While AI excels at screening vast compound libraries for potential drug candidates, it cannot overcome the ultimate bottleneck: the messy, complex, and poorly documented reality of human biology. The need for physical clinical trials remains the fundamental constraint on medical progress.

While AI excels at writing software—a domain with clear, verifiable outcomes (it compiles or it doesn't)—its application in fields like drug discovery is limited. Progress stalls where outcomes aren't easily and quickly verifiable, requiring complex, expensive, real-world testing.

Andrej Karpathy's 'Software 2.0' framework posits that AI automates tasks that are easily *verifiable*. This explains the 'jagged frontier' of AI progress: fields like math and code, where correctness is verifiable, advance rapidly. In contrast, creative and strategic tasks, where success is subjective and hard to verify, lag significantly behind.

AI's creative process mirrors Karl Popper's model of science. A generative model 'conjectures' plausible hypotheses (or hallucinates), and a verifier then attempts 'refutation' by testing them against hard criteria. This explains why AI currently excels in verifiable domains like code and mathematics, where correctness can be proven.

Judgment Labs CEO Alex Shan argues that AI agents will first dominate domains with easily verifiable results, like coding, where a solution's correctness can be quickly checked. Progress will be slower in non-verifiable fields like law or complex drug discovery, where feedback loops are long and ambiguous.

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.

Unlike coding, where AI models get immediate feedback on whether code runs, drug development faces immense delays. A biological hypothesis can take a decade and hundreds of millions of dollars to test in the clinic. This lack of rapid validation checkpoints is a core obstacle for AI's ability to learn and reliably improve drug target selection.

The path to AI self-improvement isn't uniform. It is happening first in software engineering and AI research because these fields have cheap, fast, and verifiable feedback (e.g., unit tests). This capability won't automatically transfer to domains like biology until similar closed-loop systems are built.

The tech industry mistakenly assumes AI's rapid success in coding will replicate across all knowledge work. Coding is an ideal use case: text-based, easily verifiable, and used by technical experts. Other fields lack this perfect setup, meaning widespread AI agent adoption will be much slower.