Anthropic's team of idealistic researchers represented a high-variance bet for investors. The same qualities that could have caused failure—a non-traditional, research-first approach—are precisely what enabled breakout innovations like Claude Code, which a conventional product team would never have conceived.

Eroom's Law (Moore's Law reversed) shows rising R&D costs without better success rates. A key culprit may be the obsession with mechanistic understanding. AI 'black box' models, which prioritize predictive results over explainability, could break this expensive bottleneck and accelerate the discovery of effective treatments.

In high-stakes fields like pharma, AI's ability to generate more ideas (e.g., drug targets) is less valuable than its ability to aid in decision-making. Physical constraints on experimentation mean you can't test everything. The real need is for tools that help humans evaluate, prioritize, and gain conviction on a few key bets.

Success in drug discovery hinges on a rare, intuitive 'nose for value.' Lengauer likens this to a star athlete who consistently makes the game-winning shot after many attempts. It's an unteachable gift for getting the big decisions right more often than others, especially in a context of repeated failure.

A significant number of Eli Lilly's compelling inventions came from unsanctioned projects. The company intentionally provides budget flexibility and avoids micromanagement at its R&D sites, allowing scientists to pursue their curiosity.

The future of AI in drug discovery is shifting from merely speeding up existing processes to inventing novel therapeutics from scratch. The paradigm will move toward AI-designed drugs validated with minimal wet lab reliance, changing the key question from "How fast can AI help?" to "What can AI create?"

The blockbuster drug bivalirudin was discovered as an unsanctioned "20% time" project at Biogen. This policy, allowing scientists to explore personal interests, demonstrates how institutionalizing freedom for undirected research can lead to major, company-defining breakthroughs that would otherwise be missed in a rigid R&D structure.

Profluent CEO Ali Madani frames the history of medicine (like penicillin) as one of random discovery—finding useful molecules in nature. His company uses AI language models to move beyond this "caveman-like" approach. By designing novel proteins from scratch, they are shifting the paradigm from finding a needle in a haystack to engineering the exact needle required.

Despite major scientific advances, the key metrics of drug R&D—a ~13-year timeline, 90-95% clinical failure rate, and billion-dollar costs—have remained unchanged for two decades. This profound lack of productivity improvement creates the urgent need for a systematic, AI-driven overhaul.