Reid Hoffman says Silicon Valley's got a blind spot that savvy AI investors could be cashing in on (www.businessinsider.com)

Reid Hoffman warns that Silicon Valley’s long-running “everything should be done in software” mindset has created a blind spot: the richest near-term AI breakthroughs may come from biology and other atom-focused domains where bits meet atoms. On an a16z podcast he argued that fields investors deem too slow, complex or regulated—like drug discovery, gene editing and lab automation—are fertile ground because AI doesn’t need to solve problems perfectly to be transformational. Even low-success-rate predictions (Hoffman cited models that are right ~1% of the time) can massively reduce experimental search spaces, turning a “needle in a haystack” into a tractable, high-value hit-finding problem. That view matters because momentum and capital are already shifting into medical AI: Cathie Wood and incumbents such as Microsoft and Nvidia are investing in cloud AI for hospitals, diagnostics, medical imaging, sequencing and CRISPR workflows. For the AI/ML community the implication is clear—opportunity lies in multidisciplinary systems that integrate ML models with wet labs, simulation-then-validation pipelines, and regulatory-aware product design. Startups and investors who build expertise across biology, automation, and model-driven experiment planning could capture a long runway to build iconic companies, even if development cycles and validation remain slower than pure-software ventures.