ICYMI: Microsoft’s charge “implies a more than $12 billion quarterly loss at OpenAI, said Firoz Valliji, an analyst at Bernstein.” That “would mark one of the largest single-quarter losses for a tech company in history.” @jessefelder.bsky.social $MSFT www.wsj.com/livecoverage...

Poi quando dici che le norme fatte per il contrasto al terrorismo (o alla pedopornografia, o altri reati esecrabili che sembrano meritare misure eccezionali) finiranno per essere usate anche per cose ordinarie, ti dicono che è la fallacia della china scivolosa… milano.repubblica.it/cronaca/2025...

Giusto in tempo per passare a occuparci di filtri sull’età e altre idee geniali nostrane

The games business is about games. If you’re in the games business, and somehow (as often happens) your company somehow is now focused on something else (AI, XR, “IP,” acquisitions, etc.) you’re in for a period of struggle. Everyone will wonder what’s wrong. But you know. It’s about games. Gameplay.

The same websites that spent the last few years becoming an unreadable mess of pop ups and banners are now complaining that users will do literally anything including installing apps or believing AI slop summaries to avoid having to browse through them.

Quelli di AGCOM, non contenti degli incredibili successi di Piracy Shield, hanno pubblicato una lista di 48 siti porno (gli piace proprio fare delle liste di siti da bloccare), che dal 12 novembre dovranno verificare l’età degli utenti. Hanno detto come? Pensato a un’infrastruttura? Ovviamente no.

TL;DR: Italy's Piracy Shield, an automated system for addressing streaming copyright infringement, continues to face serious issues as AGCOM extends its operation without resolving the existing problems.

Italian Senator left Jeffrey Sachs literally speechless and shocked by calling him a liar and a propagandist. It happened on Prime Time Italian television. Priceless 🤩 [ENG SUB] (The 2nd part in the next tweet)

Reminder: one of their *own investors* is amongst those who has lost touch with reality as a result of their product. And they still haven’t slowed down.

I know if you follow my feed you might mistake me for being an anti-AI Luddite, but you would be mistaken. My initial career path was AI. In the early 2000s. My PhD was on applications of ML to security, and if you take most of my publication track you will find applications of AI to cybersecurity.

Click through the link and watch the slop video. I mean it, watch it. Whenever you hear a tech bro saying “our AI system can do <whatever>”, this is their standard. This dude thinks this is an amazing videogame.

LLMs are not a labor replacement technology, we've seen this come out of every study on their use in business contexts. They've been falsely marketed as such, though, and therein lies the problem for all of us, from job seekers to the cos. pushing the tech as it causes myriad internal issues.

Remember this can also have some long-term consequences that are really serious. Certain diagnosis or medications can have implications on future employment as well as the ability to hold certain certifications and licenses. Getting your medical record corrected after this will be an nightmare

🤦‍♂️

We are deeply saddened by the loss of Greg Newby, who led Project Gutenberg ( @gutenberg.org ) with passion and purpose. Greg’s belief in free and open access to knowledge continues to inspire us and so many others working to preserve our shared culture online. gutenberg.org/about/newby....

Mail di MooneyGo dopo violazione dei dati. Di nuovo si consiglia di cambiare regolarmente le password. Non serve. Non aumenta il livello di sicurezza. Usate un password manager. Password lunghe e complesse. Password diverse per ogni servizio. Le cambiate solo se sono violate.

The internet was on fire. 🔥 One small library affecting billions of systems. Log4Shell was the biggest security vulnerability of all time. Now, Log4J maintainer, Christian Grobmeier tells us what it felt like inside the flames 👉 github.blog/open-source/...

PSA: we're aware that Signal is down for some people. This appears to be related to a major AWS outage. Stand by.

You know how billionaires end up with severe cognitive deficits as a result of becoming surrounded by yes men who constantly tell them their every idea is genius? What if we made a bot that just does that to everyone. I think that would be a good idea.

Jesus christ. AI coding platform Augment code had to jack up prices because 22.5% of their users were spending 20x of what they paid, and even after raising the prices, they're still running at a loss. reddit.com/r/AugmentCod...

Apple announces a "major evolution" of its Security Bounty program, doubling the top award to $2M for exploit chains that could be abused for spyware attacks (Lily Hay Newman/Wired) Main Link | Techmeme Permalink

It’s amusing to see people that clearly have no technical understanding of how LLMs work pontificate on how they work. They’ve clearly picked up some of the appropriate language/concepts, but it’s uncanny valley level stuff where their use is off in subtle ways that expose that they have no clue.

Wat

We're so relieved to see Germany reaffirm its opposition to the dangerous Chat Control proposal--the one that would mandate mass scanning of communications. Germany's long been a solid champion of privacy, and the news that it was considering backing mass surveillance was alarming. 1/

“Concerns over an AI bubble bursting have grown lately, with analysts recently finding that it’s 17 times the size of the dotcom-era bubble and four times bigger than the 2008 financial crisis.” Hang onto your butts. This “correction” is gonna hurt. futurism.com/artificial-i...

I have endless circular arguments about this. Them: “It’s written a summary.” Me: “No, it’s written something tuned to look like a summary.” Them: “But it looks like a summary.” Me: <sigh> We’re not used to computers lying to us.

Germany has agreed to stop ChatControl for now, due to huge amounts of public pressure. Good job! The bad news is that it could come back as soon as December, and the German government has interpreted the feedback as a need to “moderate” the proposal.

“Use this technology that has been proven to make your work slower overall to be 5 times faster in building the other technology that no one actually wanted and everyone just pretended never happened”

I’m happy to announce a new #preprint! 🧑‍💻📝🎉 Quantum states often show up with hidden structure. What if a state is built from just a few elements of a larger, #non-orthogonal, #overcomplete dictionary? Can we exploit that sparsity to beat standard #tomography costs? 🧵⬇️ /n

No. It’s most definitely not. This is ridiculous. Absolutely ridiculous.

jesus christ

Memories.

Today, my research group @ Georgia Tech released a paper on vulnerabilities in Tile --- the second largest device finding network after Apple's AirTags. You can read about it in Wired, reporting by @kimzetter.bsky.social! www.wired.com/story/tile-t...

Equating “AI” with the output of LLMs may easily be the death blow to the discipline. What’s coming is not merely a winter, it will be a nuclear wasteland. Hope the money grift was really worth it folks!

Who could have possibly predicted this, except any engineer even superficially trained in robotics?

Very happy to see a full room at our #RecSys2025 tutorial "Hands on #QuantumComputing for Recommender Systems" presented by me and Paolo Cremonesi. We covered gate model QC and #quantumannealing with examples on community detection and feature selection. Materials: qclef.dei.unipd.it #recsys

They also likely used a computer.

OpenAI “admits” what literally anyone with any training in the subject always knew. Generative AI models work like this. They are built like this. “Hallucinations” is all they do all the time.

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

Who could have ever predicted this?! Except any engineer with even a basic understanding of robotics, I mean.

Io non so nemmeno come commentare. La Serie A (quelli di Piracy Shield, ricordate? Quelli disposti a bloccare mezza Internet pur di "fermare la pirateria") ha deciso di... trasmettere su YouTube le partite, gratis, in un certo numero di nazioni: www.reddit.com/r/singapore/...

ROTFL alignmentalignment.ai

Here we go again 🙄 we have written an open letter detailing how incredibly misguided is the ChatControl proposal: csa-scientist-open-letter.org/Sep2025

NEW: X has started rolling out it's end-to-end encrypted chat for all users. I was invited to start using it yesterday, so I asked a couple of experts if users should trust it. The answer: no, nobody should trust it at this point. There are several reasons why, which they explained to me here.

Silently thanking the travel gods that I didn’t choose to connect through JFK today…

“The depictions in the media and from clinicians suggest that people are mainly experiencing delusions, and for that reason, clinicians prefer ‘AI-mediated delusions’ rather than the snappier but less accurate ‘AI psychosis’ to describe the phenomenon.”