$DECENTRALIZEDAI - Sam advocates for open, decentralized AI platforms to drive innovation and experimentation across multiple domains, believing this approach will yield the best possible strategies in AI development.
$RLSCALING - Sam believes reinforcement learning (RL) is particularly well-suited for AI development, as it allows models to explore different strategies, verify attempts, and improve through learning.
$MODULAR - Sam is enthusiastic about highly modular sparse models in AI, which involve small, specialized models that can be trained and improved in narrow domains and then combined for broader applications.
Bearish:
$AILOCKIN - Sam expresses concern about major AI companies' strategies to create user lock-in, viewing this as a potential risk in the AI industry that could hinder innovation and user freedom.