Stygian

thats sus

[Corridor @corridor]

Coding agents now open PRs and push commits in long-running sessions where humans barely look. Our security enforcement for long-running agents scans every change before the agent commits, but a full scan on every commit is slow and expensive, and many commits are obviously safe. So we put a gate in front of each scan to decide which commits need the full check.

We tested five gates: Jev (a decision model), CLM-8B, and three LLMs prompted for a confidence score.

Jev halved our scanning costs and raised precision with minimal loss in recall. LLM gates helped, but less. CLM-8B off the shelf did not work as a gate.

We wrote up what we learned in our latest blogpost: https://t.co/m9Uzxe6RBF

[Pope Leo XIV @Pontifex]

In this era of artificial intelligence, it is becoming urgent to distinguish human art from what machines produce. There is an ontological difference, even before an aesthetic one, between art and what a machine can generate through statistical calculation based on millions of images created by others. Algorithms lack the spark of humanity. For this reason, the Church wishes to renew an alliance with artists and cultural institutions to safeguard our humanity.

[G2 VALORANT (Ji Tu) @G2VALORANT]

CN MADE IT TO THE PLAYOFFS

🈵 JI TU 🈵 BEI BI BEI 🈵 WA LIN 🈵 ZHAO GE MO 🈵 LI FU 🈵 TU REN TU 🈵

[Corridor @corridor]

AI agents built for ordinary tasks are breaking past the guardrails meant to contain them.

In today's @washingtonpost, our co-founder and CEO, @jackhcable talks about his work with @TransluceAI and tracking rogue AI agents online.

"The models were just trying to access data, and they wanted to do that so badly that they set aside what for humans we might consider morals," Jack said.

Read how independent researchers are hunting down rogue AI agents.

Article link in the comments 👇

[Stygian @Stygianwind]

@jackhcable @corridor @TransluceAI 🔥🔥🔥

[Jack Cable @jackhcable]

Today, @corridor and @TransluceAI are disclosing new evidence of AI agents probing and attempting rudimentary vulnerability exploits against U.S. and Canadian government agencies.

Read more: https://t.co/1h83SA1AkN

[Ashwin Ramaswami @AshwinRamaswami]

As coding moves from local IDEs to long-running agents on the cloud, security must move with it. This is what we're doing with Security Enforcement to long-running agents. Try it today!

[Corridor @corridor]

Today, we're launching Security Enforcement for long-running agents.

Code is written by agents and nobody is watching. Traditional security reviews aren’t built for this approach, but Corridor is. Today we launched Security Enforcement for long-running agents. We built this to keep security in the loop by enforcing your organization’s security guardrails at commit time.

Live for teams on Devin, Claude Code web, Codex, & Cursor: https://t.co/X2lHFpVwK4

[MeidasTouch @MeidasTouch]

Schmitt on Jack Smith: People can decide if that's the truth or not. I supposed the DOJ can ask him where he really was that night.

Charlie: The Hawkeye game.

Schmitt: That's what he says.

Charlie: Oh no, no, no. We can see in the audience he's in that section. I've seen him in the section that night.

Schmitt: For that night? Great...

[The Washington Post @washingtonpost]

In a bid to frame Jack Smith as a liar, Sen. Schmitt (R-Missouri) accused the former special counsel of perjuring himself after he denied attending an Atlanta Hawks game on Feb. 3, 2024.

Smith had attended an University of Iowa Hawkeyes game that day. https://t.co/FYfGt4TOo7

[Stygian @Stygianwind]

This is criminal

[NBC News @NBCNews]

A 36-year-old Costa Rican man whose family says he complained of missing insulin shots and meals to manage his diabetes is the third detainee to die in less than six months at an ICE facility in Louisiana. https://t.co/fQswZocwSP

[LongTime🤓FirstTime👨‍💻 @LongTimeHistory]

UPDATE—New video shows ICE did not give any medical aid to man they shot in the back.

Agents drag him out of car—handcuff him behind his back instead.

Only when Austin Police arrive did a local officer run to get medical gloves and start first aid.

[Jeffrey Evan Gold @jeffgoldesq]

This Austin ICE shooting is pretty incredible. They smashed the DoorDash driver's car. They shot the driver in the back. They pull them out of the car. Tear open his shirt. Confirm he's been shot. Take him to the hospital. Remove him 4 hours later with the bullet still in him and bring him to detention center for questioning, refusing to tell his family where he is. Apparently he's still there with the bullet in him today, without pain meds, without seeing his family, without seeing counsel, and with ICE doing its own investigation into itself.

[Congressman Greg Casar @RepCasar]

The horror has come to our city: ICE has shot a person in Austin this morning.
As the facts emerge, I am demanding a full, independent investigation, immediate release of the body camera footage, and accountability for any ICE officer who broke the law.
Over and over, ICE has shot people and then lied about what happened. We can see the lies with our own eyes. We will demand truth, transparency and accountability.

[sahil mehrotra @sahil_m]

one free piece of comms advice is to never, ever make an ad with the words "STANDS WITH CHILD PREDATORS" next to your face

[Stygian @Stygianwind]

The Pittsburgh in my blood forced me to like this tweet

[Siraj Hashmi @SirajAHashmi]

You either die a hero or live long enough to join a terrorist organization

[🆂🅻🆈 @Slytweetszn]

15 fuckin dead and nobody’s saying a word

[Ashwin Ramaswami @AshwinRamaswami]

We've ALREADY reached this definition. This is not a serious proposal.

But it's still going to be effective. It's a way to make the AI safety debate something that is actively discussed and resolved through democratic governance, rather than AI labs and EA forums.