Xiao Xue

An artist paints a portrait stroke by stroke, and then someone offers them a flash camera — one click, and it's done. That's roughly the shift from multi-step models to one-step models in generating images of galaxies. Typical neural networks for image synthesis work like a watercolorist: they apply a layer, blur it, add details.

A lighthouse spins in dense fog. Its beam cuts through the gloom, but the fog itself remains invisible. That's roughly how astronomers search for dark matter: it doesn't emit light, yet it can affect the signals that reach us.

A low rumble coming from underground could be the noise of a distant construction site, or it could be the echo of an ancient earthquake. Astronomers faced a similar puzzle in 2023 when they detected ultra-low-frequency gravitational waves — the trembling of spacetime itself, rippling across the universe.

In 2034, a rocket leaves a spaceport with an unusual payload: two spheres the size of a large globe, each weighing nearly a ton. No engines, no electronics. Just a mirrored surface of hundreds of corner-cube reflectors — tiny prisms that return light exactly to where it came from. This is the future space tuning fork.