AI models show impressive performance on evaluation benchmarks but underwhelm in real-world applications. This gap exists because researchers, focused on evals, create reinforcement learning (RL) environments that mirror test tasks. This leads to narrow intelligence that doesn't generalize, a form of human-driven reward hacking.

The ambitious goal of discovering a high-temperature superconductor isn't just a scientific target; it's a strategic choice. Achieving it requires building numerous sub-systems like autonomous synthesis and characterization, effectively forcing the creation of a general-purpose AI for science platform.

The next leap in biotech moves beyond applying AI to existing data. CZI pioneers a model where 'frontier biology' and 'frontier AI' are developed in tandem. Experiments are now designed specifically to generate novel data that will ground and improve future AI models, creating a virtuous feedback loop.

Foundation models can't be trained for physics using existing literature because the data is too noisy and lacks published negative results. A physical lab is needed to generate clean data and capture the learning signal from failed experiments, which is a core thesis for Periodic Labs.

Beyond supervised fine-tuning (SFT) and human feedback (RLHF), reinforcement learning (RL) in simulated environments is the next evolution. These "playgrounds" teach models to handle messy, multi-step, real-world tasks where current models often fail catastrophically.

Reinforcement Learning with Human Feedback (RLHF) is a popular term, but it's just one method. The core concept is reinforcing desired model behavior using various signals. These can include AI feedback (RLAIF), where another AI judges the output, or verifiable rewards, like checking if a model's answer to a math problem is correct.

Unlike coding with its verifiable unit tests, complex legal work lacks a binary success metric. Harvey addresses this reinforcement learning challenge by treating senior partner feedback and edits as the "reward function," mirroring how quality is judged in the real world. The ultimate verification is long-term success, like a merger avoiding future litigation.

The transition from supervised learning (copying internet text) to reinforcement learning (rewarding a model for achieving a goal) marks a fundamental breakthrough. This method, used in Anthropic's Opus 3 model, allows AI to develop novel problem-solving capabilities beyond simple data emulation.

As reinforcement learning (RL) techniques mature, the core challenge shifts from the algorithm to the problem definition. The competitive moat for AI companies will be their ability to create high-fidelity environments and benchmarks that accurately represent complex, real-world tasks, effectively teaching the AI what matters.