Rado RC1

Exploitation, hardware, embedded, reverse engineering, automotive security. Pwn2Own Master of Pwn Flashback team (@FlashbackPwn).

[Rado RC1 @RabbitPro]

@LiveOverflow That’s the Vulnerability Research.

[Rado RC1 @RabbitPro]

I won a chess game with a local Qwen3.8-27B. The model was surprisingly weak which I didn’t expect.

Now I need to think what it means for my Vulnerability Research model selection.

[Rado RC1 @RabbitPro]

@icrackthecode Yes. Speed and the pressure to deliver at speed. So less incentive to really understand the problem rather put more compute or quickly swap target to something else.

[Rado RC1 @RabbitPro]

Vulnerability Research is changing rapidly with the AI development. The impact on VR is massive.

After speaking with many experienced researchers I can feel a division of excitement and nostalgia to “the good old days”.

What’s your take? Few of comments that I hear:

[Rado RC1 @RabbitPro]

@alexjplaskett I assume the skills will be hard to maintain if not taken care of. There was a paper showing the more AI is used by human the less creative human becomes. Same might apply to VR creativity if we’re not maintaining muscle memory and not practicing.

[Rado RC1 @RabbitPro]

@crawl3er Yes the gap has narrowed. Question is can we lift the ceiling now?

[Rado RC1 @RabbitPro]

@viperbjk That sort of makes sense considering it’s the next token prediction machine but at the same time it’s becoming insanely good at that 😅

I assume the good old days is the pre AI era when hard and challenging problems were solved by humans and skills.

[Rado RC1 @RabbitPro]

- miss days when skill matter more then the compute power
- AI at VR will be like Nessus
- FOMO, everyone must be so much ahead
- on-prem vs self-hosted
- it will be harder to find and attract new VR researchers with proper skills, the new gen gap

[Rado RC1 @RabbitPro]

- We are guiding the research and it’s a force multiplier
- from prompt to a full scale fuzzing campaign in few minutes
- it’s like having an intern researcher on steroids that works 24hrs a day
- automated AI is benefiting more junior researchers then senior researchers

[Rado RC1 @RabbitPro]

@FuzzySec @nebusecurity Wow! Amazing!

[Flashback Team @FlashbackPwn]

Our flagship Hunting Zero-Days in Embedded Devices course is now ONLINE and ON-DEMAND! Learn how to reverse engineer embedded devices and exploit them remotely!

In this course, we will show you our tips, tricks and techniques to analyse embedded system hardware, extract its firmware, reverse engineer it, find vulnerabilities and exploit them remotely.
All of this backed up with over 14 hours of video content and tons of hands-on, long and hard exercises.

But it's not going to be easy... this is not your cookie cutter "do this, do that" course. You will be throw in the deep end, and have to survive. You've been warned!

More information below (& a massive launch discount code in the thread)
Join us at https://t.co/eOSWiKA7G9

[Rado RC1 @RabbitPro]

@mrgretzky @FlashbackPwn 🙏

[Flashback Team @FlashbackPwn]

We have been quiet, but we have not forgotten you. We're back, and with us we bring our flagship Hunting Zero-Days in Embedded Devices course... but this time, it's online and on-demand!

More information below (& a massive launch discount code in the thread) 👇

[Rado RC1 @RabbitPro]

Thanks Brussels… we love to watch AI revolution from the sidelines…

[[email protected] @daviddiaul]

Offensive cybersecurity is changing fast. ⚡️🧠

I’m looking for people in #offensive #cybersecurity who want to move beyond traditional vuln research and start integrating AI/LLM workflows into real #vulnerability #research.

Want to help shape what comes next? Reach out.

Job links in the reply 👇

#OffensiveSecurity #VulnerabilityResearch #LLM #AI #CyberSecurity

[Rado RC1 @RabbitPro]

@FuzzySec Worse case scenario to play bingo 😅

[Rado RC1 @RabbitPro]

I had a lot of discussions at offcon26 about the state of the AI, the impact it creates and the future of offensive security industry.

Opinions from “it’s a new AFL” to “we’re doomed”.

Are we going to meet all together next year at offcon27? Are you bullish or bearish on AI?

[Rado RC1 @RabbitPro]

@viperbjk You are probably right. However a lot of impact and experimentation before we settle on what’s possible and what’s not. With winners and losers as we define it.

[Rado RC1 @RabbitPro]

@saidelike @offensive_con I really liked Natashenka’s story how Gemini refused to help her :p

[Raelize @raelizecom]

We let Claude reproduce our EM fault injection attack on Google's TV Streamer 4K — from restricted adb shell to root — using only our presentation slides and the tool manuals as input.
Time to root in <15 minutes. 🤯
Full write-up 👇
https://t.co/Wgl3EjiGF9