Nick Saraev and Jack Roberts discuss OpenAI's reported mathematical discoveries, distinguishing the promise of verifiable reasoning from the work still needed to check and reproduce claimed results. The conversation then turns to realistic AI influencers and the difficulty of deciding what to trust when synthetic people become convincing.
Nick Saraev and Jack Roberts examine model routing through Grok and SpaceX, arguing that access and distribution can matter more than using a company's own model for every task. They compare Mistral Large 4's specialist results with broader reasoning performance, then consider local AI hardware costs and the practical value of generated outputs.
Watch on YouTube
