Shawn Wang: belief

Transcript context

…Yeah. The old Fairly, like, I would say, involved, like, features to ship. So the team has been working for a month, all of them. The one which I would say is the newest for OpenAI is the real time API. For a couple of reasons. I mean, one, you know, it's a new modality. Second, like, it's the first time that we have an actual, like, WebSocket based API. And so, I would say that's the one that required, like, the most work over the month. To get right from a developer perspective and to also make sure that our existing safety mitigation that worked well with like real time audio in and audio out. Yeah, what design choices or what was like the sort of design choices that you want to highlight? Like, you know, like I think for me, like, WebSockets, you just receive a bunch of events. It's two way. I obviously don't have a ton of experience. I think a lot of developers are going to have to embrace this real time programming. Like, what are you designing for, or like, what advice would you have for developers exploring this? The core design hypothesis was essentially, how do we enable, like, human level latency? We did a bunch of tests, like, on average, like, human beings, like, you know, takes, like, something like 300 milliseconds to converse with each other. And so that was the design principle, essentially. Like, working backward from that, and, you know, making the technology work. And so we evaluated a few options, and WebSockets was the one that we landed on. So that was, like, one design choice. A few other, like, big design choices that we had to make prompt caching. Prompt caching, the design, like, target was automated from the get go. Like, zero code change from the developer. That way you don't have to learn, like, what is a prompt prefix, and, you know, how long does a cache work, like, we just do it as much as we can, essentially. So that was a big design choice as well. And then finally, on distillation, like, and evaluation. The big design choice was something I learned at Skype, like in my previous job, like a philosophy around, like, a pit of success. Like, what is essentially the, the, the minimum number of steps for the majority of developers to do the right thing? Because when you do evals on fat tuning, there are many, many ways, like, to mess it up, frankly, like, you know, and have, like, a crappy model, like, evals that tell, like, a wrong story. And so our whole design was, okay, we actually care about, like, helping people who don't have, like, that much experience, like, evaluating a model, like, get, like, in a few minutes, like, to a good spot. And so how do we essentially enable that bit of success, like, in the product flow?…

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