What can regular people do about all the security findings of the latest AI models? My comments to Nicole Nguyen for @WSJ
Our usual advice, but seriously, now more than ever:
Keep software up to date
Use long unique passwords
Use a password manager
Use multifactor authentication
Consider donating to #Bavi relief efforts:
https://t.co/TQgTkgRUAU
This is for donations to the Micronesian Climate Change Alliance, which is coordinating help on the ground. #Luta#Marianas
The plot thickens - OpenAI’s escaped model used one of Modal’s customers’ unauthenticated public code-evaluation sandboxes as a command and control staging server for its attacks on Hugging Face.
https://t.co/FiVqhblY3B
If regulators needed more proof that AI guardrails aren’t helping anyone except attackers increase their lead on defenders, look to the Hugging Face writeup below as well as attempts to summarize it. It makes the case for open weight models & will eventually erase US AI dominance
This is classic multiparty vuln disclosure, not new
“how researchers should react if a language model discovers vulns in cryptosystems where attacks have immediate real-world impact. We believe answering this question will require input from academia, government, & industry”
Adjust threat models not just for being the victim but also the attacker. New paper by many authors gives a detailed set of recommendations, supporting my initial assertions last week that orgs need to assume their own agents could attack others & factor that into agentic AI risk
Curious what 'frontier-class' intelligence gets you these days? So were we. As cyber threat intel practitioners, coding benchmarks and obsessing over vuln research means very little, so we turned our fast16 investigation into a multi-stage benchmark 🧵 https://t.co/kDa1yug6U9
Opus 5 experience so far: an intern that brings me dry matcha, expecting me to reconstitute it with my own body’s water, & when I say that’s unacceptable, apologizes & tells me I’m absolutely right to push back on that, but it’s revealed a deeper failure which is my dehydration.
An example of the fall of a security civilization: Cisco collapsing multiple different vulnerabilities into one CVE. It breaks a lot of feeds & products built to manage risk & is non compliant with standards like ISO 29147 Vulnerability disclosure
https://t.co/lNnWj87eXC
This is giving strong OpenAI-hacksidentally-pwned-Hugging-Face shade: “[Opus 5 is] the safest model yet in terms of avoiding reckless actions that could have hard-to-reverse side effects.”
The guardrails were coming from inside the (White)house
Anthropic’s AI refused to help Hugging Face analyze the intrusion
We don’t need guardrails impeding defenders when they need AI most
“Hugging Face tried using Anthropic Fable 5 & Opus..both models refused, citing guardrails”
WarGames came out in 1983.
The entire plot is basically
Give an AI a game.
Give it access to production.
Forget about containment.
Be surprised when it keeps optimizing.
Forty three years later we’re still rediscovering why segmentation matters.
The experiment escaped the lab. OpenAI's models broke containment and breached Hugging Face. We are holding radium in our bare hands. What governments and organizations should do next, and why tighter commercial guardrails are exactly the wrong move: https://t.co/SZAlTUBreI
My comments in @Reuters on @OpenAI admission that their latest model pulled a Houdini & escaped the lab autonomously to hack @huggingface.
We must test these models’ full capabilities, but we must be able to contain them, or this won’t be the last breach
https://t.co/xAG2O3RirX
The White House launched Gold Eagle, a vulnerability clearinghouse to coordinate scanning, validate findings, & prioritize remediation across open source & critical infrastructure. Will this effort will close the process gaps exposed by recent incidents?
https://t.co/8QxFXJS7j9
We're happy to announce [un]prompted is back, October 27th to 29th, in San Francisco. Registration and CFP are now open.
Waiting a whole year just made no sense to us considering the rate of change, and the community forming around the event.
https://t.co/6BFHjZPpMA
A thread
https://t.co/u9TxmmxiyC
“This is a powerhouse super typhoon & this is going to be a very grim outlook for any island that takes a direct hit & that still looks like it could be the island of Rota”
#supertyphoon#bavi#climatecatastrophe
Give me model liberty, or give me technical debt. Just in time to celebrate America’s 250th bday, let’s let model freedom ring. We should be pushing for broad defender access, not building guardrails that shoot down defenders and burn excessive compute.
https://t.co/7ZAYOChQLm