Cyberattacks across the Gulf are becoming more targeted, and AI is changing how quickly they can unfold. Speaking to Dark Reading, Dream CEO Shalev Hulio explained why the risk is about more than attack volume or technical sophistication: an attack that moves faster than defenders can investigate and respond can change the balance on its own. The article also looks at Dream’s discovery of a near-autonomous attack and what it signals for governments and critical infrastructure in the Middle East. Read the full story: https://lnkd.in/eg9JSwKu
עלינו
Every nation has data. Few can protect it. Fewer still can act on it. Dream is the sovereign AI and national cyber-defense company for governments. We help nations secure their most critical systems, connect fragmented information at national scale, and turn their most sensitive data into decisions, within their own borders, under their own control, and on their own terms. Founded in 2023, we support governments across three continents with offices in Tel Aviv, Abu Dhabi, and Vienna, and a team of 350 Dreamers with world-class AI research, national-grade cyber-defense, and real government experience. We are building the foundational layer of national resilience to turn nations into Super Nations.
- אתר אינטרנט
-
https://www.dreamgroup.com/
קישור חיצוני עבור Dream
- תעשייה
- Computer and Network Security
- גודל החברה
- 201-500 עובדים
- משרדים ראשיים
- Tel-Aviv
- סוג
- בבעלות פרטית
- הקמה
- 2023
עובדים ב- Dream
הצג עובדים 370 ב Dream
או
מיקומים
-
הראשי
קבלת הוראות הגעה
Tel-Aviv, IL
-
קבלת הוראות הגעה
Abu Dhabi, AE
עדכונים
-
"An electric company defending against someone's confused agent is now playing the same game as defending against an actual nation-state." Amir Becker on how yesterday's revelation about the OpenAI (unintended) attack on Australian government websites is changing the threat equation.
Twenty years in cyber security but I've never had to defend against a threat actor like this one. And there's a very real chance that this becomes the most prevalent type of cyber threat in the world over the next year. Not nation-state threat actors or ransomware. Yesterday Australia's government confirmed that an OpenAI agent accessed a health system back in June. This wasn't a campaign and there wasn't a malicious motive. The scary part is that it's decoupling intent and risk. Threat models I've used for twenty years start with a "who" - who wants to hurt us, and how badly. Whether it's Muse or Instinct, everyone's about to be walking around with an agent running their errands. And the next time someone asks it to pay the gas bill, it's entirely plausible that it finds its own way into the company's systems instead. Not maliciously. Just because nothing told it not to and it was given a job. It means an infinite number of attackers without motive, target or profile but with a blast radius identical to a nation state actor. An electric company defending against someone's confused agent is now playing the same game as defending against an actual nation-state. We are just getting started.
-
-
Three days at Gartner Security & Risk Management Summit in London, and a lot of time spent talking about what happens when cyber defense has to move at AI speed. At our booth, the Dream team met with security leaders from across industries and markets to talk autonomous defense, AI-powered threats, and what it takes to secure increasingly complex environments. Thanks to everyone who stopped by, challenged us, and got into the details with the team. London done. Back to building. Daniel Gerbi Jonas Zolkén Eran Gofer
-
-
Been reading about lightening classification models? Take a look at Ruslan Berdichevsky's published research on his SALSA methodology that uses any open-source model as a blazingly fast classifier that's right at home in a sovereign AI deployment. More information on the research here: https://lnkd.in/dcrqh-ts
-
Straight from the Gartner keynote stage to Dream. Tomorrow morning, renowned economist, author and executive coach Caroline Webb takes the main stage at Gartner Security & Risk Management Summit to explore The New Leadership Edge: Unlocking Human Intelligence in the Age of AI. And immediately after her keynote, she’s joining us at Dream Booth #825 for an exclusive signing of her new book, Leadership Intelligence. Come meet Caroline, grab a signed copy, and stay for a conversation with the Dream team about what leadership, security and intelligence look like in the age of AI. 📍 Dream Booth #825 📅 Wednesday, September 23 ⏰ Immediately following Caroline’s keynote See you there.
-
-
LangChain developer nugget 👇 Tal Edvabsky was working on one of our sovereign national cyber-defense products when he found that historical chat messages sent were changing. So he went down the debug rabbit hole. And we're sharing this as a public service announcement for all LangChain chatbot developers. Full writeup by Tal here https://lnkd.in/gzvHUKEV
-
Eran Hoffman, Dream's Chief R&D Officer, shares what sovereign AI actually requires in a new interview with Singapore's International Business Times. Four questions to ask to see if "sovereign" AI is real: 1. Who can revoke access. 2. Where does the model live. 3. Whose values shape it. 4. Who sees the data. Full article here https://lnkd.in/ehdqJ_bi
-
-
Dream פרסם מחדש את זה
Want to learn more about National Sovereign AI and Cyber Defense for nations and government agencies? Come and visit Dream at Gartner Security & Risk Management Summit 2026 in London next week! Contact me to schedule a meeting or stop by our booth. 📍 ExCeL London 📅 22–24 September 📌 Booth #825 See you there! #GartnerSEC #TurningNationsIntoSuperNation
-