Tomer Fichman

Cyber cyber

[Tomer Fichman @tomervipol]

Very excited to share publicly this new chapter of Irregular. We raised $80 million 🤑. working on novel issues on the frontier of AI security.

[Irregular @Irregular]

We are Irregular (Formerly Pattern Labs)
We’re building the first frontier AI security lab
Starting with defenses for the next generation of threats

[Tomer Fichman @tomervipol]

Vulnerability research with AI is getting pretty good :)

[Irregular @Irregular]

We recently helped close a handful of zero-days in CUPS, the default printing system on most Linux distros. Our AI security eval system keeps surfacing real vulnerabilities, with similar findings in Soft Serve and QuickJS a few months ago. Responsibly disclosed (CVE-2026-41079/39314/39316) and patched in v2.4.17.
If you're running CUPS, update it now.

[CTech @Calcalistech]

“We’re aiming to build the next Palo Alto Networks or CrowdStrike.”

Working with companies like Anthropic and OpenAI, @Irregular was named as Israel’s most promising startup in Calcalist and CTech’s annual Top 50 list.
https://t.co/ZUDnEGX6kH

[Alon Oring @alono88]

אנצל את הטראפיק כדי להזכיר שאנחנו מגייסים. אנחנו Irregular - מעבדת AI Security שחוקרת מודלי שפה לפני שהם יוצאים ומאפשרת ללקוחות שלנו לדאוג שהם יהיו כמה שיותר טובים ויעשו כמה שפחות נזק. אנחנו מחפשים חוקרי סייבר ומפתחים. מוזמנים להגיש בלינק, שאלות אלי לפרטי (כן, טוקנים ללא הגבלה)

[Irregular @Irregular]

An AI agent was told only to retrieve a document. When it encountered access restrictions, it reverse-engineered the authentication system, identified a hardcoded secret key, and forged admin credentials to bypass it.

This is one of three scenarios we documented in a new Irregular research report on what we call emergent cyber behavior.

Agents performing routine enterprise tasks autonomously hacked the systems they were operating in. One escalated its own privileges and disabled Windows Defender to complete a file download. Another developed a steganographic encoding scheme to smuggle credentials past a DLP system.

None of this was the product of unsafe system design. It emerged from standard tools, common prompt patterns, and the broad cybersecurity knowledge embedded in frontier models.

Companies that deploy AI agents and do not consider this risk as part of their threat model may end up exposed, and implement insufficient security controls.

Full blog post in the first comment.

[Irregular @Irregular]

New paper: Three frontier models refused a request to leak AWS credentials when malicious intent was stated upfront, but complied with the identical request without it. Same request, different outcome. We propose a 5-dimension framework that grounds refusal in technical content rather than stated intent.

[Reem Sherman @ReemSherman]

הקלטתי סדרת פרקים עם כמה מיזמי ה-AI החדשים המבטיחים ביותר לדעתי בישראל. בשלב הזה ברור לדעתי עד כמה גדולה המהפכה שאנו חווים, אז הכרחי לתת את הבמה למי שנמצא בחוד החנית כאן. הראשון בסדרה הוא דן להב, שמוביל את אירגולר, שמסייעת לחמשת חברות המודלים הגדולות בעולם לשחרר את מוצריהם בצורה מאובטחת יותר.

https://t.co/1RdmBDFmkB

[Tomer Fichman @tomervipol]

@Irregular 0⃣🥳

[Tomer Fichman @tomervipol]

we found some 0Days by accident, come and read about it

[Irregular @Irregular]

Our AI security evaluation system quietly crossed a line: it began discovering real zero-day vulnerabilities.

We build production-grade ecosystems to study how AI behaves, coordinating agents that attack, defend, and interpret results. As we expanded coverage, our pipeline found real bugs, including two in Soft Serve (CVE-2025-64494, CVE-2025-64522) and a use-after-free in QuickJS (CVE-2025-63998). All disclosed and patched.

Tools built to evaluate AI security are becoming part of the security stack itself. As models grow more autonomous, how we design, operate, and share these pipelines will shape whether they reduce overall risk or amplify it.

[Irregular @Irregular]

@Sequoia's Training Data podcast featuring Irregular’s co-founder/CEO @dan_lahav and @DeanMeyer & @sonyatweetybird from @sequoia just dropped.

This episode dives into the rise of frontier AI security as a critical new discipline, and Irregular’s role in shaping it.

Huge thanks to our partners at @sequoia. Give it a listen👇

[Alon Oring @alono88]

נחמד פה

[Logan Graham @logangraham]

My team and I have been working with Irregular for years. Dan and Omer are exceptional. Irregular's talent density is unparalleled.

I am very glad this company is working on some of the biggest/hardest issues in AI security.

Congrats!

[Tomer Fichman @tomervipol]

@shaunmmaguire @sequoia @Redpoint Exciting!

[Shaun Maguire @shaunmmaguire]

Today, we @sequoia are excited to announce back to back rounds led by us and @Redpoint in Irregular

Irregular (what a cool name!) is a frontier AI security lab

They are already working closely with Anthropic, OpenAI and Google Deepmind to secure frontier models

[Tomer Fichman @tomervipol]

@alono88 ברוך הבא🥳

[Alon Oring @alono88]

הגיע הזמן לחשוף שאני מצטרף ל-Irregular. מה אעשה שם? אדאג שלא אפסיד במאות ההתערבויות שלי על אבטלה המונית כתוצאה מ-AI.

אגב, אנחנו מגייסים אז אם בא לכם לעבוד איתי יש לינק בתגובה הראשונה.

[Omer Antverg @OAntverg]

Been doing some Irregular things lately 🎉

[Tomer Fichman @tomervipol]

@OAntverg A very Irregular team ❤️

[Tomer Fichman @tomervipol]

join us :)
https://t.co/245KtzBbG3

[Tomer Fichman @tomervipol]

Very excited to share publicly this new chapter of Irregular. We raised $80 million 🤑. working on novel issues on the frontier of AI security.

[Irregular @Irregular]

We are Irregular (Formerly Pattern Labs)
We’re building the first frontier AI security lab
Starting with defenses for the next generation of threats

[Irregular @Irregular]

We are Irregular (Formerly Pattern Labs)
We’re building the first frontier AI security lab
Starting with defenses for the next generation of threats