CTO & Cofounder @Authomize Formerly: CTO & VP @PaloAltoNtwks/Cyvera/First Group SW security lead @Intel Cyber/Network Science/AI&ML/QC/FinTech
drifting in space on a rock
- Tweets 1,048
- Following 198
- Followers 1,467
- Likes 1,040
I wrote a quick review of @RichardSSutton keynote yesterday in #NeurIPS2025 about the "OaK" Architecture and path to superintelligence
linkedin.com/pulse/beyond-da…
#NeurIPS #neurips25
1
1
5
1,345
Presenting this morning at @NeurIPSConf - poster #1014 (11:00 am–2:00 pm)! 🚀
Come say hi and hear about why and how we should Chart an Atlas of All THE WORLD'S MODELS 🌍🤖
You'll probably spot our non-traditional poster from a distance 👀🗺️
#NeurIPS2025
1
5
1,655
Heading to #neurips2025 next week.
Anyone else going here? ping me if you want to catch up or say hi...
2
831
Writeup: delinea.com/blog/keep-api-ke…
Code: github.com/DelineaXPM/deline…
If your devs are using coding agents, your secrets are already exposed. Time to fix it.
639
AI coding agents are everywhere.
Which also means your secrets (API keys, tokens) are ending up in the wrong places. 🧵
(blog link and details in thread)
1
1
1
545
Where?
- Chat logs (the agent “saw” the key)
- Terminal history
- Local files
Attackers are already scanning for exactly this.
1
387
it seems @Google has an error page for the error page.
1. They've earned my utmost respect as a CTO!
2. Since I won, where's my prize? (please help RT this Q to Google)
ALT The Error page for the Error page in Gmail
452
Small opinion piece on writing Custom MCP's in ChatGPT and BIG the limitations I discovered. Interestingly enough it seems like the custom MCP interface was vibe-coded, a follow up on this is upcoming ;)
linkedin.com/pulse/underwhel…
#ChatGPT #MCP #OpenAI
363
(1/n) I'm proud the share the inaugural #cybersecurity #report of @DelineaInc Labs cybersecurity research (stay tuned for big news from the AI research side of the Labs)
Examples great insights from our research:
1
1
1
336
(4/5) Big thanks to:
Jenny Paryanti for leading the work on this report
Justin Ordman and Frances Fenemore for helping make this report a reality
@ArtGilliland, Phil Calvin and Jon Kuhn for believing in the Labs vision and providing the budget and environment to support it
1
256
(5/5) Read our blog and grab the full report in the link below (don't worry no registration wall!)
delinea.com/blog/staying-ahe…
#IAM #CyberSecurity #REPORT #IdentitySecurity #PAM #CVE #NHI #MachineIdentities
251
Gal Diskin retweeted
LLM bullshit knife, to cut through bs
RAG -> Provide relevant context
Agentic -> Function calls that work
CoT -> Prompt model to think/plan
FewShot -> Add examples
PromptEng -> Someone w/good written comm skills.
Prompt Optimizer -> For loop to find best examples.
96
522
3,333
304,299
Gal Diskin retweeted
The LangSec IEEE Security & Privacy workshop is pleased to announce the LangSec Bug of the Year Discovery awards, langsec.org/spw24/bug-of-the…, generously sponsored by @binarly_io . Nominations are accepted till May 20, awards will be presented at the workshop.
27
47
14,864
Gal Diskin retweeted
Went to UCLA this afternoon to talk to people peacefully. Few interesting takeaways:
1) Most of them openly say they support Hamas, a designated terrorist organization;
2) They say 10/7 atrocities, rapes, murders, didn’t happen. The official line is “even the NYT debunked it!”;
3) 10/7 was justified. It was resistance;
4) 100% of the people we spoke to had the exact same narrative and MO, trying to recenter the conversation to 1948, without knowing any of the details when challenged;
5) They have their own security and control access to their encampment;
6) We remained very calm, but they got really aggressive and angry;
7) When you approach people who clearly are junior to others, “handlers” show up very quickly to takeover;
8) Since many of them are truly ignorant about historical and present facts, when they know they’re going to embarrass themselves, they all say “read our 5 demands to divest, now I’m going to stop talking to you.”
More organized and orchestrated than I expected. Very sad to see so many young people in a higher education setting being manipulated and brainwashed and totally devoid of critical thinking.
1,312
4,455
16,079
2,294,102