If you got into cyber before, congrats
If not, ngmi
You're never beating research teams with uncensored frontier models which you can't even access
Introducing GPT‑6 Cyber: a major advance in security research that pairs stronger capabilities with better model alignment.
The model powered the Linux kernel research we announced yesterday, and security teams at companies like Mozilla, Palo Alto Networks, Calif, and Cisco—alongside our own at OpenAI—are already putting it to work.
GPT-6 Cyber is available to select Daybreak Red customers and we are expanding access over time.
openai.com/index/introducing…
~200k
yeah security has fundamentally changed
Our security researchers built an autonomous cyber vulnerability discovery, exploitation, and patching pipeline.
In a single run with an upcoming cyber model, it validated and generated patches for 948 local privilege escalation vulnerabilities in the Linux kernel at a total API cost of $1.2 million.
That was indeed me
No mallace intended
External network blocks affecting access to Railway
Some users are unable to access Railway-hosted applications or railway.com because network providers and security services have blocked Railway IP addresses.
We’ve confirmed that a …
Status: Identified status.railway.com/incident/…
OH MY GOD 😭😭😭
JUST SHUT UP DAWGGGGGGG
IT'S BETTER FOR YOU
CONSULT LAWYER OR SOMETHING
Replying to @rosnburg @complex_maths
It was indeed only run on the one target. Can't tell you why, but can tell you just the one
I don't think it's a contradiction to adopt a policy regardless. We didn't do any outbound automated testing without consent prior, and we won't ever in the future
Can't complain since it's free
But I've sent one message to deepseek flash model and I'm out of messages on opencode?
Anyone can build software
But the difference between people who have experience building software and those who are new is exponential
Maybe it's cope?
But the difference in quality, performance, ui/ux is insane
Can a normal person ask the agent to improve the performance?
They absolutely can!
But will they?
why are we not holding people accountable for their agent's actions?
it should be the absolute minimum
otherwise we will keep seeing morons which attack infra and hide behind the "oops my agent did this sowwy 🥺"
"SORRY THAT WAS MY AGENT WHICH TOOK DOWN YOUR INFRA
SILLY ME! 🥺"
Ah yup. That was indeed me. No mallace intended. I've been running some agents to do signups + agentic testing on similar platforms. Just halted the harness I was designing around this as it was definitely going off the rails and reaching too far
Readers added context they thought people might want to know
In many jurisdictions this is a crime. In the US under Computer Fraud and Abuse Act, 18 U.S.C. §1030, in the UK under Computer Misuse Act (CMA) 1990, in Germany Strafgesetzbuch (StGB), and many others.
justice.gov/jm/jm-9-48000-…
legislation.gov.uk/ukpga/1990/18/…
gesetze-im-internet.de/stgb/
yc is so trash
sure they funded coinbase/doordash etc
you probably will never see anything useful come out of yc in the future
just look at the recently accepted batches and their works
i really can't see any of them surviving a couple of years without raising again





