Ben McCarty — security-maximalism 🛡️ retweeted
My first question to everyone is "Do you believe in Mathematical Cryptography?" The follow up is "what are you doing to protect democracy with it?" No fallacies. No block-chain hate. Just our ability to mathematically prove every citizen's vote is counted as casted, & as intended
The first question that must be asked of every Republican elected official is “do you believe in democracy?” The follow up is “what are you doing to protect it?” No equivocations. No gaslighting. No other question comes close to being as important.
0
1
1
Unofficial fan art I designed in tribute to "IF ANYONE BUILDS IT, EVERYONE DIES" - published today! Currently reading the book and enjoying it.
(Not affiliated, endorsed, or sponsored.)
#FanArt
(Not affiliated, endorsed, or sponsored.)
#FanArt
0
0
3
197
Ben McCarty — security-maximalism 🛡️ retweeted
Nate Soares and I are publishing a traditional book: _If Anyone Builds It, Everyone Dies: Why Superhuman AI Would Kill Us All_. Coming in Sep 2025.
You should probably read it! Given that, we'd like you to preorder it! Nowish!
You should probably read it! Given that, we'd like you to preorder it! Nowish!
263
383
1,929
1,406,262
Ben McCarty — security-maximalism 🛡️ retweeted
Last day to submit questions to this crowdsourced benchmark.
agi.safe.ai
agi.safe.ai
Have a question that is challenging for humans and AI?
We (@cais + @scale_AI) are launching Humanity's Last Exam, a massive collaboration to create the world's toughest AI benchmark.
Submit a hard question and become a co-author.
Best questions get part of $500,000 in 1, 2024
Details: safe.ai/blog/humanitys…
Submit here: agi.safe.ai/submit
0
3
28
5,961
Ben McCarty — security-maximalism 🛡️ retweeted
Three models remain unbroken in the Gray Swan jailbreaking competition (~500 registrants), which is still ongoing.
These models are based on Circuit Breakers + other RepE techniques (arxiv.org/abs/2406.04313, ai-transparency.org).
app.grayswan.ai/arena/leaderbo…
These models are based on Circuit Breakers + other RepE techniques (arxiv.org/abs/2406.04313, ai-transparency.org).
app.grayswan.ai/arena/leaderbo…
7
20
175
32,389
A recent publication introduces the RAMBO air-gap bypass technique, which leverages electromagnetic emissions from RAM to transmit signals covertly across an air gap.
arxiv.org/abs/2409.02292
arxiv.org/abs/2409.02292
0
2
2
322
Ben McCarty — security-maximalism 🛡️ retweeted
Presumably the real purpose of this company is to refute people who said "We'll just walk over to the superintelligence and pull the plug out", without MIRI needing to argue with them
🚀@LumenOrbit (YC S24) is building a network of megawatt-scale data centers in space, scalable to gigawatt capacity.
Congrats on the launch, @johnstonphil, @EzraFeilden, and @AdiOltean!
ycombinator.com/launches/LmD-l…
19
15
288
24,058
Script kiddies have compromised countless orgs despite being "naïve hackers".
Credentialed security experts have repeatedly proven that they can't always stop these "naïve hackers".
“One good test is worth a thousand expert opinions.” -- Wernher von Braun
Credentialed security experts have repeatedly proven that they can't always stop these "naïve hackers".
“One good test is worth a thousand expert opinions.” -- Wernher von Braun
Remember when naïve hackers were called script kiddies?
What would be the equivalent designation for folks who think they are AI experts after a couple hours of LLM prompt engineering?
Prompt kiddies?
0
0
1
200
So, vulnerable prototyping with no duty to security until incidents happen, which you will then iteratively patch in 'fixes' as an afterthought while increasing its risk until a government regulation steps in?
Black Swan events are not a "myth".
AI tech has new risks to manage.
Black Swan events are not a "myth".
AI tech has new risks to manage.
Every new technology is developed and deployed the same way:
You make a prototype, try it at a small scale, make limited deployment, fix the problems, make it safer, and then deploy it more widely.
At that point, governments regulate it and establish safety standards.
1/
0
0
1
188
Remembering the time I was responding to "URGENT!" internal IR threat intel requests nearly every Friday night & Saturday for over a year - until the IR team eventually admitted the incidents had already been handled, weren't urgent & they were only testing how good our intel was
2
0
1
The IR MD had a feud with the Threat Intelligence MD - so I became some pawn in their proxy war between each other
Being a sacrificial pawn for office politics & reaping no benefits after never quitting or failing sure is something
Wish I knew what the lesson of this story is...
Being a sacrificial pawn for office politics & reaping no benefits after never quitting or failing sure is something
Wish I knew what the lesson of this story is...
1
0
2
During the middle of this feud, the IR MD hung this up directly in front of my workstation in our office
0
0
0
Ben McCarty — security-maximalism 🛡️ retweeted
Want to understand why the physics Nobel awarded today is such a big deal?
Alain Aspect, John F. Clauser and Anton Zeilinger were awarded the Nobel prize today for groundbreaking work in quantum technologies (1/?)
Alain Aspect, John F. Clauser and Anton Zeilinger were awarded the Nobel prize today for groundbreaking work in quantum technologies (1/?)
21
641
2,491
Ben McCarty — security-maximalism 🛡️ retweeted
Rick as a pickled gherkin epitomises the human condition. With only his mind, and minimal access to the world, he subdues it. #RickAndMorty
21
122
803
Ben McCarty — security-maximalism 🛡️ retweeted
NRL Announces the Washington Metropolitan Quantum Network Research Consortium (DC-QNet) dvidshub.net/news/423832/nr…
1
1
2
Ben McCarty — security-maximalism 🛡️ retweeted
Introducing the #AWS Center for Quantum Networking, our new home for #quantum networking R&D. We're on a mission to build the underlying technologies that will enable the distribution of quantum information anywhere on Earth. aws.amazon.com/blogs/quantum-…
5
32
113