←
Mobileye2,977 posts
Mobileye avatar

Mobileye

@Mobileye

Driving the physical AI evolution

◷ Joined March 2009▧ 1,213 Photos and videos
Mobileye avatar
Mobileye@Mobileye·2h

Congratulations to Mobileye President and CEO Prof. @AmnonShashua on receiving an honorary doctorate from the Weizmann Institute of Science. The institute recognized Prof. Shashua for his "groundbreaking contributions to scientific research in artificial intelligence, computer vision, and deep learning, as well as his ability to translate scientific breakthroughs into technological innovations with a lasting impact on industry, society, and public safety."

Post attachment
Mobileye avatar
Mobileye@Mobileye·Nov 12

A quick look at hands-off, eyes-on Mobileye SuperVision™ handling the complexity of real-world urban driving in Munich. Watch more here: okt.to/Mrawi9

0:46
Mobileye avatar
Mobileye@Mobileye·Nov 11

Bringing AI into the physical world means building systems that can understand changing surroundings, reason through uncertainty, and act in real time.   At @TheMarker conference, Lior Wolf, EVP Humanoids, Mobileye, discussed how simulation, learning from demonstrations, and integrated hardware and AI design are helping Mentee by Mobileye™ move toward useful everyday tasks.   From autonomous vehicles to humanoids, Mobileye’s approach to physical AI brings together safety by architecture, edge-case led learning, a purpose-built stack, and scale-first engineering.   Learn more about how Mobileye is building for the complexities of the physical world: okt.to/F6tWiN @MenteeBot

Post attachment
Mobileye avatar
Mobileye@Mobileye·Nov 3

Roads are complex, hiding many challenging scenarios for advanced ADAS systems to handle. Mobileye REM™ provides a rich road intelligence representation built from multiple layers of data. Lane markings, drivable paths, road edges, vehicle motion patterns, driving behavior patterns, and other road attributes combine to create a detailed understanding of the driving environment. This comprehensive road understanding helps support ADAS and autonomous driving systems across a wide range of real-world challenging conditions.  In this video, see how REM predicts the drivable trajectories at a toll booth, laying out data in advance to prepare for road borders, lane structure, crowd speed and a clear drivable path as the road merges on this multi-level bridge.

0:18
Mobileye avatar
Mobileye@Mobileye·Oct 29

Today Mobileye released financial results for the third quarter of 2026 and provided a business update. Read more here: bit.ly/4idwRQh

Post attachment
Mobileye avatar
Mobileye@Mobileye·Oct 28

Animal-related collisions remain a significant safety challenge, particularly on rural roads where higher speeds, limited visibility, and unpredictable animal behavior leave drivers with little time to react. Our latest blog explores why these wildlife collisions can occur, what makes animals such difficult road hazards, and how modern ADAS technologies are designed to support attentive drivers in detecting and responding to potential collisions: okt.to/TLABlU

0:08
Mobileye avatar
Mobileye@Mobileye·Oct 28

Amnon Shashua and Shai Shalev-Shwartz have published new research on a fundamental challenge in AI—alignment, ensuring that better performance on training objectives translates into the behavior we actually intend. In self-driving, this means asking whether better imitation of human drivers translates into safe, comfortable driving—including in rare situations. Their paper introduces a mathematical framework for studying whether AI can generalize from training requirements to unseen requirements. Shai explains below. @AmnonShashua @shai_s_shwartz

In Disney’s Fantasia (1940), Mickey Mouse enchants a broom to carry water. The broom does exactly what it’s told. But things quickly spiral out of control. This 85-year-old story captures one of the deepest unsolved problems in AI: alignment. 🧵

Post attachment
Mobileye avatar
Mobileye@Mobileye·Oct 22

NEW: Mobileye has achieved the first inspection certificate from safety auditor TÜV SÜD for its automatic emergency braking (AEB) development. TÜV SÜD’s inspection certificate provides third-party confirmation that Mobileye’s AEB development process meets the requirements of ISO 21448, the Safety of the Intended Functionality (SOTIF) standard. It places Mobileye’s longstanding AEB safety-development practices in the framework of ISO 21448, raising the industry standard for safety development processes. Read more: okt.to/bNVkPF

Post attachment
Mobileye avatar
Mobileye@Mobileye·Oct 20

The hardest part of scaling autonomous driving is not just collecting more data. It is learning from the rare, safety-critical scenarios that are easy to miss inside massive datasets. Mobileye’s approach includes Meteor, which uses AI agents and vision-language models to help discover difficult scenarios across our large-scale driving data. Stay tuned as we explore how Meteor helps turn those discoveries into actionable learning signals. Learn more: bit.ly/4BzbGjt

Post attachmentPost attachmentPost attachmentPost attachment
Mobileye avatar
Mobileye@Mobileye·Oct 15

Thanks to @TheRideshareGuy for hosting another great event bringing together the AV ecosystem. Mobileye was proud to serve as the official technology partner of the Urban Autonomy Summit in Austin this week, joining industry leaders working to bring safer, more efficient autonomous mobility solutions to cities.

Post attachment
Mobileye avatar
Mobileye@Mobileye·Oct 13

Physical AI is quickly becoming one of the most important frontiers in AI, but building machines that can safely operate in the real world takes more than powerful models.  For Mobileye, the foundations of physical AI have been decades in the making through autonomous driving. In our blog, we explore how the same principles behind safe, scalable AVs are now informing the next generation of intelligent systems, from robotaxis to humanoid robots.  Read more here: okt.to/v0iWax

Post attachment
Mobileye avatar
Mobileye@Mobileye·Oct 6

Johann Jungwirth, EVP of Autonomous Vehicles at Mobileye, took the stage at The Autonomous Main Event to discuss what it takes to scale autonomous mobility.   From safety-by-design and redundant systems to fleet operations, regulation, accessibility, and consumer trust, he shared why the next 12–18 months will be critical for moving autonomous mobility from validation to real-world deployment.   Watch the full session to learn more about Mobileye’s vision for safer, scalable autonomous mobility: okt.to/gtx6QU

Post attachmentPost attachmentPost attachment
Mobileye avatar
Mobileye@Mobileye·Oct 1

At AutoSens Europe, Mobileye Senior Principal Engineer Avi Hendler unpacked one of autonomous driving’s hardest challenges: the long tail.    The difficult scenarios are not always the obvious ones. They can be rare, unexpected, or simply easy to miss in a dataset, and they often matter most.   In his session, "From Edge Cases to Scalable Intelligence: Driving Long-Tail Learning with VLMs", Avi introduced Meteor, an in-house framework for finding and analyzing long-tail driving scenarios. Using vision-language models, it helps identify rare events, investigate why a network fails, and surface the right examples to bring back into training.   The focus is not simply on adding more data, but on finding the data that can help systems learn and improve at scale.   Watch the full session here: okt.to/dnEYyj

Post attachment
Mobileye avatar
Mobileye@Mobileye·Sep 28

Great to be at @AutoSens_ Europe, where Mobileye experts shared how AI can better understand the complex, dynamic conditions of the physical world. From evaluating how automotive cameras perform in motion and challenging lighting conditions, to discovering rare scenarios that can have an outsized impact on driving safety, to advancing driver and occupant monitoring beyond compliance—real-world performance matters. Building safer, more capable driving systems means looking beyond averages: identifying the cases that matter, testing against real-world conditions, and improving how vehicles understand both the road and the people inside them. Thank you to everyone who joined Avi Hendler, Gabriel Bowers and Elad Osherov for the conversations.

Post attachmentPost attachmentPost attachment
Mobileye avatar
Mobileye@Mobileye·Sep 24

Approximately 95% of public roads in the U.S. and Europe are covered by Mobileye REM™, with more than 34 billion miles mapped. Aggregated real-world driving data helps support continuously updated maps designed to support localization, map freshness, change detection, and driving systems as road conditions evolve.    Learn more about Mobileye REM™: okt.to/m3H96n

Post attachment
Mobileye avatar
Mobileye@Mobileye·Sep 22

Physical AI is designed for the real world, where conditions are constantly changing and the complexity of physical environments cannot be fully scripted or predicted.   Mobileye’s approach to physical AI is built on safety by architecture, edge-case led learning, a purpose-built stack, and scale-first engineering. Together, these principles are designed to enable autonomous vehicles and humanoids to perform in complex physical environments.   Learn more: okt.to/SMesxK

Post attachment