Michael
Matias

The story

Building the Immune System for the Internet

A soldier from Israel's most elite cyber unit walks into a Zoom call and cannot tell whether the other person on screen is real. That's not a thriller plot. At a Calcalist AI conference in Tel Aviv, Michael Matias ran this exact experiment live - two videos, one real, one synthesized - and even the best algorithmic tools said both were genuine. Only one was.

Matias had already been watching this problem form for years before he co-founded Clarity in late 2022. He'd seen what AI could do from inside Unit 8200, the Israeli intelligence corps that has produced more cybersecurity founders per capita than any institution on earth. He'd worked at Stanford. He'd been a partner at Eric Schmidt's Innovation Endeavors, where "deep learning and cybersecurity" weren't separate conversations. By the time synthetic media became a headline risk, he had a decade of preparation behind a company that was two years old.

Clarity's core proposition is blunt: deepfakes are viruses. They fork. They replicate. They evolve. You need an antivirus - one that updates faster than the threat. So Clarity built dozens of small, highly specialized neural networks, each analyzing a different tell: the way lips fail to sync perfectly, anomalies in ear rendering, artifacts in the voice track, micro-inconsistencies in how AI generates eyes. None of these networks is sufficient alone. Together, they build a forensic consensus that's hard to fool.

The clientele tells you the stakes. News publishers verifying war footage before broadcast. The Israeli government after October 7, when Clarity was deployed by the Ministry of Defense to determine which circulating videos were authentic and which had been manipulated. Identity verification providers protecting hiring pipelines. Large enterprises defending against the new generation of CEO fraud - where a synthetic voice on an audio call authorizes a wire transfer that no real executive ever approved.

Origin

The Making of a Cyber Mind

Israel's Forbes named him to the 18Under18 list in 2013. He was fifteen. By the time most people have their first internship, Matias had already founded Hacking Generation Y - a youth hacking program that put tools in the hands of thousands of teenagers across three continents. He started his first company in high school. He overcame bullying so notable he wrote a book about it later.

Unit 8200 came next. The Israeli Defense Forces' elite cyber and intelligence unit is where some of the world's most consequential security infrastructure gets designed by people in their early twenties. Matias didn't just serve; he built two AI teams that received the National Defense Prize. Twice. Then Stanford, where he studied Computer Science with a focus on AI - not to change careers, but to sharpen a trajectory that was already locked.

At Innovation Endeavors, Eric Schmidt's firm, Matias sat at the intersection he'd been building toward: deep learning meets cyber. He invested. He learned how VC thinks about technology risk. He developed the antenna for which problems were genuine and which were manufactured urgency. Deepfakes registered as genuine.

In parallel, he led AI teams at Hippo Insurance. And he pressed record on Zoom calls during COVID lockdowns and turned those conversations into 20MinuteLeaders - a podcast that would eventually log over 1,000 episodes with founders, operators, and investors from across the tech world.

Timeline

The Career Arc

Clarity

The Antivirus for Synthetic Media

Clarity operates on a basic premise: the same way a biological pathogen forks and mutates, a deepfake model generates variations. The forensic fight isn't about finding one signature - it's about building a detection ecosystem that adapts. Matias describes it directly: "Clarity treats deepfakes as viruses, acting like pathogens that quickly fork and replicate. Its solution was built to replicate and maintain adaptivity and resiliency."

The product covers three modalities: video, audio, and images. On the video side, specialized networks analyze facial geometry frame by frame - lips, nose, ears, eyes - looking for the subtle rendering artifacts that AI generation leaves behind. On audio, the system hunts for anomalies in voice pattern, timing, and compression that distinguish a cloned voice from a live speaker. Images get a different forensic treatment focused on pixel-level manipulation signatures.

The business serves news publishers that need to verify footage before broadcast, government and intelligence agencies that need media authenticity at scale, identity verification providers protecting hiring pipelines from fraudulent candidates, and large enterprises defending executives and brands. Clarity offers a scanning tool via app and API, watermarking services to authenticate legitimate content, and subscription and pay-as-you-go plans for different use cases.

The $16 million seed round - led by Bessemer Venture Partners and Walden Catalyst Ventures, with participation from Secret Chord Ventures, Ascend Ventures, and Flying Fish Partners - landed in February 2024. By then, Clarity had already proved its value in the most consequential context possible: real-time government media verification during an active conflict.

Beyond Clarity

The Side Bets Add Up

Through Matias Ventures, Michael has made 80+ angel investments in deep-tech startups, with 3+ acquisitions and exits. The portfolio includes OxEye (acquired), Flow Security (acquired), Entitle (acquired), Deci (acquired), and active investments in Darrow, BeeHero, Daily.dev, and Tastewise. Seven-plus years of investment experience, built partly while doing everything else.

The podcast - 20MinuteLeaders, later rebranded as The Human Zero Day - started as a COVID project. Matias would get on Zoom calls with innovators and simply ask good questions for 20 minutes. His friends from Stanford and Unit 8200 kept listening. He kept recording. Over 1,000 episodes later, the show has covered founders, operators, and investors from the global tech ecosystem, and the archive sits on Spotify, Apple Podcasts, and his own YouTube channel.

He has also written over 150 articles for a leading Israeli financial newspaper. His book, published in 2020, includes contributions from over 100 CEOs, VC partners, and thought leaders - each chapter mapped to a pivotal moment in his own life. The narrative arc covers starting a company in high school, daily life in the IDF, participating in the performing arts, and recovering from childhood bullying. The title's pun is deliberate: in programming, age is just an integer. And integers don't define what a system can do.

In his own words

What He Actually Believes

The pattern in his public statements is consistent: genuine concern without catastrophism, confidence in the technical fight, and a long-game mentality that treats detection as an ongoing discipline rather than a problem that gets "solved." That's the posture of someone who spent years in intelligence, not someone who showed up in AI security because it's a growth market.

Recognition

Awards & Milestones

Explore further

Links & Resources

Watch

Michael Matias on Video

Matias's TEDx talk at Gunn High School - on why teenagers should run startups - was recorded years before the Forbes recognition and the $16M round. It captures the same voice: direct, unpretentious, specific.