Rayo Verweij

Philip Pullman said that anything that a child might encounter in their childhood is appropriate material for a children's story.

The Edinburgh Castle bathed in a golden sunset as seen from the cemetery below, with a ranbow curving out from the Castle

Still pretty

A picture of a foggy street

Edinburgh castle failing to render, likely thanks to the AWS outage

Delighted to be at @edfuturesinstitute.bsky.social today with our #DesigningResponsibleNLPCDT sharing our brilliant students work as part of Doors Open Day!

A free poster boards displaying children’s drawings behind a table with colouring materials

I have a little stall at the @edfuturesinstitute.bsky.social Doors Open Day today! I’m sharing some early research around children’s views of AI at school. Come say hi if you’re around!!

Sex is getting scrubbed from the internet, but a billionaire can sell you AI nudes Grok’s nonconsensual deepfakes show how hollow the online safety push can be.

"In the new ... internet, girls can’t learn about periods on Reddit and artists can’t sell smutty games on Itch.io, but a military contractor will make you nonconsensual deepfakes of Taylor Swift taking her top off for $30 a month." — @thedextriarchy.bsky.social www.theverge.com/internet-cen...

Earth is spinning faster, leading timekeepers to consider an unprecedented move | CNN Earth is spinning faster this summer, making the days marginally shorter and attracting the attention of scientists and timekeepers. But these days are not ‘short’ in terms of how many sunlight hours ...

If you've been feeling tired recently, it's because Earth is spinning faster, making some days ~1 millisecond shorter than 24 hours so you're getting less sleep. 😴

i genuinely think that part of the problem is people don’t have the mental model for how the “chat bubbles” map to autocomplete

it doesn’t help that the chat interfaces actively hide what’s going on and make the illusion opaque

Rayo standing in front of a projector screen with a slide that says “Children as Creators of Real-World AI-Powered Tools for Education”

Yay I’m at @acm-idc.bsky.social in Reykjavík this week! I spoke about my PhD plans at the Doctoral Consortium and am so excited to get to know everyone this week :)

It was absolutely wonderful to work with the Digital Leaders at St Francis Xavier’s Primary School. Thanks again for having me!

drawing from a child with bold, colourful text that ready 'AI should respect our privacy'

📣 “𝗖𝗵𝗶𝗹𝗱𝗿𝗲𝗻 𝗻𝗲𝗲𝗱 𝘁𝗼 𝗸𝗻𝗼𝘄.”

That’s the message we heard loud and clear from Members of the Children’s Parliament.

𝗠𝗮𝗻𝘆 𝗰𝗵𝗶𝗹𝗱𝗿𝗲𝗻 𝗵𝗮𝘃𝗲 𝘃𝗼𝗶𝗰𝗲𝗱 𝗰𝗼𝗻𝗰𝗲𝗿𝗻𝘀 𝗮𝗯𝗼𝘂𝘁 𝗵𝗼𝘄 𝗔𝗜 𝗶𝘀 𝗮𝗳𝗳𝗲𝗰𝘁𝗶𝗻𝗴 𝘁𝗵𝗲𝗶𝗿 𝘀𝗮𝗳𝗲𝘁𝘆, 𝗽𝗿𝗶𝘃𝗮𝗰𝘆, 𝗮𝗻𝗱 𝗿𝗶𝗴𝗵𝘁𝘀. One key worry? They don’t know what’s happening with their data and they want answers.

European Citizens' Initiative Give your support !

The petition to ban conversion therapy practices in the EU is lacking enough votes from:
- #germany (49.9%)
- #sweden (49.96%)
- #Netherlands (94%)
- #austria (25.5%)
- #Slovenia (83.14%)
- #portugal (24.46%)

If you're a citizen there you can help advance a ban on conversion therapy by voting:

Thankfully people have realised that wasn’t enough and actual programming is part of the curriculum again. Would be great if we could not make the same mistakes again when it comes to AI!

There was a similar push in the 2000s where computing lessons in schools shifted from teaching basic programming to teaching MS Office as that was deemed to be more important. And instead of teaching something fundamental about how computers work it became about prepping kids for office jobs

Co-opting the term literacy for corporate AI lessons is particularly devious because it’s hard to argue against. Everyone thinks literacy is important! But in the chaos of schools trying to adapt it’s easy for the actual meanings of these words to get confused

If literacy is the ability to effectively understand and interpret text, to me, AI literacy is understanding the limits of AI output, knowing why LLMs can hallucinate, being aware of the risks of model biases, etc etc. Not “how to use AI to offload your thinking”

I get so frustrated with the dilution of the term ‘literacy’ - in no world should ‘AI literacy’ mean ‘prompt engineering classes’ but apparently here we are

Join us at AI for All during Edinburgh Science Festival! Discover how we make #NLP shape our daily lives in a fair, accountable and transparent way. Explore exhibits, activities, and meet the next generation of researchers.

📅 Fri 11 Apr
⏰ 10am – 4pm|Free / Drop-In
📍@edfuturesinstitute.bsky.social

For example, CodeCademy makes it super difficult nowadays to get to their free courses. And on the other side the new RPF code editor doesn’t let you add external links (anchor tags) to HTML files because “safety”

Okay enough ranting for tonight. What? You thought I was about to offer that solution? Haha very funny vier april kikker in je bil

Look this is hardly the biggest problem in the world but every day I see so many people write about the dangers of genAI in education and like I hear you and probably agree with you 100% but at a certain point you gotta offer an actual alternative solution

They are genuinely brilliant, got everything they could out of Scratch and are hungry to learn more complex stuff. But intro to coding resources are either aimed at kids and super restrictive/linear or aimed at adult learners and, well, also restrictive plus you need a credit card

This is so cool. Have been wanting to get into PoE properly for so long now and I think this might actually do it

A really fun reflection, I thought, not just on AI but on the purpose and origins of language in general

“AGI” Is Impossible Either That or Defining It Is Entirely Arbitrary

“We are not in fact in danger of the machines overtaking us in what we do best. What we do best is to mess with mom, and to whisper sweet nothings to our lovers. But we are in grave danger, at present, of misidentifying what we do best”

There are *good*, positive use of GenAI. But forcing it on people -- particularly as a massive technology corporation with its fingers in many aspects of billions of peoples' lives -- creates more risk than potential benefit. Let people opt-in as they see fit, don't force it down their throat.

The image is from a "Transparency Center" document and lists guidelines regarding acceptable and prohibited content for insults. It mentions: 1. Insults about: Character, such as cowardice, dishonesty, criminality, and sexual promiscuity or immorality. Mental characteristics, including but not limited to accusations of stupidity, intellectual capacity, and mental illness, as well as unsupported comparisons among politically correct (PC) groups based on inherent intellectual traits. 2. Highlighted section: The document allows allegations of mental illness or abnormality when tied to gender or sexual orientation, referencing political and religious discourse about transgenderism and homosexuality. It also acknowledges the non-serious use of terms like "weird."

Meta literally created a LGBTQ exception for calling someone mentally ill as an insult. You can't do it for any other group except LGBTQ people.

They must've calculated the optimal way to space out these words. Like, at the exact spots people are most likely to stop using the app they drop one of these to lull users into a false sense of "oh, I guess I *can* learn this language!"

It’s cool that humans have been eating bread so long that when we touch a loaf for a moment we can sense ah this is going to be awful bread. Like badgers sensing vibrations underground through their claws.

Beat BG3 in Honour Mode last night and honestly this might be where 2025 peaks for me