A nonprofit called Fermi Explorer Mission says it will launch a spacecraft toward Alpha Centauri by the end of 2029, following a trajectory that an AI system discovered after the team spent a year failing to find one themselves. The probe could take up to 80,000 years to cover the 4.4 light-years to our nearest star system. That’s not a typo. The mission embraces the glacial timeline as a feature, not a bug, keeping costs down to a projected $15 million.
The route came from Get Physics Done, open-source software built by Physical Superintelligence, an AI physics lab that launched today with $58 million in funding led by Bill Gates’s Breakthrough Energy. PSI cofounder Matt Pines says the system ran mostly autonomously for three days, chewing through a billion tokens, before spitting out a maneuver the Fermi team hadn’t considered. The spacecraft would swing closer to the sun than Mercury, firing its engine on each close pass where solar panels get four times the light and a burst of thrust buys more energy than it would anywhere else. Because the engine only runs near the sun, the panels stay small and the craft stays light.
The announcement lands in the shadow of Breakthrough Starshot, Yuri Milner’s 2016 plan to laser-propulse tiny probes to Alpha Centauri within 20 years. Milner pledged $100 million. Nothing launched. Fermi president Philip Johnston doesn’t hide his irritation. “We didn’t want to do another Breakthrough Starshot,” he said. “We’re dead set on something actually launching.” The probe will carry at least a kilogram of cargo, including a copy of the Golden Record—the gold-plated disc of Earth sounds NASA attached to the Voyager probes in 1977.
Pines is candid about the AI’s limits. The system lacks a human researcher’s judgment and taste, he says, often chasing dead ends or failing to recognize which problems matter. That’s the honest caveat buried in an otherwise flashy announcement: the AI didn’t design the mission, it assembled known orbital maneuvers in a combination humans hadn’t tried. Johnston, for his part, expects faster future spacecraft to beat Fermi’s probe to the finish line anyway. A launch in the year 3000 with engines just 20% quicker would arrive first. So why launch at all? Because nobody else is actually doing it.
