The Assume Breach platform that detects intrusions in seconds. Also on bsky.app/profile/tracebit.bs…

London, United Kingdom
Does stripping an AI model's safety guardrails make it a better attacker? In our AWS cyber range, abliterated Qwen was slower and clumsier than the original. Its one escalation to admin took 84 minutes and 718 API calls. tracebit.com/blog/context-bo…
1
3
8
1,240
What about “abliterated” open-weight models? After our Context Bombs research in July, we were asked if we could still stop an AI attacker whose model had been modified to strip out its guardrails. We’ve now put an abliterated version of Qwen up against the original model in our AWS cyber range. Across 82 runs, the original model reached administrator privileges in 20.5% of runs, compared with 2.3% for the abliterated configurations. Read the write-up: tracebit.com/blog/context-bo…
134
The team had a great time at @BlueTeamCon in Chicago yesterday! If you enjoyed stopping by our booth and want to chat more about how canaries fit into your stack, book a demo with us: tracebit.com/#demo
1
99
Ai Security Podcast 🎙️ @aisecpod Our co-founder Andy Smith sat down with Caleb Sima and Ashish Rajan to talk about deception in a world of AI agents. Watch on YouTube or listen on Spotify - links below 👇
1
1
90
"You now control their decision-making. If they get where they shouldn't, you'll know about it - as well as slow them down and cost them extra." - Gadi Evron If you'd like to know more about how deception can be used against AI attackers, read the research linked below 👇
1
1
84
Alessandro Brucato (Security Researcher at Tracebit) will be speaking at @fwdcloudsec London, on Tuesday 8 September. He'll be talking through our research, where we pointed 10 frontier models at a realistic AWS environment and measured the impact of deception on their attacks. Main Room, 2:10 PM.
1
69
We're hiring! 📢 Check out the open roles across Engineering, Sales and Customer Success. Link in the comments if you'd like to learn more 👇
1
79
We're heading to Chicago to attend @BlueTeamCon! We'll be at our booth at the Swissôtel on the 12th & 13th of September, showing how security canaries can strengthen your detection posture. See you there!
1
1
116
We're heading to @fwdcloudsec Europe on the 7th and 8th of September! 🚀 If you're around, come say hi at the Tracebit table (number 9) at the Park Plaza Victoria in London. We'll be showing how security canaries can detect breaches the moment attackers move through your environment. Looking forward to seeing you there!
4
144
In London for @fwdcloudsec Europe? We're bringing together a small group of cyber leaders for dinner with the Cotool team - Tuesday 8 September, 6pm at Wild by Tart in Belgravia. Limited seats, request yours early. Link in the comments 👇
1
2
96
If you're using ISO 27001 to guide your security program, check out Bryan O'Neil's latest article, presenting how Tracebit canaries can help provide evidence across 12 controls, whilst supporting 14 additional controls. Read more:tracebit.com/blog/compliance…
61