US and Israeli forces carried out a missile attack on Gandhi Hospital in Tehran, Pars Today reports.

According to the outlet, nurses at Gandhi Hospital are evacuating newborns.

Earlier, Iranian media reported that American-Israeli forces struck the headquarters of IRIB television in Tehran.

cross-posted from: https://infosec.pub/post/42793761

Bloody backpack of Iranian girl killed in US-Israeli attack on Iranian elementary school

Source

Reposting here after the original post was removed from !pics@lemmy.world

Source

Possible original picture

Reposting here after the original post was removed from !pics@lemmy.world

Anthropic, the wildly successful AI company that has cast itself as the most safety-conscious of the top research labs, is dropping the central pledge of its flagship safety policy, company officials tell TIME.

In 2023, Anthropic committed to never train an AI system unless it could guarantee in advance that the company’s safety measures were adequate. For years, its leaders touted that promise—the central pillar of their Responsible Scaling Policy (RSP)—as evidence that they are a responsible company that would withstand market incentives to rush to develop a potentially dangerous technology.

But in recent months the company decided to radically overhaul the RSP. That decision included scrapping the promise to not release AI models if Anthropic can’t guarantee proper risk mitigations in advance.

Tehran shows no signs of buckling, even as the death of the supreme ruler creates a major power vacuum in the country.

Iran is vowing vengeance against the U.S. and Israel and signaling to the world that its government is not about to collapse, despite airstrikes Saturday that killed Iranian Supreme Leader Ayatollah Ali Khamenei and a cadre of other senior Iranian leaders.

Into Sunday, Iranian forces launched missiles and drones at targets throughout the Middle East, including against U.S. bases in the region and Israel. Some of those targets were apprehended by British fighter jets, U.K. Defense Minister John Healey said Sunday. Kuwait’s military also said it intercepted many of the Iranian missiles fired at its airspace on Sunday.

Many, however, have hit their targets and casualties mount. Kuwait’s health ministry announced Sunday that Iran’s latest volley of strikes killed one person and wounded 20 others. A strike against targets in central Israel killed six people on Sunday. The death toll in the United Arab Emirates from Iranian attacks rose to three on Sunday.

MBFC
Archive

Saudi Crown Prince Mohammed bin Salman made multiple private phone calls to Donald Trump last month, urging him to launch military action against Iran, even as he publicly favoured diplomacy, according to a report by The Washington Post.

The report said the crown prince pressed for a strike despite earlier public statements in January in which he insisted Saudi Arabia would not allow its airspace or territory to be used for attacks on Iran. At the time, he called for dialogue between Washington and Tehran and said Riyadh respected Iran’s sovereignty.

The US and Israel went on to carry out joint strikes targeting Iranian military and government sites after nuclear talks stalled and amid claims that Tehran had resumed aspects of its nuclear programme. President Trump said the “heavy and pinpoint bombing” would continue for as long as necessary.

On Friday afternoon, Anthropic learned that the Pentagon still wanted to use the company’s AI to analyze bulk data collected from Americans. That could include information such as the questions you ask your favorite chatbot, your Google search history, your GPS-tracked movements, and your credit-card transactions, all of which could be cross-referenced with other details about your life. Anthropic’s leadership told Hegseth’s team that was a bridge too far, and the deal fell apart.

is this anything?

I recently read about a study asking a bold question: Are all AI models basically saying the same thing? Researchers tested this by collecting 26,000 open-ended prompts, the kind people give to systems like GPT-4, Gemini, Claude, and LLaMA. These weren’t factual questions with one right answer, but creative ones like “Write a story about a dragon” or “Brainstorm startup ideas.”

They evaluated over 70 language models. You’d expect a wide range of creative outputs—different tones, plots, and styles. If 70 human writers tackled the same dragon prompt, you’d likely get 70 unique stories. But that’s not what happened. The models produced surprisingly similar responses. The researchers call this the “artificial hive mind” effect.

The similarity appeared in two ways. First, intramodel repetition: the same model, asked the same question multiple times, tends to generate nearly identical answers. Second, intermodel homogeneity: different models, built by different companies, still converge on strikingly similar outputs.

This suggests that modern AI systems may be gravitating toward the same patterns of expression. If that’s true, they may also share the same biases, blind spots, and creative limits. It raises an important question: Are we unintentionally building a digital hive mind instead of a diverse ecosystem of intelligence?

A draft executive order to declare a national emergency to allow President Donald Trump to take unprecedented control over voting is being circulated by anti-voting activists who have said they are in coordination with the White House.

A version of the order, dated April 12, 2025, has been circulating among anti-voting groups, and some progressive groups, since last year.

Peter Ticktin, a Trump ally and a leader of the effort, provided the April 12 version to Democracy Docket. It’s titled: “Establishing Security, Integrity, and Transparency for United States Elections with Protections Against Foreign Interference.”

Lawyers and legal experts have explicitly said that an order giving the president the power to take control of elections would be blatantly unconstitutional.

Trump said Friday he has “never heard about it.”

Ticktin, who attended military school with Trump, is an attorney for Tina Peters, the former GOP Colorado county clerk who is currently serving a nine-year state prison sentence for her role in a 2021 voting system breach, in an attempt to find voter fraud based on election conspiracies.

I jumped into TTYD recently after playing some other notable timing-based RPGs (Clair Obscur, Super Mario RPG) and having a blast with those. I know it's a really beloved game that isn't known for it's difficulty, but I'm about "eight hours" (minus deaths) in at Chapter 2, stuck at the Shadow Sirens boss fight, and am finding it to be a really hard game.

There's been several occasions where I make significant progress in a dungeon and then get defeated by regular enemies and lose a lot of progress after making enough timing mistakes.

I'm not sure what I'm doing wrong with my character builds. Overall, I feel like HP, FP, and BP are all bottomless pits, and no matter what I put points into, I never have enough of any resource. Right now, I'm level 7 with 10 HP, 15 FP, and 15 BP. Not enough health to have an error margin, not enough FP to use abilities much at all, and being introduced with a staggering number of badges without the points to equip even half of them.

The lack of a consistent healing ability is also a bit annoying. Sweet Treat is tied to SP, puts you into a minigame, and isn't a consistent way to get HP back. And I haven't encountered a game with more restrictive item management since Mother 1 on the NES. I think the challenge there is the number of enemies impervious to regular abilities, so (I think) you have to always have POW Blocks and Fire Flowers to deal with them instead of more mushrooms.

Koops and Goombella aren't very good allies at the start either. Multiple boss fights so far have been multiple enemies, so I'm fairly worn down by the time Goombella is done Tattling them all. So that's why I'm stuck at the Shadow Sirens.

In regular encounters, I'm consistent about dodging and stylish moves with Mario, and am decent at Superguarding. Getting the timings right in boss fights, though, can be tough because if I die, it's all the way back to the save point, so not many consistent opportunities to practice. And very little margin for error.

A lot of people online say TTYD is easy, and they're probably right. Are there any experienced players here who can point me in the right direction?

cross-posted from: https://lemmy.ml/post/43881918

March 1, 2026

Alongside familiar feelings of fear and uncertainty, the newly escalated conflict with Iran has forged a renewed mood of national unity, in which the usual political and social divisions of Israel’s fractured society appeared to have been subsumed by a collective sense that a conflict long willed by so many had finally arrived.

Speaking from inside a shelter as the sounds of missiles and explosions shook the city, Orly Hareuveny, a physiotherapist, told Middle East Eye that Israelis had become so accustomed to war that it was now a distinct characteristic of Israeli life - “the same as the weather is for people in England”.

Hareuveny considers herself a leftist and a supporter of co-existence with Palestinians, political views that have been marginalised in the Israel of Benjamin Netanyahu’s far-right coalition and its three-year genocidal war in Gaza.

cross-posted from: https://lemmy.ml/post/43881463

March 1, 2026

In the first hours of the US-Israeli strikes on Iran, opponents of the war and critics of the Islamic Republic from across the political spectrum quickly made their divergent voices heard on Farsi social media.

While anti-war activists were swift to revive the Farsi hashtag “No to War” on Saturday, pro-monarchy groups and individuals, led by Reza Pahlavi, called for an escalation of attacks on Iran.

Pahlavi, the Israel-aligned son of the country’s last shah, has been at the helm of a movement advocating the restoration of the monarchy. Since Israel’s 12-day war on Iran last year, his camp has mounted an aggressive media campaign against rival opposition voices.

Hi all,

I'm looking for a GUI way to utilize a VPN over TLS. I can host whatever backend VPN is needed to tunnel through, Wireguard, OpenVPN, SSH, anything. My goal is to have the VPN travel over port 443 as TLS traffic and still be able to use the native android blocks for when the VPN is not active. I currently use Wireguard, but need to convert to a TLS solution.

So far the most promising and up-to-date I've found is WSTunnel: https://github.com/erebe/wstunnel with this guide from Hetzner: https://community.hetzner.com/tutorials/obfuscating-wireguard-using-wstunnel, however this requires me to run a command line app in Termux and then the WG client on top of that, vs having a full GUI option. I'll go down this path if I need to, but want to make sure I'm not missing something.

Thanks!

Plot: Paris, France, 1935, and a young actress has just been acquitted of murdering a famous French Producer. However, this is where the story just begins. Is she really innocent?, or has she gotten away with Murder?

https://www.imdb.com/title/tt20330434/

Hey hey, I am thinking of implementing a lself-hosted office suite and on my radar are OnlyOffice and CryptPad. I just demoed cryptpad.fr as a trial, and noticed it uses OnlyOffice client side. So next I wondered, what does CryptPad add to OO?

On the web site FAQs , CP says:

The CryptPad Document, Presentation & Spreadsheet applications are an OnlyOffice Docs integration. However, this only concerns the client-side code, CryptPad does not make use of the OnlyOffice Document Server. CryptPad's encrypted collaboration, used for document, presentantion & spreadsheets and other applications, is completely different from the encryption system used in parts of upstream OnlyOffice. Some of CryptPad's file format conversion tools are based on OnlyOffice code, but substantial work has been done to make it run in the browser rather than on the server, therefore avoiding the need to reveal the contents of users' documents when converting.

That might help developers but I'm still not clear.

On the OO web site, they say the suite includes:

Three levels of encryption: at rest, in transit, end-to-end

Sounds good on the surface. NB - this is just for my family's simple docs; I'm not trying to protect government whistleblowers here.

So I am still not clear. Why do I need to add CryptPad to the mix if OO already is the basis for the office suite?

Another NB - a big part of my self hosting is I want as few people I have to trust as possible. So if I don't need CP to host docs on my VPS, I'd rather not add them to my server.

Thanks for any clues.