June 29 – July 2, 2026 · San Francisco, CA · imported from ai.engineer's public schedule feed

AI Engineer World's Fair 2026 — unofficial import demo

Unofficial demo. This programme was imported from the AI Engineer World's Fair's own public schedule feed to show vibeboard at real conference scale. Not affiliated with, or endorsed by, the organisers.

All sessions
Data QualitySession

Training Frontier Models to Out-Think Hackers

Uri Rolls, Thom Wolf

When
Wednesday, July 111:40 AM – 12:00 PM · 20 min
Where
Track 9San Francisco, CA · imported from ai.engineer's public schedule feed
Google Calendar

About this session

We will give a surprisingly optimistic talk about AI and cyber, and why we believe it is not the end of cybersecurity as we know it, but an opportunity to empower defenders and build a lasting edge over attackers.

Cyber is a battle of skill and speed, and the rising tide of frontier models is allowing human attackers to move faster and cheaper. That combination of skilled hackers and breakthrough LLMs is a real threat, while defensive systems are still expected to operate at scale with limited human intervention, constrained by what models can do out of the box. But the answer is not fear or despair. Just as high-quality data transformed software engineering, the right cyber training data can teach models to turn from weapons being used against us into tools that protect us.

Speakers (2)

Uri Rolls
Uri Rolls

CEO, Arithmetic

Uri Rolls is Co-Founder and CEO of Arithmetic, an AI data company building cyber worlds for frontier models. Just as AI transformed software engineering, Arithmetic is working to bring the same leap to cybersecurity by creating post-training data from real vulnerabilities, teaching models to reason across black-box search spaces, turn reconnaissance into exploitation, and take correct defensive action. Previously, Uri conducted black-hole imaging research at Harvard and the Smithsonian, and spent several years in Israeli Intelligence.

Thom Wolf
Thom Wolf

Co-founder and CSO, Hugging Face

Co-founder and Chief Science Officer of Hugging Face. Previously physics research and attorney.

More in Data Quality