How AI plotted an interstellar journey to Alpha Centauri

The Fermi Explorer Mission, a nonprofit organization, announced plans to launch a spacecraft toward Alpha Centauri, our nearest star system, by the end of 2029. Alpha Centauri is located 4.4 light-years away, which equals about 25 trillion miles. The mission aims to send a small spacecraft carrying at least one kilogram of cargo, including artistic and scientific payloads, messages, and a copy of the Golden Record—a gold-plated disc sent with NASA’s Voyager probes in 1977 as a potential message to extraterrestrial life. The spacecraft is expected to take up to 80,000 years to reach its destination, as it will travel at a much slower speed than previous proposals like Breakthrough Starshot, which aimed for a fifth of the speed of light but never launched.

The Fermi team struggled for a year to find a feasible trajectory for their low-cost mission until they turned to Physical Superintelligence (PSI), an AI physics research lab cofounded by Alex Wissner-Gross. PSI used an AI system called Get Physics Done, which breaks complex physics problems into smaller tasks and runs simulations using models like Anthropic’s Claude or OpenAI’s GPT. After three days of autonomous research, the AI proposed a novel trajectory: the spacecraft would first slow down to swing close to the sun—closer than Mercury—where its solar panels would receive four times the light, allowing short bursts of thrust to generate more energy than elsewhere. This approach keeps the spacecraft lightweight and cost-effective, with a total budget of $15 million funded by private donors.

In 2016, billionaire Yuri Milner launched Breakthrough Starshot, a competing interstellar mission aiming to send tiny probes to Alpha Centauri at one-fifth the speed of light using powerful lasers. Milner pledged $100 million for a proof of concept, but a decade later, no launch has occurred. The Fermi Explorer Mission deliberately avoids this approach, prioritizing an actual launch over speed. Philip Johnston, the mission’s cofounder, states, 'We’re dead set on something actually launching.' The Fermi mission accepts that its spacecraft will not be the first to arrive, as future advancements in propulsion could make later missions faster. Even a 20% speed improvement a thousand years from now would still allow a later probe to reach Alpha Centauri 10,000 years sooner than Fermi’s.

The Fermi Explorer Mission is not solely about engineering ambition but also seeks to address the Fermi Paradox, a question posed by physicist Enrico Fermi in 1950. The paradox questions why, given the vast number of stars and planets in the galaxy, we have not yet detected signs of intelligent extraterrestrial life. If intelligent life is common, it should have spread across the galaxy by now. The mission’s launch would prove that humanity can and wants to reach another star, challenging the two main explanations for the paradox: either interstellar travel is too difficult, or intelligent civilizations tend to self-destruct before spreading. Philip Johnston suggests that if the latter is true, 'once you hit superintelligence, that for some reason is self-destructive,' possibly due to a 'great filter' that prevents civilizations from surviving long-term.