideshow Bob Foundation and University of the Fellowship of the Ring and the Galactic Triad. LLMs warned people the illness was real. go.nature.com/4mDXXka

@patrickfreyne.bsky.social with all the right words, in the right order: "language is ultimately how we build community with one another. A world of LLM communication is a colder, blander one devoid of incidental human empathy and original thought."

I'm almost there ...

OpenAI: We’re burning money like the Joker. A miracle needs to happen for us to turn a profit Microsoft: Please please use our AI systems, we’re teetering on the edge here Anthropic: I wonder what’ll kill us first: lawsuits, regulations or model collapse Media and universities: AI is here to stay

More and more organizations are coming out against AI in schools.

He was ever the optimist. Time will tell if he was a prophet

HISTORICAL AFFIRMATIVE ACTION: These people have been dispossessed of the fruits of their own labor for centuries. They were denied the opportunity to build generational wealth and influence. We will strive to correct this. MAGA AFFIRMATIVE ACTION: Give me money because I like to say the N-word

Fascinating paragraph from Scott Alexander, where he admits he’s the dog that caught the car

To adapt the FDA language... "This encounter has not been evaluated by anyone. This PR stunt is not intended to depict, simulate, mimic, or reproduce any actual delivery.”

“The researchers found that 40% of workers had encountered workslop within a month, and then spent an average of 3.4 hours a month dealing with it – which the study estimates adds up to $8.1m in lost productivity for a 10,000-person organization.” www.theguardian.com/technology/2...

Looks like AI is not the magical tool that CEOs make it out to be.

BREAKING: Elizabeth Warren has introduced a bill that would allow you to file taxes for free directly with the IRS. The bill has support on both sides of the aisle, and would bring back and expand the free taxing filing program that the Trump administration killed off.

So my husband is doing interviews to hire a new HR person and already it’s become VERY obvious which candidates are using chatbots to conduct their interviews. It’s…not going well for them.

The kids are alright

The bottom 95% of taxpayers will see tax INCREASES on average this year, driven primarily by expanded tariffs and expiration of health insurance credits. Meanwhile, the richest 1% alone will receive $117 billion in tax CUTS in 2026. Never forget who Trump governs for.

Guardrails are a scam! “We have also continued to strengthen ChatGPT’s responses in sensitive moments, working closely with mental health clinicians.” It's impossible cos that's full blown cognition: either the lie is the automation (instead indentured labour is used) or they're fully fibbing. 1/

A completely unhinged post from the head of Norway's oil fund: to hunt down the "leaders" of groups resisting AI implementation in companies and "remove" them archive.ph/r0u2r

You cannot stop this with guardrails, it just has to be shut down. Anything else is lying.

I've been quietly removing tech from my school library since I took over 6 years ago. I thought people would be annoyed, but it's been great. Adults have barely noticed, and the students have responded well.

A couple weeks ago, Nicholas Keung at the Toronto Star reported on how generative AI is being used to assess immigration applications in Canada...and how it's led to refusals because of stuff the AI tool added in, that humans (clearly) failed to check. www.thestar.com/news/canada/...

GenAI dipshits are still at that stage of the scam where they keep insisting how useful it is as an accessibility aid (the next stage is insisting it helps the global south), and no it fucking isn't. GenAI actively *destroys* accessibility tools like captions and alt text.

Join us for a discussion of "The AI Con: How to Fight Big Tech’s Hype and Create the Future We Want." Register today: bit.ly/4sYGeWU Featuring @emilymbender.bsky.social + @alexhanna.bsky.social and hosted by @histoftech.bsky.social

Blood on our hands. www.nytimes.com/interactive/...

There is no McTry.

I wouldn’t say I’m a luddite, but, even working in IT, I struggle to find a use-case in which AI performs adequately. And, remember kids, it’s a *terrible* search engine

why do I find it so wildly offensive and paternalistic when I see folks suggest that LLM use can help 'disadvantaged' people to 'write better' or 'save time'? because it is such a sad underestimation of people's abilities, a surrender to a thoughtless monoglot ideology, a celebration of ignorance

I believe the frame that minoritized scholars are helped by unethical tech is a trap to avoid. We better fight the system of inequities ourselves & not buy into industry frames trying to sell their discriminatory, polluting, and exploitative tech as "inclusion". bsky.app/profile/iris...

I know AI psychosis isn't an official DSM diagnosis (yet), but from what we've seen so far, this man is getting way high on his own supply.

“Feel free to adapt and reuse. No attribution needed (…) If you make an adaption that you also want to share, I’d love to know about it, too.” irisvanrooijcogsci.com/2026/04/08/r...

🚨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/

Once again: this is exactly what LLMs and chatbots are designed for. This is them doing their designed function. Mass production of meaningless text so someone who wants to do harm has cheap access to the aesthetics of justification www.theguardian.com/environment/...

Impossible to disagree with a single word that Ben Rhodes (Obama era NSC official) is saying here:

The matter-of-fact way she asks what investors should do if Trump destroys Iran. Absolutely wild. "Let's talk about tonight, this deadline that President Trump has set 8 p.m., has threatened to destroy a civilization. How does an investor process that? Is it a bigger upside risk or downside risk?"

I'm not sure how you can be against removing this man from office right now. What will it take? A nuke being dropped? It's too late then. The purpose of impeachment and the 25th Amendment is to prevent that type of thing from happening, not just to react to it.

No matter how you feel about cocaine, its use is growing more widespread and it's important that educators understand how they could integrate it as a tool in their classrooms.

hear hear!

Microsoft CoPilot isn't meant for real work. Even Microsoft says so. Any university that pushes it as an analysis tool needs to wake up.

That's totally disrespectful. Imagine you worked hard your entire life to create a company that manufactures trash cans only to have someone throw Pam Bondi's face in it.

Two summer schools focused on skills for cognitive modeling and mathematical psychology will each receive $20,000 grants through this year’s William K. & Katherine W. Estes Fund. @psychonomicsociety.bsky.social Learn more about these projects by @peterkvam.bsky.social & Eric-Jan Wagenmakers.

OMG

When it comes to writing, the process is the purpose. You don’t have to always like or enjoy the process, but if you don’t respect it enough to do it yourself, there is no purpose. AI will never fill that void. www.thehandbasket.co/p/refusing-t...

I seriously think this is one of the biggest disconnects. LLMs constantly and confidently produce misinformation, and it requires enormous amounts of vigilance to work with them under high standards of rigor. If people don't care about facts or rigor, of course they'll think this is miracle tech.

Wait wait wait, *this* Stephen Marche?

This Administration is spending billions on a war with no strategy instead of lowering people’s health care and child care costs.

I regret to inform you that CNN now has the article claiming that datacenters put out more heat energy than a nuclear bomb.

hahaha — and I cannot stress this enough: LMAO

"A massive seven-year project exploring 3,900 social-science papers has ended with a disturbing finding: researchers could replicate the results of only half of the studies that they tested" www.nature.com/articles/d41...

guardrails are a scam