Fast radio bursts are the time domain's sharpest mystery: millisecond flashes of coherent radio emission bright enough to be seen across billions of light-years, releasing in a thousandth of a second what the Sun radiates in days (Petroff et al. 2019; Cordes and Chatterjee 2019). The population splits stubbornly: most burst once and vanish; a minority repeat, some with periodic activity windows; and the two classes differ in spectra, durations, and polarization in ways that resist a single engine.
The standard inventory strains across the population. The 2020 detection of an FRB-like burst from a Galactic magnetar secured magnetars as one engine, but the most active repeaters require energy budgets and burst rates that push magnetar models to their limits; one-off events in passive, old-star host galaxies sit poorly with the young-magnetar channel that star-forming hosts support; compact-object merger channels predict one-offs but struggle with rates; and using FRBs as cosmological probes, the community's great hope for mapping the missing baryons, requires understanding source populations whose diversity keeps outrunning the channel inventory (Zhang 2020; Luo et al. 2024). Each new survey decade adds hosts and behaviors the tuned channels did not predict.
The standing is data-rich and theory-poor: thousands of bursts cataloged by CHIME and ASKAP, localization improving rapidly, and the engine question still open across at least two apparently distinct populations.