Sarah Szalavitz💡's Avatar

Sarah Szalavitz💡

@dearsarah.bsky.social

Accountability Architect💡 Artisanal Database Maker🔎 🧾 Storyteller/Writer/Producer🎙️✍️🎥 Avenger, Academic, Lawyer🗳️ Redistributive Justice ⚖️ Choice Optimizer🤖 Researcher & OG Prompt Engineer🧌 Whistleblower🚨 Human Palantir🔮 https://linktr.ee/realdearsarah

119,387 Followers  |  1,148 Following  |  37,588 Posts  |  Joined: 30.04.2023
Posts Following

Posts by Sarah Szalavitz💡 (@dearsarah.bsky.social)

Preview
The Data Centers Have Arrived at the Edge of the Arctic Circle As AI labs gorge themselves on compute, data center operators are flooding north in search of cheap and plentiful energy.

The Data Centers Have Arrived at the Edge of the Arctic Circle flip.it/88k4rS proving once again that yes it really can always get dumber.

02.03.2026 10:27 — 👍 12    🔁 3    💬 0    📌 0
According to OpenAl, the limitations it has imposed are more enforceable than Anthropic's because it will only provide the DOW with its technology via the cloud, rather than installing it directly on hardware. OpenAl personnel will also be kept involved so that they can see how the DOW is using its technology. This will allegedly allow the company more oversight and control of its Al systems.
"We don't know why Anthropic could not reach this deal, and we hope that they and more labs will consider it," wrote OpenAl.
However, an excerpt of the contract shared by OpenAl indicated that its technology will only be barred from use in autonomous weapons or to surveil U.S. citizens where such use is illegal. In fact, the agreement appears to lay out circumstances where OpenAl's tech would be allowed for these purposes, such as where human control over weapons isn't required by DOW policy or law.
"The Department of War may use the Al System for all lawful purposes, consistent with applicable law, operational requirements, and well-established safety and oversight protocols," the contract reads, per OpenAl. "[A]ny use of Al in autonomous and semiautonomous systems must undergo rigorous verification, validation, and testing to ensure they perform as intended in realistic environments before deployment."

According to OpenAl, the limitations it has imposed are more enforceable than Anthropic's because it will only provide the DOW with its technology via the cloud, rather than installing it directly on hardware. OpenAl personnel will also be kept involved so that they can see how the DOW is using its technology. This will allegedly allow the company more oversight and control of its Al systems. "We don't know why Anthropic could not reach this deal, and we hope that they and more labs will consider it," wrote OpenAl. However, an excerpt of the contract shared by OpenAl indicated that its technology will only be barred from use in autonomous weapons or to surveil U.S. citizens where such use is illegal. In fact, the agreement appears to lay out circumstances where OpenAl's tech would be allowed for these purposes, such as where human control over weapons isn't required by DOW policy or law. "The Department of War may use the Al System for all lawful purposes, consistent with applicable law, operational requirements, and well-established safety and oversight protocols," the contract reads, per OpenAl. "[A]ny use of Al in autonomous and semiautonomous systems must undergo rigorous verification, validation, and testing to ensure they perform as intended in realistic environments before deployment."

OpenAI CEO Sam Altman responds to (criticism of & concerns about) deal with “Department of War” flip.it/Zrl4gg Unsurprisingly, Altman is provably & blatantly lying.

As Altman knows, Anthropic refused to permit “any lawful use”—but OpenAI agreed.

Hope @robbonta.bsky.social files an injunction…

02.03.2026 10:12 — 👍 3    🔁 1    💬 0    📌 0

Related: again, it’s fine that we don’t agree—but this reply, yes, read literally, crosses into condescending & insulting…& everything is awful enough… so again, I’m going to leave it here. bsky.app/profile/klea...

02.03.2026 00:46 — 👍 0    🔁 0    💬 0    📌 0

Or we disagree on what “the broader picture” is & the complexities of “seizing” Anthropic &/or replacing its 7 founders.

Obvi Trump dgaf about laws or Congress—but its always been easy for the Commander in Chief to lawlessly war crime.

But if Trump tries to seize a 380b US ai co, he’ll fail fast

02.03.2026 00:34 — 👍 0    🔁 0    💬 1    📌 0
Iran’s soccer president casts doubt on World Cup participation following strikes The US-Israeli joint military campaign could prevent the nation from playing in the tournament in the US this summer.

Iran’s soccer president casts doubt on World Cup participation following strikes

01.03.2026 22:35 — 👍 179    🔁 42    💬 19    📌 16
Post image Post image

Freedom is not a geopolitical byproduct.

It cannot be subcontracted.

It cannot be bombed into existence.

Liberation is built from within - by those who bear the consequences long after foreign states move on.

10/13

01.03.2026 22:36 — 👍 6    🔁 1    💬 1    📌 0

While I’m not going to defend this or any of the prior presidents who dropped bombs without congressional approval, I do not understand why you think bombing Iran is a useful comparison to seizing a US startup.

I hope we can agree to disagree; I’m going to leave it here now.

01.03.2026 22:36 — 👍 1    🔁 0    💬 0    📌 0

Sure.

That doesn’t change the situation with Anthropic though, regardless of what he says on social media…or even if Trump signs an EO declaring that he (or🇺🇸) now owns Anthropic (&/or, while he’s at it, all AI cos).

It’s fine that we don’t see this the same way—time’ll tell:)

01.03.2026 22:32 — 👍 0    🔁 0    💬 1    📌 0

TikTok required legislation—there’s no chance Congress passes a law to force the sale of Anthropic, a US startup.

In other cases (US Steel, Intel), the companies voluntarily agreed to give Trump a %.

Hegseth already has retaliated—he declared them a supply chain risk & cancelled their contract.

01.03.2026 22:24 — 👍 0    🔁 0    💬 0    📌 0

Text is tough; since we don’t know each other, I’m not sure how else to read messages other than literally.

There was a law (signed by Biden!!!) requiring TikTok’s sale, based on its Chinese ownership—so for this & other reasons, it is not comparable?

That’s not my read; but it’s ok to disagree.

01.03.2026 22:21 — 👍 0    🔁 0    💬 0    📌 0

Anthropic is valued at $380b (FWIW US TikTok is $14b)—& Anthropic’s contracts with the USG aren’t a significant % of their biz.

Anthropic doesn’t need DoD or to stay in the 🇺🇸—but DoD proved in less than 24 hours that they currently need Anthropic & also will not abide by the ToS or rule of law.

01.03.2026 22:13 — 👍 0    🔁 0    💬 0    📌 0

I watch pretty closely—& am critical of both the AI industry & Trump.

What you are suggesting isn’t comparable to what’s happened with TikTok (or Intel. or US Steel etc)—Trump’s USG hasn’t taken over the operation of any co…& if they did, in AI, (aside from Grok?) workers would just quit.

01.03.2026 22:04 — 👍 0    🔁 0    💬 2    📌 0

Me too! So did the nurse who vaccinated me yesterday:)

01.03.2026 21:56 — 👍 1    🔁 0    💬 0    📌 0
Preview
Inside this downtown LA utility box, there’s a tiny theater Artist S.C. Mero is taking a familiar form of public art to a different place.

Inside this downtown LA utility box, there’s a tiny theater flip.it/e1lzEi cc @sonaar.bsky.social @safyhallanfarah.bsky.social

01.03.2026 21:55 — 👍 7    🔁 3    💬 0    📌 0

Maybe you’re joking…but as much as as I am an abolitionist except for 1%ers & also a critic of AI, that’d destroy the US AI industry, possibly overnight. And the US economy is propped up by the AI hype bubble & its warehouses…

Maybe that’s the fastest way to destroy the US & its future economy?

01.03.2026 21:41 — 👍 1    🔁 0    💬 0    📌 0

“I’m begging everyone to please be normal about XYZ Public Figure with Big Platform”

I’m begging everyone to recognize that Public Figures with Big Platforms who punch down on trans people, especially small accounts, reveal themselves to be toxic, narcissistic — name your poison

01.03.2026 21:13 — 👍 9    🔁 1    💬 1    📌 0
Preview
US military reportedly used Claude in Iran strikes despite Trump’s ban Trump calls Anthropic a ‘Radical Left AI company run by people who have no idea what the real World is all about’

DoD reportedly used Claude in Iran strikes despite Trump’s ban

Trump calls Anthropic a ‘Radical Left AI co run by people who have no idea what the real World is all about’
www.theguardian.com/technology/2... so DoD will keep using Anthropic—in violation of the contractual ToS,Trump’s order & laws…

01.03.2026 21:18 — 👍 8    🔁 3    💬 1    📌 1
Preview
Thermal Drone Footage Shows Musk’s AI Power Plant Flouting Clean Air Regulations Images confirm xAI is continuing to defy EPA regulations in Mississippi to power its flagship data centers.

A resident was forced to abandon the home that's been in her family for generations after xAI parked gas turbines across the street. "I do feel like xAI is playing by a different set of rules."

01.03.2026 21:01 — 👍 206    🔁 75    💬 2    📌 2
Post image

“my dog is in dubai”

01.03.2026 21:06 — 👍 2533    🔁 158    💬 86    📌 26

For folks asking: I was making a joke….& since folks are asking, it was apparently not a good one:)

01.03.2026 21:01 — 👍 3    🔁 0    💬 2    📌 0

Vaccine Seeking Behavior is the new Drug Seeking Behavior.

01.03.2026 03:53 — 👍 8    🔁 1    💬 5    📌 0

If you are an elected official that opposes military strikes on Iran, the head of CBS News will clap back at you. Will be interesting to see if she does this for other objectors, or just Muslim politicians, but it does raise the question how much her views of the Middle East shape the news.

28.02.2026 21:18 — 👍 2978    🔁 572    💬 43    📌 13
Preview
X Is Drowning in Disinformation Following US and Israel’s Attack on Iran WIRED has reviewed hundreds of posts on X that promote misleading claims about the locations and scale of the attack.

“Almost all of the most viral posts reviewed by WIRED on Saturday came from accounts with blue check marks, meaning they…could be eligible to earn money based on how much engagement their posts generate, even if the content is false.”

From @davidgilbert.bsky.social

28.02.2026 22:45 — 👍 1248    🔁 458    💬 36    📌 42
Preview
No Clear Endgame in the Conflict Between Afghanistan and Pakistan

No Clear Endgame in the Conflict Between 🇦🇫 and 🇵🇰

Pakistan’s airstrikes in Afghanistan showed its overwhelming superiority in conventional warfare, but the Taliban have a lethal repertoire of guerrilla tactics. www.nytimes.com/2026/02/28/w... that & 🇺🇸 🇮🇱 & 🇮🇷 are bombing the rest of the Middle East

28.02.2026 22:27 — 👍 4    🔁 3    💬 0    📌 1
Preview
Israel Beats Trump to the Mic With Major War Update Israeli officials confirmed that the Ayatollah was killed after Trump launched an unauthorized, full-scale war in the Middle East.

Israel Beats Trump to the Mic & Confirmed the Assassination of the Ayatollah.

Following the Ayatollah’a refusal to exonerate Trump for appearing 1000s of times in the Epstein files, Trump & Bibi launched an unauthorized, full-scale war in the Middle East. www.thedailybeast.com/israel-beats...

28.02.2026 22:22 — 👍 15    🔁 5    💬 2    📌 1

why are things simultaneously so fucking ridiculous yet terrifying

28.02.2026 13:03 — 👍 160    🔁 30    💬 5    📌 2

that top DOD officials spent the entire day feuding with a tech company and the boy scouts before launching a war in the middle of the night should probably inform your opinion about how well planned and executed this will be

28.02.2026 08:22 — 👍 3257    🔁 845    💬 31    📌 22
Preview
New 'Ride the D' Metro shirts sold out almost immediately. Because of course they did L.A. Metro really wants people to take their soon-to-be-opened rail line extension, and have come up with a virally popular marketing method to spread the word — a new line of merchandise proudly embl...

LA Metro's new 'Ride the D' shirts quickly sold out. Because of course they did www.latimes.com/california/s... the tshirt is the perfect souvenir for the worst version of reality—& obvi the headline is false!

Ride the D is a sellout—but these shirts aren’t sold out shop.metro.net/products/rid...

28.02.2026 10:32 — 👍 1    🔁 0    💬 0    📌 0

Afaik OpenAI had already announced the deal—it’s mentioned in the piece
bsky.app/profile/dear... Also maybe it’s just me..but was there any question that Altman would agree to the DoD’s terms? Or any doubt that Altman has no problem with using OpenAI to spy on US citizens or for autonomous warfare?

28.02.2026 08:55 — 👍 0    🔁 0    💬 0    📌 0
Preview
Statement on the comments from Secretary of War Pete Hegseth Anthropic's response to the Secretary of War and advice for customers

Statement on the comments from Secretary of War Pete Hegseth from Anthropic www.anthropic.com/news/stateme...

28.02.2026 08:45 — 👍 4    🔁 1    💬 1    📌 0