Posted 4/9/2024, 1:01:06 PM
Startup Symbolica Raises $33M to Build Structured AI Models That Require Less Data and Compute
- Symbolica AI aims to develop more efficient "structured" AI models that require less data and compute than today's models, founded by ex-Tesla engineer George Morgan
- Current deep learning methods are not sustainable long-term due to exponentially increasing costs, according to Morgan
- Symbolica's models incorporate symbolic/mathematical abstractions to better encode knowledge and "explain" reasoning
- The startup raised a $33M funding round led by Khosla Ventures, though some experts are skeptical symbolic AI alone is the solution
- Symbolica plans to provide consulting for enterprises to build custom symbolic models, with applications like automating software development