The Never-Sleeping Crewman: Costs, Consumption, and Nanobots (AI for Sailing Boats)
29 de marzo de 2026 · 36 min · 4K visitas
Man reads AI emergency email template on tablet in cabin
Qué cubre este vídeo
Max, aboard the Beneteau Oceanis 50 Y2K in Australia, installs and demonstrates an AI agent called Nanobot as a permanent onboard crew member, interfacing with it via Telegram under the name Enea.▶ 7:52 Nanobot runs on 1 GB of RAM, written in 4,000 lines of Python — a fraction of the 400,000-line OpenCLO framework Max first attempted and abandoned because it required at least 8 GB and took days to configure.▶ 6:28 Nanobot was operational in under an hour.▶ 6:48 Max connects it to SignalK for live wind data, a Cerbo GX for battery monitoring (SOC, voltage, current, amp-hours), and solar production figures; Enea reads 597 W / 44 A solar output and 2.2 kWh charged that morning.▶ 14:49 Critically, Enea can remotely switch the battery charger on and off through the Cerbo, demonstrated live.▶ 15:20 It also pulls weather forecasts from Open Meteo, retrieves satellite imagery of the boat's position, and stores boat documents — including the Pantenius insurance certificate — sending them as Telegram attachments on demand.▶ 17:53 Max demonstrates a simulated collision claim: Enea drafts a complete incident email template to Pantenius with emergency contact numbers drawn from the policy document.▶ 20:57 Max notes he has not yet connected Enea to his email or calendar due to security concerns.▶ 17:56 A lightweight local AI agent running on minimal hardware can monitor onboard systems, control equipment remotely, and handle boat administration through a standard messaging app.
Barcos en este vídeo
Más de Ale & Max - Y2K Sailing Adventure
De Ale & Max - Y2K Sailing Adventure — mira el barco del canal, la tripulación y cada vídeo analizado.