AI's initial pharma application focused on molecule design, driven by commercial incentives. New molecules are patentable and thus more easily fundable. This occurred despite incorrect target selection being the primary cause of drug development failure, representing a less immediately commercializable but more fundamental problem.

The discovery-based model of finding highly impactful single targets like HER2 or PD-1 is becoming unsustainable as the low-hanging fruit is picked. The field must shift toward an engineering-first approach, designing complex, multi-functional therapeutics to achieve specific clinical objectives, much like high-tech fields.

The paradigm for drug development is shifting from being "first" or "best" in a category to being "last in class." Using AI, the goal is to design a molecule with such intentionality and specificity that it becomes the final, definitive therapeutic for a disease, rendering subsequent improvements unnecessary.

Combining two payloads in an Antibody-Drug Conjugate (ADC) introduces a major risk: new, synergistic toxicities not seen with either agent alone. This complicates dose-finding and safety assessment, requiring developers to anticipate and monitor for entirely novel side effects.

Unlike traditional therapies, the safety of multi-specific antibodies cannot be optimized later via dose adjustments. Critical safety profiles are determined at the initial design stage, and early flaws can prevent a molecule from ever reaching therapeutically effective doses.

Increasing a biologic's binders from two or four to six or twelve is not an incremental improvement. It creates 'emergent properties of scale.' This high valency allows for sophisticated control over 3D spatial geometry at the cell surface and eliminates the design trade-offs inherent in simpler multispecific molecules.

As biologics evolve into complex multi-specific and hybrid formats, the number of design parameters (valency, linkers, geometry) becomes too vast for experimental testing. AI and computational design are becoming essential not to replace scientists, but to judiciously sample the enormous design space and guide engineering efforts.

Designing therapeutics with immense combinatorial complexity is impossible through rational design alone. The optimal approach is to first use human biological hypotheses to narrow the vast search space. Then, employ large-scale screening and data analysis to optimize within that constrained space, navigating variables too complex for human comprehension.

Instead of screening billions of nature's existing proteins (a search problem), AI-powered de novo design creates entirely new proteins for specific functions from scratch. This moves the paradigm from hoping to find a match to intentionally engineering the desired molecule.