Exclusive: Cracking The Only Engineering System That Has Ever Worked

Embryos know something we don't, but Polyphron thinks AI can help us learn

In late 2023, Matt Osman went for a preventative MRI.

His life was good. His AI company, Treat, was humming, and he was looking to start a family. He was young, active, and had no medical issues. He had no reason to expect anything was wrong. Still, as a precaution, he decided to get some tests done.

Not long after the MRI, Osman was walking down Orchard Street in the Lower East Side of New York City when his phone buzzed. An e-mail from the screening clinic arrived first and then, ten seconds later, a call. They never call with good news. The radiologist saw a complex mass on his pancreas, the size of a golf ball. The pancreas isn’t a good place to find anything, and the clinic recommended he see an oncologist immediately.

What followed was a year of unbearable uncertainty. Biopsies were inconclusive, and two surgical oncology teams couldn’t agree on what to do next. One surgeon wanted to resect immediately, another warned of dangerous complications that could damage enough pancreatic tissue that he’d almost certainly become diabetic. With a split ticket, Osman opted to monitor, watching for changes in shape or size. If it stayed put, the likelihood of it being benign would ratchet up over time.

It did stay put. Osman is now in the clear. But the experience rewired something inside of him. “It was the thing that started me on the journey of realizing that human tissue is one of the most valuable substances on earth,” he told me, “and that we now have a plausible path to making it on demand.”

That path is Polyphron, the company Osman co-founded with Fabio Boniolo, a Harvard-trained computational biologist. They’re building a platform to grow functional human tissue—not by painstakingly reverse-engineering each tissue type, but by training artificial intelligence models on developmental biology and letting them learn what embryos already know how to do. Success here could mean another viable path to human life extension.

Ultimately, the bulk of human morbidity and mortality isn’t whole limb loss or sepsis. What makes us sick is scarred hearts after cardiac events, damaged lungs from smoking, or sclerotic kidneys that can’t filter like they used to. Polyphron is simply asking whether it’s possible to get to those damaged organs earlier, with smaller interventions, matched to the patient, without waiting for catastrophe.