Gurley argues initial LLM versions are structurally flawed around personal memory and RAG approaches don't solve it.
“I do think that the initial versions of these models are structurally flawed around providing personal memory and allowing someone to become dependent on one of these things.”
Gurley states LLMs are great at text but horrible at numbers, citing Microsoft CTO.
“LLMs are great at text, they're horrible at numbers. And even Kevin Scott on a podcast he did ironically with Sequoia, that's over this past week, he made that point.”