I could go on.

In all these cases, I say “could” — not “should,” not “will.” I have nuanced and widely varying personal opinions on which of all these possibilities are probable, and which are desirable.

I’m just pointing out there we do in fact have choices, and the very specific trajectories AI vendors lay out are not at all inevitable.

A/end

@cstross Touché 😊 A river delta probably (that is the 'os' part of the name), but not so much a swamp. I suppose it is hard to blame the original builders of Bangkok, but understanding of the regional water mechanics should have been there for a while now. Willingness to do the right thing about it is probably not much there, as most such places in the world.

@catsalad
False marketing. It's clearly in a case of unbroken glass.

Postmortem of a little community hobby wiki struggling to survive an extinction-event-tier DDOS purely because they banned one guy for using Claude on the wiki https://blog.xkeeper.net/the-cutting-room-floor/tcrf-2026-ddos-postmortem/

@lowqualityfacts I'd like one star that says "MOSTLY"

When your job as a consultant on "cats eyes" turns out to be a hands-on role ...

I’m glad they’re putting this in the spotlight, really glad, but Mother Jones absolutely biffed something here, something a lot of folks are getting wrong:

The quote about how AI is “the largest theft of labor in human history?” That’s not “how tech leaders talk when they think no one is listening.”

No no no. •That• is the kind of email you send when you think the company is doing something terrible, and you want your objections preserved in writing. 🧵

https://www.motherjones.com/politics/2026/09/openai-chatgpt-microsoft-copyright-legal-case-documents-revelations/

@matteopzz @diggita @thamichele @linux La frammentazione fa male anche al discorso disabilità. L'accessibilità all'informatica per ciechi, ipovedenti, disabili motori e (soprattutto) cognitivi, ha bisogno di standard rigidi. Ora non entro nel tecnico, ma a livello comunicativo pare un paradosso "l'inclusione ha bisogno di regole" perché includere vuol dire guardare qualsiasi eccezione che ci sia a questo mondo. Il che è per definizione utopico ma si cerca almeno di coprire più esigenze possibili. Linux non lo fa completamente, e questo rende difficile avvicinarsi a chi il computer non lo usa (solo) per produttività e svago, ma è una vera e propria "estensione" del corpo, insomma una protesi. E la comunità di sviluppo Linux, almeno in molte sue parti, ha la tendenza di dire "se hai bisogno di qualcosa, fattelo tu" con una narrazione "volere è potere" che non potrà mai coprire il background culturale, sociale, economico, di tutti.
L'altro giorno (lo trovate nel mio profilo) ho avuto in mano un tale, Robin Kipp, cieco, che è riuscito a rendere parlante un FlipperZero, uno di quei dispositivi hardware che servono a far funzionare telecomandi e cose strane. Però Robin ha, probabilmente, un background ingegneristico di tutto rispetto. Che io non ho.
Io, sto imparando ad "aggiustarmi le cose" grazie al codice assistito, insomma quello che il marketing chiama "vibe coding" come fosse un pene a pile. Rendo l'idea? Però fin che è un plugin WordPress gioco anche; ma non azzarderei mai a proporlo al pubblico né tanto meno a mettere le zampe su un materiale delicato come Linux. Rischierei solo di indispettire gli sviluppatori ancora di più e rendergli la questione accessibilità ancora più antipatica di quello che già gli è.
Spero di poter incontrare di persona qualcuno con cui parlarne al sfscon a Bolzano il prossimo novembre, così da scambiarci idee davvero davanti a un tavolo.

My quibble matters for two reasons:

1. The way that the AI-pushing tech leaders like Sam Altman and Elon Musk and Satya Nadella actually talk when they think no one is listening is AFAICT completely different from that Hecht quote. They are delusional, megalomaniacal. They wouldn’t gloat about getting away with labor theft; the concept of “labor theft” does not even exist for them. How can it be theft when they deserved it, when they are messianic figures whose wealth is proof of their infallible rightness?

If we want to oppose this destructive AI bubble, we should understand what we’re opposing.

2/

@farah I know 😩

The #ICanHazPDF hash-tag is active on the Fediverse.

I remember that hash-tag (ICanHazPDF) starting in (I think) the year 2011.

...

A lot of scholarly papers are paywalled. ICanHazPDF is a way for those who don't have "academic access" to ask for a copy of the paper.

...

ICanHazPDF was a play on the "I Can Has Cheezburger" (ICHC) meme.

#ICanHasCheezBurger #ICHC #ScienceFedi

Podcasting used to be a hobby for most people it's quickly becoming a business for many.

I talk with Claricast's Joe Eftihiou about this on this Tuesdays Entrepreneur's Enigma podcast.

#podcasting #podcastingBusiness

One of the cool things of the RC2014 Assembly 2.0 was the official badge which in itself is just the PCB with the LCD printed on the silkscren so you could write your name there with a pen - but you could also get an upgrade kit to turn it into a real LCD module. Which I did, courtesy of @rc2014 - and I just finished an early version of a driver for it to be used with #dxforth - lcd-clr, lcd-xy and lcd-print are already working as the photo shows.

#rc2014
#rc2014Assembly

It’s been 7 months running my own #fedify powered #ActivityPub server
There are tons of improvements to make, if not an entire refactoring of the current codebase.

It works but there is quite a few quirks that make it unfit for general usage, unless you want to experiment :)

🔗 https://rmendes.net/notes/2026/09/26/a2fb7

@rick that's so cool!

@Soyweiser @techtakes I am currently thinking that time travellers would be better employed if they took ship to South Africa to kill babies Musk and Thiel.