On this episode of The Blindspot, Jay dives back into the world of large language models—what they can do now, and where they’re headed. He breaks down how the chatbots you’re probably using are already smart enough to walk you through building a boat, and why the cutting-edge models are operating on a whole different level—reasoning through frontier science, math, even leukemia research. Jay wrestles with being a humanities guy in an age of machine-driven breakthroughs, then walks us through how he used OpenAI’s latest model to outline a nine-book series starring a character named John Parker Belasko. He ends with a blunt truth: today’s models are the worst they'll ever be. The future? It’s all about how many tokens it’ll cost to solve your next big problem. Enjoy