@poritzj

146 Followers
287 Following
336 Posts
Trying to work to make the world a little better in the ways I can; mostly in the open education and FLOSS movements at the moment. More at https://poritz.net/jonathan
Websitehttps://poritz.net/jonathan

I realized I hadn't updated in several years my post https://mathematikoi.net/p0/index.html on the textbook pain multiplier. Still haven't, but at least I've made the new graphs, the basic one I attach here. It shows the crazy cost increases of textbooks in the US from 1980 to 2025 hit a bit of a wobble 2017-2024 due to #OER, I assume, but then jumped way up again recently.

Note some recent BLS data is missing, I guess from US federal govt chaos.

#OpenEd #openeducation

Number of AI chatbots ignoring human instructions increasing, study says https://www.theguardian.com/technology/2026/mar/27/number-of-ai-chatbots-ignoring-human-instructions-increasing-study-says?CMP=Share_iOSApp_Other

Ok, so this is a problem—but not for the reason the headline claims. The AI agents are not «ignoring instructions», but are following instructions and because they are programmed to act semi-independently without checking for new instructions every step along the way, the consequences of the instructions they are given are well beyond what is foreseen. We have had fables warning us about this for thousands of years—think of King Midas not being able to forsee the consequences of everything he touches being turnes to gold—but we stubbornly refuse to listen to our own warnings.

Number of AI chatbots ignoring human instructions increasing, study says

Exclusive: Research finds sharp rise in models evading safeguards and destroying emails without permission

The Guardian

Thanks so much for all the suggestions here!

I actually don't have access to the bike for the next 16 hours, so I can't follow up immediately. But tomorrow AM I'll check and post what I find, in case anyone is curious.

Any #brompton experts here?

I was folding my (P line) Brompton the other day and heard a small "ping" sound. Then I saw this small, black, anodized, Phillips-head screw on the ground - but I can't find where it fell off from!

Anyone have any ideas?

The bike still works fine, at the moment: it gets me around the streets of Pisa quickly and comfortably, easily on the train even when crowded, even up&down the 4km long, 350m high hill near my house -- a reliable companion!

I just unsubscribed from the newsletter of @dajb .

Recently I've felt that his writing has more followed the establishment's talking points -- e.g., I criticized a post of his which he said we have to update our mental models rather than, as I insisted, keep fighting for justice.

It turns out genAI is a big part of his writing process now: maybe that's the difference I noticed!

I have better things to do with my time than read AI slop, even when a nice man edits and types it up for the web.

RE: https://scholar.social/@olivia/116229013687072877

Unsurprisingly, from the first sentence it's a lie:

"In the five years since Anthropic began, AI progress has moved incredibly quickly."

No, actually it hasn't. We're still using transformers, developed in late 2017; in fact, we're **hopelessly dependent upon them with nothing else in sight.**

Progress has stalled, despite endless hype. For example, where are those "agents?" Where are they? Can someone, anyone point to a successful, profitable agent? Anyone?

This is all such a terrible joke.

There was an old country that swallowed some wood;
We should recall it had loads of wood!
It swallowed the wood to catch the coal,
It swallowed the coal to catch the steel,
It swallowed the steel to catch the oil,
It swallowed the oil to catch more coal,
It swallowed more coal to create The Cloud,
That squiggled and jiggled and tickled its crowds.

It swallowed The Cloud to catch AI;
I don’t know why it swallowed AI – perhaps it’ll die

#energyTransition #lyricallyDissAI

“I’m just so exhausted by both sides of the AI argument” imagine being the kind of piss baby that sees a mass of people who are righteously furious at a system designed from the top down to fuck them over, and decides the anger is just as much of a problem as the exploitation

RE: https://social.coop/@cwebber/116217477822586442

The whole "AI" industry is just based on the hope that they can deskill people fast enough that they'll have to rent back cognitive support systems. Forever.

It's like Uber. Just that they don't try to break existing transportation infrastructures but your brain.

Reup: Marco Rubio Will Let 2.5 Million Children Die so Sarah Rogers Can Fund Far Right Extremists in Europe emptywheel.net/marco-rubio-...

Marco Rubio Will Let 2.5 Milli...
Marco Rubio Will Let 2.5 Million Children Die so Sarah Rogers Can Fund Far Right Extremists in Europe - emptywheel

The same week we got new figures of how many deaths -- 9.4 million, of which 2.5 million would be young children -- Marco Rubio's destruction of US foreign aid may cause, we got new details on the slush fund for far right extremists in Europe that Rubio will spend instead.

emptywheel