Neurosymbolic AI outperforms chatbots and product search | E2327
Summary
This episode features Zach from Ontology, who demonstrates a neurosymbolic AI model that enables precise, real-time aesthetic search for e-commerce by learning from user mood boards without retraining. The hosts discuss how this approach offers greater trust and efficiency than traditional LLMs, which often hallucinate. The second half introduces Ashi Disanayaka of Spacium, a startup building in-orbit refueling stations. She explains their zero-boil-off technology and high-precision robotic actuators, which allow spacecraft to extend missions and carry more payload by refueling in space.
Topics discussed
Intro: Neurosymbolic models vs LLMs and sponsor ads
Demo: Ontology's visual search and real-time learning
Business model: API, revenue, and aesthetic search
Sponsor: Vanta security and compliance platform
Neurosymbolic AI: Trust, cost, and AGI potential
Company evolution: From product to platform
Sponsor: Agree contract-to-cash automation
E-commerce: Mood boards, UGC, and viral loops
SEO vs AEO, AI slop, and Twitter monetization
Intro: Spacium and the need for space refueling
Sponsor: Why Security on-demand team
Space fuels: Storables, cryogenics, and zero boil-off
Launch economics and electric propulsion benefits
Startup culture: In-house testing and rapid iteration
Robotic actuator precision and docking challenges
Next steps: Xenon transfer and industry standards
Impact on spacecraft design and SpaceX IPO effect
Bay Area ecosystem and Spacium's work culture
Listen ad-free on Castria