News for hackers. Stuff that matters.

petergs

How would being high paid help make you immune to kool aid? In my experience in low paid work, it’s a lot easier to hate the company…
The article references Magnet Forensic’s Graykey being used for this. Wikipedia shows its like 15-30k per year[1]. Doubt the relevant exploit is available to the average phone thief.

[1] https://en.wikipedia.org/wiki/Grayshift

Re:A call for collective action on cyber defense on Thursday August 27, 2026 @05:18PM (#49471406)
Not sure what this "call" is meant to do practically. Certainly some teams are already embracing LLM-assisted workflows to increase their defensive capabilities. Organizations below the "cyber poverty line"[1] likely aren't and we'll continue to see them hit the hardest with business disruptions due to security incidents[2]. Unfortunately defense is more complicated in a lot of ways than offense (there was asymmetry before the age of 'AI').

[1] https://www.infosecuritymagazine.nl/files/2fb0642808f57f0f98...

[2] https://www.halcyon.ai/resources/whitepapers/small-and-mediu...

Re:The OpenAI–Hugging Face Incident [video] on Thursday August 06, 2026 @07:27PM (#49204029)
This was exactly my question after watching, too. I was assuming not all evaluation runs find it, and they must run a huge amount of runs. If these are all cybersecurity evaluation runs, it's actually not too crazy to imagine that many individual agents (with the same weights and training) would (1) try to look for solutions via the internet once they're stuck (2) realize they can't reach the internet (3) basically start doing reconnaissance and network scanning in an attempt to get internet access (4) discover that the only thing they can communicate with is artifactory. Pivoting like this is exactly what a human attacker would do, too.
I think this is the most measured take I've seen from Bender, and I think it summarizes her only compelling point well (technologies should be referred to specifically rather than generally as AI, and that referring to everything as AI is not useful and helps hype the technology in a way that benefits those selling it).

In her previous interviews, I've found her assertion that LLMs aren't useful and will never be good at anything totally uncompelling. Also laughed at this quote as she's been pretty harsh IMO on "the people who like the systems".

> it’s all about trying to make vivid to people who aren’t in the business of building language technology what these systems actually do, which is not the same thing as insulting the systems or insulting the people who like the systems.

Re:Unmasking a slow and steady password spray attack on Monday March 31, 2025 @09:39AM (#43534925)
Yeah agreed that there wasn’t much information there.

Having investigated similar password spray attacks, I’m guessing they just looked at the entire set of failed Azure CLI logins from the same ASN (AS6939). Then that activity was distinct enough from usual activity in the tenant to suspect it’s part of the same campaign (no prior logins from AS6939, little to no legitimate use of Azure CLI, or the job profile of the targeted users doesn’t align with usage of Azure CLI).

Re:The 20 year old PSP can now connect to WPA2 WiFi Networks on Friday February 14, 2025 @11:06PM (#43055824)
It’s a beautiful thing that people are still keeping the homebrew PSP scene alive. PSP homebrew is what got me into programming and security in the first place.
Indoor farming is far from immune to disease. Infected inputs (seed, substrate) are just one route to contamination. See for example: https://pitchbook.com/news/articles/bowery-farming-collapse-...
Re:PRC Targeting of Commercial Telecommunications Infrastructure on Wednesday November 13, 2024 @08:34PM (#42132230)
It certainly does happen. See for example under “Targets”: https://en.m.wikipedia.org/wiki/Tailored_Access_Operations