Summio

YouTube video

Most Businesses Will Fail AI Transformation. Andrey Doronichev on the Future of Technology

A Conversation with Andrey Doronchev on Breakthroughs in Biotech, the Future of AI, Entrepreneurship, and the Search for Meaning in Life.

9 min read

Available in

Summary preview

AI, Biotech, and Reinventing Life: The Journey of Andrey Doronchev

Hey there! Today, we've got someone on the show whose career is basically one big experiment, a constant chase for paradigm shifts. Andrey Doronchev is a serial entrepreneur who kicked things off by installing pirated Windows at 13 and grew up to be a key player behind mobile YouTube and Android. He spent 14 years at Google, famously turned down $10 million in stock options, but never stopped launching his own ventures across everything from fintech to gaming. Now, his company, Biopptic, is using AI to develop medicines, already patenting molecules for cancer and Parkinson's. How did he manage to reinvent himself, and why is biotech the next trillion-dollar industry? Let's dive in!

Introduction: From Windows to YouTube and Beyond

Andrey sets the stage right away, joking that they've "cured billions of people" – a nod to a future where AI in biotech becomes a reality. He walks us through his journey: first the internet, then mobile phones as a new platform, followed by VR/AR, and now, transforming the pharmaceutical industry with AI. His consistent strategy? Betting on cutting-edge technologies where paradigm shifts are possible. It's risky, sure, but that's exactly how massive companies are born.

A Breakthrough in Treating Rare Diseases

The conversation kicks off with Biopptic's recent win: Stanford scientists, using their AI models, identified the most potent candidates for treating Friedreich's Ataxia. This was a massive validation because Stanford has unique facilities for testing molecules on real patient cells. While Biopptic can generate molecules for any target, the actual testing on living cells is the phase that traditionally takes years and costs millions. The ability to speed this up is a game-changer.

How AI Generates Molecules

Andrey explains their AI, Biopptic, takes the amino acid sequence of a target protein and a vast space of theoretically synthesizable molecules (billions of compounds) as input. A neural network, trained on scientific data, predicts which of these molecules will be active against the target. Unlike highly specialized models, their system is 'target agnostic' – it works with any target, even if it hasn't seen it before. This approach led to finding 53 active molecules, two of which proved better than existing ones.

Biotech: More Than Trillions, It's About Complexity

Despite biotech's massive potential, Andrey stresses its incredible complexity. The drug development process is a long, arduous journey through many stages: molecule discovery, optimization, preclinical, and clinical trials. Each step demands huge investments and time. AI can dramatically accelerate only the earliest stages. Simulating even a single human cell is an unimaginably complex task, akin to modeling an entire city. So, talking about digital patient twins is still a long way off.

Biopptic's Exit Strategy and Business Model

Biopptic's core business strategy revolves around service revenue. Their AI agent, the 'Biopptic Agent,' acts as a 'cursor' for the pharmaceutical industry, helping automate complex processes like competitive intelligence or new drug discovery. Pharma companies pay for using this agent, providing a steady income stream. Additionally, there's a 'scientific bets' division, where they take a share of intellectual property (IP) in deals with partners, like the one with Stanford. It's a long game, but potentially highly profitable, with possibilities for licensing developments to big pharma or selling royalties to aggregators.