CISO and faculty by day, adversary emulation/tools by night, bad jokes and memes all the time.
Arizona, USA
- Tweets 15,347
- Following 1,977
- Followers 2,543
- Likes 113,803
Pinned Tweet
Super excited to present w/ @Shammahwoods at @BSides_ABQ this Friday on “Breaking In: A Story of Why Initial Access is Just Like Catching Pokemon.”
Hope you can join us for a fun look at recent trends in initial access vectors, complete with a demo of some tactics we expect to see more of, reporting on detectability levels, as well as a special Easter Egg from some super wild adversary research Ezra recently conducted.
Our talk will be Friday, September 25, 2026 at 12:00 PM at UNM Continuing Education - Albuquerque, New Mexico.
Special thanks to the many people who have shared examples, samples, attacks, and scenarios they observed (too many to tag everyone and risk forgetting someone).
Thanks to @BSides_ABQ and @m19o__ for the opportunity to present at BSides Albuquerque!!
5
4
14
1,115
💯👇
Just throwing it out there but Offensive AI Conference (OAIC) starts in about a week. This venue is good as any for cybersecurity and AI people to actually talk about working together.
offensiveaicon.com/
Fantastic read and as someone who has worked with ML and cybersecurity this is exactly what I have long suspected.
Cybersecurity and AI need to work better together
We need to help establish standardized training environments, transparency for events, we need to also create measurement systems for capability that helps us understand when to worry and when things are impressive but not end of the world.
Us in cybersecurity are extremely happy to help doing this, please reach out
3
15
1,099
Mike Manrod retweeted
There will come a time (maybe we’re there) where AI safety folks are deliberately dismissing cybersecurity out of spite, because humans are humans.
The irony is, if they asked their favorite AI chat bot, it would tell them you need safety + cybersecurity.
So is their AI wrong?
1
1
750
Mike Manrod retweeted
Folks, I have some bad news.
I thought I had one box of stickers left to scan.
I have three. 😅
👉 hackerstickers.org
4
1
10
790
Mike Manrod retweeted
Claude Sonnet is outperforming Astra. And Opus...
What in the world just happened.
OpenAI better have some pretty impressive tricks up their sleeves tomorrow.
1
1
77
🚨 LOLRMM update inbound.
Bespoke RMMs going to make a splash. You ready?
bunny-lab-io.github.io/Borea…
4
29
1,477
The universe must have gotten the notice it was Monday, and decided to give it the full and enthusiastic expression of everything a Monday can be.
3
20
459
Recently my depression was really really bad. I expressed to two people in my life that I didn’t want to leave. Only one of them picked the phone up immediately and called me. The other person texted back that they hoped I felt better soon.
If you are ever the recipient of someone saying “I don’t want to live anymore,” please believe them. And if you can’t be bothered, then stop acting like you give a damn and find someone else who loves them to help.
8
2
75
3,718
Mike Manrod retweeted
⚠️Observed phishing URLs delivering RMM payload:
RMM: ScreenConnect
Theme: DocuSign
URL: hxxps://aminacompliance[.]info/
Download URL: hxxps://leiber[.]top/Bin/ScreenConnect.ClientSetup.msi
#ThreatIntel #Phishing #RMM
3
7
445
OMFG… actual footage of @Shammahwoods and my drafts the last 24 hours before a talk. 🤣🤣🤣
piped.video/shorts/YqPPyGDfi…
7
221
A fun strategy to find old classics lost to the ages, is to pick a full album of something classic you love in your favorite streaming app (I used Apple Music), and then let it auto-play what comes next when the album ends. Sometimes it is craptastic, but also end up with some super cool legacy finds this way.
In this case, I started with Extreme, 3 sides to every story, and ended up with all sorts of cool stuff.
4
129
Mike Manrod retweeted
It’s funny because:
Most people wouldn’t know
LLMs make mistakes a lot!
And when you run dangerous experiments without appropriate safety & security measures, shit goes wrong!
“I don’t understand how OpenAI can have all these security incidents, we’ve not had any problems with the agents we run”
5
3
42
3,464
Mike Manrod retweeted
Discussions with AI: An analogy
The mom: hey can you go in my bedroom and get my charger please?
Toddler: I checked the livingroom and couldn't find it
M: it's in my bedroom please go get it
T: I checked the kitchen and didn't find it
M: please just go to my bedroom and get my charger
T: I'm sorry but it's not in the garage so that means it doesn't exist
M: It does exist, it's in my bedroom
T: I'm sorry but it's not in the bathroom, you must be mistaken. It doesn't exist
M: For the love of God it's in my bedroom!
T: Can you get the charger and give it to me, so I can give it to you?
M: I don't have it, it's in my bedroom. I've been asking you to go get it
T: I couldn't find the charger in the closet, so I brought you a cookie
M: I asked you to get my charger from my bedroom
T: That's right, you did! Where is it?
M: It's in my bedroom
T: I'm sorry but it's not in the attic, you must be mistaken
M: Let's start over - go get my charger from my bedroom
T: here's how you get your charger: step 1: go into your bedroom. Step 2: get your charger
M: I'm asking you to get my charger
T: I'm sorry I can't do that I'm not supposed to go to the car by myself it's against the rules
M: But it's in my bedroom
T: I'm sorry I forgot what we were talking about
3
13
496
