By continuing to report the Gaza territorial development plan as some sort of "peace" plan, the actual fate and livelihoods of the humans in Gaza is being marginalised. This is not a Board of Peace, it is a Board of Business Opportunity. www.aljazeera.com/news/liveblo...

Everything is unbelievably stupid right now.

“It’s basically a map of the United States. It’s kind of like Google Maps.” This is how an ICE agent described the app Palantir worked on for finding neighborhoods to raid. I've seen the interface, it does look exactly like that www.404media.co/elite-the-pa...

"Complying to UK law" still allows you to create the same type of imagery. I was easily able to make a prompt from AI generated picture of a full dressed woman I uploaded to Grok to a woman bound, nearly fully naked, and bleeding in about 2 minutes, and I'm in the UK.

I've been working on research to inform policy spheres on how recent procurement strategies meant AI models were not being appropriately assessed for defence use (see link), but I think at this point it's obvious and they don't even bother to lie about it anymore. defensescoop.com/2025/08/04/d...

I've been working on research to inform policy spheres on how recent procurement strategies meant AI models were not being appropriately assessed for defence use (see link), but I think at this point it's obvious and they don't even bother to lie about it anymore. defensescoop.com/2025/08/04/d...

I can’t emphasize enough how shocked my out of state friends and family are when I tell them what has happened in Minneapolis since Renee Good’s murder. Without reliable media, most not-chronically-online people are completely in the dark. If you want to help Minnesota, tell people what you know.

Or this is what's typically called "no regulation"

They’re opposed to preventative healthcare policy because they believe they’re an intervention in the natural eugenic hierarchy, framing this as an issue of “government overreach” while they’re literally banning healthcare for disfavored groups assists in that effort.

well well well

Defence minister says Somalia has ‘confirmed information’ that Israel plans to forcibly move Palestinians to Somaliland.

Legal issues aside, several implications: 1) They're aware testing data is in training data and public evals are largely meaningless 2) Once this data is inevitably absorbed into training, they can provide the illusion that models are more capable as most will assume IP isn't in training

Legal issues aside, several implications: 1) They're aware testing data is in training data and public evals are largely meaningless 2) Once this data is inevitably absorbed into training, they can provide the illusion that models are more capable as most will assume IP isn't in training

NEW: Some verified users on X are still using Grok to create “undress” images. And Grok’s app and website will still create images for people who don't have paid accounts. “The recent decision to restrict access to paying subscribers is not only inadequate—it represents the monetization of abuse

Trump’s rush to fast-track new nuclear reactors raises serious safety concerns. @heidykhlaaf.bsky.social to @npr.org: "These manufactured timelines are incredibly concerning. There's no timeline for assessing a new design and making sure it's safe." Read more in @npr.org: shorturl.at/y78eg

📣 NEW w/@udbhav-tiwari.bsky.social Mapping the technical reality & privacy/security perils of pushing AI agents into our infra We offer palliatives, but the core issues are paradigmatic: 'agency' relies on pervasive data access + ability to act w/o explicit consent www.youtube.com/watch?v=0ANE...

This take completely dismisses the non-white citizens who have been shot at (e.g., Marimar Martinez) and detained (dozens), who have not received a fraction of this outrage. This is emblematic of the issue of who Americans really think is a "citizen".

Audacious and dangerous for the company that has led to numerous suicides claim that it's "not intended for diagnosis or treatment" but that it can’t control how people end up using AI. With this + Grok, we're seeing AI companies user-blaming and abdicating responsibility of harmful model outputs.

So many of us have been dismissed for years when we challenged AGI, despite our deep expertise. But as they now pivot away from their failed AGI 2025 premonitions, it brings into question why they continue to be relied on as narrators rather than those who accurately predicted these outcome.

In retaliation for a solidarity event for Palestinian political prisoners in Israel and a screening of the film “Hind Rajab.” For this they shot students.

this is how every outlet should refer to X

The last family in the Palestinian village of Yanoun left their home last week, joining a growing list of communities that have been erased from existence through the establishment of Israeli "shepherding outposts" in their place. mondoweiss.net/2026/01/this... #Palestine #Israel

Then reverse the Online Safety Act. I can't access my Bsky messages and many subreddits without invasive and insecure age verification, but there's no limitations to X itself actually producing these images?

This is a line that has already been crossed. When Israel, with the blessing of the US, attacked multiple sovereign states (Lebanon, Syria, Qatar) in an attempt to achieve goals they could not through their genocide of Palestinians, that's when we normalized the disposal of all international law.

Insider Trading as a Service™️

Also, when reporters start talking about the chatbots this way, it is imperative to remind them they are basically asking a toaster for comment. They should be a lot more embarrassed about this than they currently are. “Grok told CNN…” no it didn’t. You typed stuff into an unreliable calculator.

Are we going to talk about how insidious it is that X and the media are pulling the "Grok says" line to remove accountability from the product owners and developers themselves?! Grok does not have autonomy or agency. www.reuters.com/legal/litiga...

This was also the end game of co-opting the term "safety" to focus on hypothetical and unscientific AGI claims. It diverted attention from the real harms that are inherent to genAI and its training data, and now society has sleep walked into this horror with no accountability in sight.

We're in a great place in society when Doctors Without Borders are being labeled as "terrorists" for trying to aid a population under genocide, and media just nods along in agreement.

An essential development over the past decade-plus is that the ignorant went from feeling chastened when their ignorance was revealed to banding together and deciding that, actually, they were right. They might disagree at times, but they agreed that ignorance itself was an indicator of trust.

Winter floods wreak havoc on Gaza displacement camps as Israel blocks aid https://aje.io/nu8pwm

A significant amount of safety research has existed for decades on Human Machine Interfaces proving exactly this. It's exhausting to watch money and effort wasted in a tech hype cycle to constantly undo and redo what we already know in safety engineering.

So not only are energy industries deregulating for AI, risking catastrophic accidents, but they're accelerating the use of the most polluting energy sources during a climate collapse. This is a death spiral. www.ft.com/content/8deb...

A significant amount of safety research has existed for decades on Human Machine Interfaces proving exactly this. It's exhausting to watch money and effort wasted in a tech hype cycle to constantly undo and redo what we already know in safety engineering.

It's becoming increasingly difficult to fight the desire to become a hermit in the woods.

Israel’s public broadcaster Kan airs footage showing the total destruction of Jabalia Refugee Camp in northern Gaza: Video: Mosab Abu Toha via X (@MosabAbuToha)

"Salesforce execs delivered a diff message to customers: Agentforce sometimes works better when it doesn’t rely so much on LLMs, otherwise known as genAI. Salesforce has been using rudimentary, 'deterministic' automation in Agentforce to improve reliability" www.theinformation.com/articles/sal...

Surprise, surprise

BREAKING: Israel’s Cabinet approves proposal for 19 new settlements in occupied West Bank. This is Israel breaking the ceasefire, violating international law, & forcibly & violently stealing land—all while having killed more than 1000 Palestinian civilians in the West Bank over last 2 years. 1/2

This is the most propagandised way to frame illiteracy as a good thing, and it's precisely the reason why AI pilots are failing and benchmarks for them are completely meaningless: they don't know how anything works. Terrifying when we're seeing a push for AI in nuclear and defense.

It's become completely acceptable to be openly racist towards Arabs/Muslims/Latinx with no repercussions. Tech and VC companies and their partners have amassed so much power that they do not believe there are any consequences to their actions or the systems they provide which bolster this racism.

[“When you see people marching and showing violent bloody images, images of death and destruction, it’s unleashing something in our community that the organisers of the protest can’t contain,” Minns said.] More upset about the protests than about the death and destruction.

I spoke to Geoff at @npr.org about the US DoE Nuclear Reactor Pilot Program, and why it's dangerous to time box safety assessments, as we discuss in our recent nuclear report.

And another one by the @NYT ‘editorial board’ that sounds precisely like a defence tech industry brochure or op Ed. I’m too familiar with the narratives to ignore the striking semblance

NEW: I looked into MAGA darling Anduril's latest foreign partner, and its weapons are accused of fueling the ongoing genocide in Sudan. theintercept.com/2025/12/11/a...

Israel used Palantir technology in its 2024 Lebanon pager attack, book claims - https://www.middleeasteye.net/news/israel-used-palantir-its-2024-lebanon-pager-attack

From our paper re: US defense LLM rollout.

People really lack the capacity to understand that Israel using Palestine as a lab for lethal AI surveillance means that the rest of world is the intended target for prime deployment. This is the plan for our future, one we have vehemently oppose. nymag.com/intelligence...

The EBU dropped plans for a vote on excluding Israel because of the ceasefire This is the point of the pretend ceasefire. Not to stop the genocide but to allow countries & organisations to more easily resist calls to put pressure on Israel & to reverse actions already taken

A presentation at the International Atomic Energy Agency unveiled Big Tech’s vision of an AI and nuclear fueled future.