Anton Osika: belief

Transcript context

…Okay, that sounds like product market fit to me. You said that you did a lot of work on the backend. I saw you tweet about this that you guys figured out some kind of unlock on scalability, like a new scaling law that allowed you to build something like this. What can you talk about there that on the technical element allowed you to build something new and the successful? There are many scaling laws I would say when you build AI systems and this one in particular is about when you put in more work, the product reliably gets better and better. And what you've seen generally when you have AI building something is that it can get stuck in some place. It is super good in the beginning and then it gets stuck. What we did was to painstakingly identify places where it got stuck and there is different approaches but address different ways how we do it but address the places where it gets stuck, tune the entire system quantitatively and having a very fast feedback loop to improve it in the areas where it got stuck. The most important areas, it still does get stuck sometimes, but that's the scaling law and we're still early in that scaling law, I would say. And so when you talk about things getting stuck, it's like the AI agent just saying, I don't know what to do from this point or they introduce some kind of bug. Is that an example of getting stuck?…

Stored transcript either side of the excerpt. The highlighted words are the published quote; the surrounding text is unedited source, never generated.