이 에피소드에 관해
PiLogic is building mathematical, expert-guided inference engines that run in milliseconds to diagnose and save satellites when standard rules fail.
💼 Are you a strategic investor exploring dual-use tech, critical infrastructure, or advanced sovereign capabilities? Let’s connect. Book a 1:1 here: 🔗 https://calendly.com/podcast-psion/30min
In this episode of Tech Command Investing, host Alain Gavin speaks with Johannes Waldstein, Co-Founder and CEO of PiLogic. While heavy investment continues to flow into generative models that require massive data centres and are prone to hallucination, PiLogic takes a contrarian approach. The company develops "exact AI" using Bayesian inference to calculate precise probabilities without relying on vast historical datasets. This enables spacecraft to identify anomalies, such as thermal runaways or faulty sensors, using a fraction of standard computing power.
We examine PiLogic's recent two-year agreement with the U.S. Air Force Research Laboratory, which tests the engine's ability to diagnose complex electrical failures before a complete system breakdown occurs. Johannes also explains the mechanics of PiLogic's "Resolve" engine. This system provides a software-based sensor upgrade capable of accurate radar tracking through the severe noise and GPS jamming observed in recent conflicts in Ukraine and the Middle East.
Finally, the conversation covers the transition to the "orbital edge". With hardware companies now launching data-centre-class GPUs into orbit, we discuss why space-based infrastructure relies heavily on highly efficient, exact machine reasoning to maintain critical operations.
📚 Books mentioned in this episode:
The Book of Why by Judea Pearl Modeling and Reasoning with Bayesian Networks by Adnan Darwiche
📊 Figures:
View the figures mentioned in this episode by checking the download link here: https://www.dropbox.com/scl/fo/cesbv7b5bd073qv9w1q02/AEqUdeJ4lIv6_3NLlTqPAmY?rlkey=9bz93njhmbblbh7e3l034apmz&st=rn8mty0e&e=1&dl=0
