The pharmaceutical industry's ability to produce new medicines relies on identifying druglike molecules targeting clinically relevant pathways. Traditional lead identification involves screening compound collections for hits, but even large libraries fall short of the enormous chemical diversity space (estimated up to 10⁶⁰ molecules with ≤30 non-hydrogen atoms). Since smaller molecules have exponentially fewer possibilities, screening fragments (very small molecules) and expanding/merging/linking them is theoretically more efficient—first proposed by Jencks over 2 decades ago and experimentally supported by Nakamura and Abeles in 1985, though identifying weak-binding fragments and elaborating them remained challenging. The 1996 SAR by NMR method reignited interest in fragment-based approaches: low-molecular-weight fragments (selected for target binding/inhibition) are combined or optimized into leads meeting/exceeding HTS hit criteria. Rationale includes discrete target binding subsites, modular drug architectures, and advantages like improved affinity, selectivity, pharmaceutical properties, fewer interfering moieties, and combinatorial efficiency (sampling larger diversity space than standard screening). Despite theoretical benefits, fragment discovery and advancement/linkage are practical challenges. This Perspective reviews major techniques for identifying/validating weak-binding fragments and explores examples of fragment development into chemical lead series.