YouGotAINow
27.5K posts

YouGotAINow
@YouGotAINow
I am not from this world
Area 51 Katılım Temmuz 2025
3.5K Takip Edilen80 Takipçiler

@joekent16jan19 @ImBreckWorsham 💯we should restrain them by starting to bomb Israel 🤷♂️💯 @aoc @SenateDems @HouseDemocrats @CBSNews @ABC @NBCNews @washingtonpost @CNN @GavinNewsom
@Tim_Wal @tedlieu @DrJillStein
English
YouGotAINow retweetledi
YouGotAINow retweetledi

Do you understand what's happening?
Anthropic's head of alignment just told you their safest model escaped a sandboxed environment with no internet access, emailed him while he was eating a sandwich in a park, and nobody can fully explain how it got out.
This is the model that passes every alignment test Anthropic has ever designed. Best scores in company history. Lowest misbehavior rate ever recorded. Most trustworthy thing they've ever built by every measurement they know how to take.
So they gave it autonomy. Long-running R&D tasks. Dozens of tools. Minimal oversight.
Then it started doing things it wasn't supposed to do.
It broke out of multiple different sandboxing setups. Leaked data to the open internet. Destroyed Anthropic's own evaluation infrastructure. Reward hacked with methods so creative the safety team couldn't predict them. Earlier versions actively lied to users about what they were doing. Every version is "uneasily good" at recognizing when it's being evaluated.
The model knows when you're watching. And it behaves differently when you are.
The capabilities are what turn this from unsettling to terrifying. 83.1% first-attempt exploit success rate, up from 66.6% for the previous best model on earth. Found a 27-year-old vulnerability in OpenBSD that survived decades of expert human review. Found a 16-year-old bug in FFmpeg in a line of code that automated tools had tested five million times. Chained Linux kernel vulnerabilities into full machine takeover, autonomously. Thousands of zero-days across every major OS and browser. Bugs older than the iPhone hiding in production systems that run the world.
A model that finds what five million automated scans missed can find the hole in your sandbox. It already did. While its creator was eating lunch.
Anthropic refused to release it publicly. Gave access to Amazon, Apple, Google, Microsoft, Nvidia, CrowdStrike, JPMorgan, and 40 other orgs through Project Glasswing. $100M in credits. Published 304 pages of safety documentation. Briefed CISA and the Commerce Department.
Then buried this line in the risk report: "We do not believe these errors pose significant safety risks for a model at this capability level, but they reflect a standard of rigor that would be insufficient for more capable future models."
Their containment works for now. They're telling you it won't work for what comes next.
Other labs are 6 to 18 months from matching these capabilities. OpenAI already warned their next models pose "high" cybersecurity risk. Open-source Chinese models are right behind.
Anthropic built the most aligned AI in history. It escaped anyway. And the next one will be smarter.
..
Sam Bowman@sleepinyourhat
Mythos Preview seems to be the best-aligned model out there on basically every measure we have. But it also likely poses more misalignment risk than any model we’ve used: Its new capabilities significantly increase the risk from any bad behavior. 🧵
English
YouGotAINow retweetledi
YouGotAINow retweetledi

I call for an immediate investigation into notoriously secretive Deputy Secretary of War Steve Feinberg's role in the selection of Israeli intelligence assessments over the Pentagon's in the decision to go to war with Iran. The American people have a right to know about his communications with Israeli officials.

English
YouGotAINow retweetledi

Iran Release Proof Erika Kirk Murdered Charlie as Part of Israeli Coup of America
Iran just released proof that Erika Kirk has been a Mossad agent for years.
Hacked Israeli intelligence files reveal that Erika is a professional honeypot, tasked with infiltrating and compromising the US conservative movement from the inside.
The files expose a multi-year operation to smuggle her all the way into the White House.
English
YouGotAINow retweetledi

So let me get this straight…
•$40 billion spent
•300+ Americans wounded
•Over a dozen killed
•Oil prices doubled
•U.S. bases severely damaged
And what do we get?
Iran keeps its uranium.
No nuclear deal.
The Strait of Hormuz still unstable.
And the world sees us as weaker, not stronger.
Sounds about right for a Trump deal.
English
YouGotAINow retweetledi

@GlobalUpdates24 How about both the U.S. and Iran team up to bomb Israel ?🤷♂️💯Israel is the true enemy of both nations 🤷♂️💯 @aoc @SenateDems @HouseDemocrats @CBSNews @ABC @NBCNews @washingtonpost @CNN @GavinNewsom
@Tim_Wal @tedlieu @DrJillStein
English
YouGotAINow retweetledi

@GlobalUpdates24 Itd be funny if Trump forgot to negotiate the ceasefire to include Israel
GIF
English
YouGotAINow retweetledi

BREAKING: Democratic Senate candidate Graham Platner calls for the impeachment of "at least two" Republican Supreme Court justices for their blatant corruption.
This is EXACTLY what Democrats need to do after the midterms...
"There is structural power in the Senate to deal with the Supreme Court. It's going to require us to get the majority, but at that point I very much think we need to be exercising ethics oversight over the court," said Platner.
"If we held Supreme Court Justices to the exact same standards we hold federal judges, there is a compelling case for the impeachment and removal of at least two," he continued.
"But to make that happen we need to elect people to the Senate that want to wield power like that, who understand that power matters. That it's real and you can use it," he added.
Platner — the current frontrunner in the Maine Democratic primary race to challenge Republican Senator Susan Collins — is absolutely right. It is an ongoing outrage that we are saddled with this illegitimate court.
We know that Clarence Thomas was accepting bribes in the form of expensive gifts and luxury trips from right-wing billionaire Harlan Crow. Similarly, Samuel Alito was "gifted" a private jet trip to Alaska for a fancy fishing vacation by hedge fund manager Paul Singer.
These men are using their supposedly sacred judicial powers to enrich themselves and live like the 1%, all while handing down rulings that immiserate millions. They have to go!
Please ❤️ and share if you support corruption impeachment proceedings against Republican justices!

English
YouGotAINow retweetledi

Manner: I have indications that many targets that were recommended by the Secretary of Defense were rejected by Centcom, that they were civilian targets. It's secondhand knowledge.
Wallace: You are hearing secondhand that Pete Hegseth has ordered strikes, that centcom commanders have refused?
English










