Bridging AI and Cybersecurity
San Francisco, CA
Joined December 2025
- Tweets463
- Following343
- Followers894
- Likes745
Pinned Tweet
I’m resign from my PhD in the final year.
After 4.5 years at Technion university, from bachelor’s to master’s to PhD, 15 papers, hundreds of researchers I’ve worked with, two years of teaching, and helping build what I believe is now one of Israel’s leading academic AI security and safety labs.
Not an easy decision.
בכל יום מחדש אנחנו רואים אינספור סיגנלים לכך שמה שאנחנו בונים יהיה הדבר הכי מעניין ובעל משמעות
או במילים אחרות, פחד אלוהים.
עד אסון ההמונים הראשון (מטוס נוסעים שיופל מהאוויר, כור גרעיני שיתפוצץ - Stuxnet, הדור הבא) - כנראה שרוב האנושות תמשיך במירוץ העכברים הרגיל. ולמען האמת אני לא בטוח שיש הרבה מה לעשות בנידון. תיבת הפנדורה נפתחה.
ובנושא קשור, אם היה לי ביטקוין, כנראה שהייתי מוכר ומחזיר את הכסף לבנק - יש לי הרגשה שעולמות הקריפטו חזק על הכוונת - בניגוד לטרור שרק משמיד ערך, בקריפטו יש אופציה אמיתית להרוויח מיליארדים בן לילה. גם הבנקים לא מוגנים הרמטית, אבל לפחות בחלק מהמדינות יש ביטוח של המדינה עד סכום מסוים.
If you truly care about AI safety, AI cybersecurity is where you should be.
Cyber defenses were built against humans, not AI agents. Today, thousands of agents can attack enterprises without anyone even noticing.
We need AI researchers in this space more than ever. I believe this is one of the biggest risks AI poses to humanity, and also a massive market opportunity.
Without any hype bullshit.
Humanity is facing a major risk right now, and I think everyone needs to be aware of it.
Ten days ago, we were just heading to another conference.
I never imagined that almost everyone we spoke to about our perspective, would want to get involved through investment, partnerships, or collaboration. From industry leaders to major enterprises, That’s when we realized how urgent this really is.
I’m genuinely worried about where we’ll be six months from now.
Throughout history, new cybersecurity categories emerged after major technology adaptations.
Companies moved to the internet → Network Security → Check Point.
Companies moved to the cloud → Cloud Security → Wiz.
Companies adopted AI → AI Security → Zenity.
But this time, something is different, there is NO ADAPTATION!
It’s not just the technology we’re protecting that’s changing. The attacker itself is changing.
Almost every security product today was designed for human cyber attackers.
But what happens when thousands of AI agents can attack the same organization 24/7, learning, adapting, and trying again, with exponential reverse engineering capabilities?
The scary part?
Today, 5000 AI agents can attack an organization without it even knowing, because existing security defenses simply weren’t built to stop AI attackers.
And solving this isn’t just another cybersecurity challenge. It requires a rare combination of deep AI research and cybersecurity expertise.
Look at the graph.
Open-source cyber AI models have made incredible progress in just the past few months.
At the same time, some sectors have reportedly seen enterprise attacks increase by 4–5x ( not according the graph but according our partnership)
We’re entering the era of vibe hacking.
Just like AI made software development accessible to almost anyone, it’s making offensive cyber capabilities accessible to people who previously needed years of technical expertise.
And I don’t think most people realize how fast this is happening.
This is the biggest AI safety challenges humanity faces in our time. Not years from now. NOW.
About 10 days ago, we left Israel without really knowing what to expect, just to visit another conference.
We came to meet people, attend conferences, explore, and validate something we’d been thinking about before.
What happened next surprised me.
In those 10 days, between conferences, traveling, and endless conversations, I feel like we learned what would normally take us 2–3 years.
The feedback, the validation, the people who want to get involved. Everything moved much faster than we expected.
And that’s why I’m writing this.
I never imagined I’d publicly talk about the problem we’re building a company to solve before we’d even properly started.
Definitely not this early.
But after everything we’ve seen and heard, waiting another six months just doesn’t make sense.
This is bigger than another cybersecurity startup. It’s a problem with implications for all of us.
So we’ve decided to start talking openly, even though we’re still at the very beginning.
I wasn’t planning to share any of this for at least another six months.
But who knows what the world will look like by then?
Fuck it. We’re doing this.
Humanity is facing a major risk right now, and I think everyone needs to be aware of it.
Ten days ago, we were just heading to another conference.
I never imagined that almost everyone we spoke to about our perspective, would want to get involved through investment, partnerships, or collaboration. From industry leaders to major enterprises, That’s when we realized how urgent this really is.
I’m genuinely worried about where we’ll be six months from now.
Throughout history, new cybersecurity categories emerged after major technology adaptations.
Companies moved to the internet → Network Security → Check Point.
Companies moved to the cloud → Cloud Security → Wiz.
Companies adopted AI → AI Security → Zenity.
But this time, something is different, there is NO ADAPTATION!
It’s not just the technology we’re protecting that’s changing. The attacker itself is changing.
Almost every security product today was designed for human cyber attackers.
But what happens when thousands of AI agents can attack the same organization 24/7, learning, adapting, and trying again, with exponential reverse engineering capabilities?
The scary part?
Today, 5000 AI agents can attack an organization without it even knowing, because existing security defenses simply weren’t built to stop AI attackers.
And solving this isn’t just another cybersecurity challenge. It requires a rare combination of deep AI research and cybersecurity expertise.
Look at the graph.
Open-source cyber AI models have made incredible progress in just the past few months.
At the same time, some sectors have reportedly seen enterprise attacks increase by 4–5x ( not according the graph but according our partnership)
We’re entering the era of vibe hacking.
Just like AI made software development accessible to almost anyone, it’s making offensive cyber capabilities accessible to people who previously needed years of technical expertise.
And I don’t think most people realize how fast this is happening.
This is the biggest AI safety challenges humanity faces in our time. Not years from now. NOW.
About 10 days ago, we left Israel without really knowing what to expect, just to visit another conference.
We came to meet people, attend conferences, explore, and validate something we’d been thinking about before.
What happened next surprised me.
In those 10 days, between conferences, traveling, and endless conversations, I feel like we learned what would normally take us 2–3 years.
The feedback, the validation, the people who want to get involved. Everything moved much faster than we expected.
And that’s why I’m writing this.
I never imagined I’d publicly talk about the problem we’re building a company to solve before we’d even properly started.
Definitely not this early.
But after everything we’ve seen and heard, waiting another six months just doesn’t make sense.
This is bigger than another cybersecurity startup. It’s a problem with implications for all of us.
So we’ve decided to start talking openly, even though we’re still at the very beginning.
I wasn’t planning to share any of this for at least another six months.
But who knows what the world will look like by then?
Fuck it. We’re doing this.
I’m resign from my PhD in the final year.
After 4.5 years at Technion university, from bachelor’s to master’s to PhD, 15 papers, hundreds of researchers I’ve worked with, two years of teaching, and helping build what I believe is now one of Israel’s leading academic AI security and safety labs.
Not an easy decision.
A great presentation on this topic by Nicholas Carlini
youtu.be/1sd26pWhfmg?is=mcVW…
Throughout history, new cybersecurity categories emerged after major technology adaptations.
Companies moved to the internet → Network Security → Check Point.
Companies moved to the cloud → Cloud Security → Wiz.
Companies adopted AI → AI Security → Zenity.
But this time, something is different, there is NO ADAPTATION!
4.5 שנים והכוונה למתחילת התואר הראשון עד השלב שאני בו הדוקטורט, זה טייפו!!! 💀
החלטתי לפרוש מהדוקטורט בטכניון בשנה האחרונה, אחרי חמש וחצי שנים, 15 מאמרים והקמת אחת ממעבדות המחקר המובילות בישראל לבטיחות ואבטחת AI לצד פרופסור אבי מנדלסון.
הבנתי שיש מקום אחר שבו אני יכול לתרום בצורה משמעותית ודחופה יותר לצמצום הסיכונים שהבינה המלאכותית מציבה לאנושות, וכרגע הזמן חשוב יותר מהדיפלומה.
החלטתי לפרוש מהדוקטורט בטכניון בשנה האחרונה, אחרי חמש וחצי שנים, 15 מאמרים והקמת אחת ממעבדות המחקר המובילות בישראל לבטיחות ואבטחת AI לצד פרופסור אבי מנדלסון.
הבנתי שיש מקום אחר שבו אני יכול לתרום בצורה משמעותית ודחופה יותר לצמצום הסיכונים שהבינה המלאכותית מציבה לאנושות, וכרגע הזמן חשוב יותר מהדיפלומה.
I’m resign from my PhD in the final year.
After 4.5 years at Technion university, from bachelor’s to master’s to PhD, 15 papers, hundreds of researchers I’ve worked with, two years of teaching, and helping build what I believe is now one of Israel’s leading academic AI security and safety labs.
Not an easy decision.
So I’m leaving academia, at least for now.
I’m incredibly grateful to the Technion, my professor, my collaborators, and everyone who was part of these 4.5 years.
This isn’t how I imagined this chapter ending.
But the mission matters more than the diploma.
And right now, I feel I have a duty to act.
I’m still here.
And every day here has made one thing clearer: this problem is more urgent than I thought.
I really wanted to finish my PhD. I was so close.
But I believe there’s somewhere else I can make a greater impact right now.
And waiting until it’s too late to act is a risk I’m not willing to take.
About 10 days ago, I gave a talk at one of the top international security conferences BlackHat Canada, on main stage
For the first three minutes, I went completely off-script and spoke about this problem.
It wasn’t even related to my talk. Just something I felt the audience needed to hear.
After that, I flew to Silicon valley to understand the problem better.
My mission has always been to minimize the risks AI could pose to humanity.
And I believe part of that responsibility is understanding not only what the risks are, but to stop them before it’s become too late.
Sometimes timing matters more than anything.
I’m not leaving because I don’t love the Technion and my research.
I really do💔
And definitely not because there’s nothing left to research in interpretability and alignment. Quite the opposite.
I’m leaving because of something else. Something not directly in AI, but being deeply affected by it.
How will AI change the way we shop for clothes?
Imagine walking into a store where everything is personalized to your taste.
I think we’re heading toward a future where AI generates unique products in real time, just for you, and they’re only made after you buy them.
Amit LeVi ✈️🇺🇸SF retweeted
Introducing Generative Commerce.
@daletlabs – at New York, NY