However, there are doubts about the viability of symbolic AI. Critics argue that these models depend on highly structured data, making them brittle and context-dependent. Despite this, Morgan maintains that Symbolica is well-positioned for the future, as its models are relatively small and cheap to train and run. He anticipates that the company's staff will double by 2025.
Key takeaways:
- Ex-Tesla engineer George Morgan founded a startup, Symbolica AI, to develop novel AI models that can achieve greater accuracy with lower data requirements, training time, and cost.
- Symbolica AI's approach is based on structured AI models, also known as symbolic AI, which encode the underlying structure of data instead of trying to approximate insights from large data sets.
- Symbolica AI recently came out of stealth mode and secured a $33 million investment led by Khosla Ventures, with participation from Abstract Ventures, Buckley Ventures, Day One Ventures, and General Catalyst.
- Despite skepticism from some quarters about the viability of symbolic AI, Morgan believes that structured and explainable outputs with formal reasoning capabilities will be necessary to meet future demands in the AI field.