Analytics

TBD (by 3/13/2024)

• Through their exposed intuitive APIs, LLMs (e.g., ChatGPT, etc.) credibly promise customizable “AGI” functionality

  • LLMs are universal approximation function generators
  • Formal axiomatic frameworks only need predicate functions from LLMs

• (Not in scope): Reinforcement Learning (RL and “Q-Learning” strategies) of an “AGI” could effectively help seeking formal proofs