The public database of protein structures (PDB) is small and grows slowly. To train more powerful models, Genesis leverages physics simulations to model small molecule behavior, creating a large, high-quality synthetic dataset that isn't possible for more complex protein-protein interactions.
Similar to how an LLM uses a 'chain of thought' to reason, Genesis's model 'thinks' by iteratively refining an in-memory representation of a crystal structure. This process is guided by physics-based principles, significantly improving the final prediction's accuracy.
Genesis's focus on sub-one-angstrom accuracy came from direct experience. When applying models to active drug discovery programs, it became 'pretty obvious' that the standard two-angstrom benchmark was inadequate. This highlights the gap between academic benchmarks and real-world utility.
Moving beyond simulation, Genesis uses a cycle where their AI proposes molecules, a pharma partner synthesizes and tests them in a wet lab, and the experimental outcomes are used as feedback to retrain the generative model. This is akin to RLHF but with physical experiments.
The community standard of two-angstrom accuracy for protein-ligand predictions is insufficient. At that resolution, critical details like an aromatic ring's orientation can be wrong, rendering the model's output misleading for drug design. Genesis argues one-angstrom accuracy is the minimum for practical utility.
In the past, AI drug discovery startups often had to build their own drug pipeline to succeed. Now, a market shift is occurring where large pharmaceutical companies are actively acquiring or licensing specialized AI models and platforms, validating the business model of being a pure AI provider to the industry.
Sergey Edunov, former Llama team lead, claims that LLM architectures have not fundamentally changed since the 2017 Transformer paper. He pivoted to drug discovery AI because the model architectures required for physical sciences are more diverse, complex, and present more interesting research challenges.
Modeling small molecules might seem easier than large proteins, but the chemical search space for drug-like small molecules is astronomical (10^60). This vastness makes finding a correct binding match computationally far more complex than for more specific protein-protein interactions.
AI agents amplify both the strengths and weaknesses of their underlying models. Before reaching a certain accuracy (e.g., sub-1.9 angstrom for molecules), agents produce 'slop' and are counterproductive. Once that threshold is crossed, their ability to automate and explore becomes transformative.
Genesis achieved its sub-one-angstrom accuracy not through one algorithmic trick, but by making it a core objective from the start. This obsessive focus on the right metric guided countless small, compounding decisions across data, infrastructure, and modeling.
Despite testing millions of compounds, high-throughput screening methods suffer from enormous false-positive rates. The actual predictive value for a re-synthesized molecule is extremely low, creating an opening for high-fidelity AI models to provide cleaner, more reliable predictions.
While GANs failed for protein systems, diffusion models became the key primitive. Now, the frontier of diffusion research is in specialized scientific areas like 3D structure prediction, surpassing the innovation seen in more mainstream AI applications like image generation.
Many diseases have well-understood genetic causes but lack effective treatments. Genesis CEO Evan Feinberg argues this makes drug discovery the most impactful area for AI, as it directly addresses the bottleneck of creating selective therapies for known targets where no medicine currently exists.