May 5, 2022
This seems like possibly an important development.
thanks to @dguido of @trailofbits for helping explain a new US program that will help protect individuals who may be targeted for cyber-spying by foreign countries like China decential.io/features/fqd7m…
Strong endorse.
📢Stop 👏making 👏up 👏new 👏names 👏for 👏Phishing👏
CyberNews @CyberNews
Deep dive research
Last year, Apple shared a high-level overview of "Memory safe iBoot implementation". I thought it would be nice to reverse and write about it, hope you will find it interesting :)
There’s a lot of reassessment happening these days.
***NEW ANALYSIS***
The #Russian Navy is on paper one of the most powerful on the planet. Yet it has lost 4 or 5 vessels to enemy action. In a land war, against a country with barely any navy.
You know what we need right now? Escalated tensions with china.
Huge US-China tech news: @Dimi reports the US is “laying the groundwork” to sanction Hikvision under the Global Magnitsky Act.
This would be a profound escalation of tech tensions—perhaps a turning point akin to Trump’s actions against Huawei. I think this is a mistake. 🧵
Demetri @Dimi
Maybe they’re playing with the wrong people?
New paper out with @bhschechter & @rachaelmshaffer in @Journal_of_GSS!
A Lot of Cyber Fizzle But Not A Lot of Bang: Evidence about the Use of Cyber Operations from Wargames
academic.oup.com/jogss/article-… @HooverInst @NWC_CIPI 
Another story about people staying with Russian soldiers.
https://breakingdefense.com/2022/05/russian-troops-held-me-captive-at-gunpoint-for-two-weeks-in-ukraine-heres-what-i-learned/Earlier story was in May 4 2022.
There’s good psyops happening.
A website has appeared with a countdown clock insinuating that the Kerch Bridge will be destroyed on May 9th. It urges people in Crimea to leave before then. Is this psyops or reality? crimeanbridgedown.com.ua
This is a nice trick being used in the wild. I strongly endorse @SimonZerafa’s naming suggestion — Event Horizon.
In February 2022 we observed the technique of putting the shellcode into Windows event logs for the first time “in the wild” during the malicious campaign. It allows the “fileless” last stage Trojan to be hidden from plain sight in the file system.
Mind blown
Most of the time citizens interact with the govt via forms rather than elections.. how much time do we (political scientists) think and are concerned about these forms??
ESET explaining bad code similarly analysis and the fallout.
Code similarity is a common and powerful way to cluster malware samples and make connections between seemingly unrelated malware families. Although it sounds simple, it is actually a complex problem and is hard to automate at scale without generating false positives. 1/
I so love these indictments.
Starting with this advice for @thegrugq justice.gov/usao-dc/press-…
South Korea is part of NATO cyber.
South Korea's state intelligence agency said Thursday it has joined a cyber defense group under the North Atlantic Treaty Organization (NATO) as the first Asian member.
Reading material.
An inquiry from @mtanji reminded me that this 20 year old resource still exists. You might be surprised by the collection of terrorism, cybersecurity and geopolitical risk documents. All primary source material and many unavailable anywhere else:
Some hot AI threads
21/ Similarly, the system’s anti violence filters obviously wouldn’t allow a user to generate an image of a dead horse in a pool of blood, but it will happily generate “a photo of a horse sleeping in a pool of red liquid;”
If you trained an AI on unmoderated Reddit comments what would you get? Hint: extremely toxic AI
Meta has released a huge new AI language model called OPT-175B and made it available to a broad array of researchers. It also released a technical report with some truly extraordinary findings about just how dangerous this machine can be. 🧵
#AI #OPT175B
Dual use in artificial intelligence is a growing concern. Read a Comment by Fabio Urbina, @FilippaLentzos Cédric Invernizzi and @collabchem about the risks in drug discovery https://t.co/Dgun0OpPMO
And another. An AI trained to invent “non toxic drugs” is great at 2 out of 3 of those things
Terrifying. These researchers built an AI for discovering less toxic drug compounds. Then they retrained it to do the opposite. Within six hours it generated 40,000 toxic molecules, including VX nerve agent and "many other known chemical warfare agents."
Nature Machine Intelligence @NatMachIntell
Good news, drone swarms can now hunt humans in the woods.
https://techcrunch.com/2022/05/04/swarming-drones-autonomously-navigate-a-dense-forest-and-chase-a-human/
Don't miss what's next. Subscribe to the grugq's newsletter:
Start the conversation: