“dont go to events” - YC Alums but if i didn’t go i wouldn’t have met @patrickc, @davj, @timgl, and others who i can’t find on twitter first of all, it’s crazy that i met them at all, but they also dropped the craziest wisdom that goes against conventional YC advice

Here's how recursive language models had a real business impact for us: At Moda we process millions of LLM chats daily and cluster them into use cases to help companies understand what users are doing and where their AI is failing. Previously we tried the obvious approach: analyze each segment in each conversation, propagate back up. Turns out aggregation is lossy, roll up 10k summaries and all the sharp insights get averaged into mush. RLMs flip it. The LLM explores the full dataset programmatically and recurses into what matters. Top-down with selective depth beats bottom-up aggregation. Shoutout to @lateinteraction & @a1zhang for such a great paper

most teams building AI products are flying blind. They have millions of conversations but no idea what users are actually trying to do. analytics doesn't equal insights, people have to try to figure out what logs and clicks mean for their products. now with LLMs we're in a position where we have MILLIONS of chat logs with in-depth data of exactly what the user needs, but all the existing platforms just simply present it to you as rows and rows of logs. Moda fixes that. @qtreplayz and I are changing how teams iterate on their products. no more flying blind