Quincy Larson
👤 PersonAppearances Over Time
Podcast Appearances
Yeah, so I listen to the community and what people in the community seem to think is important. And then I also talk to employers. I talk to professors. I talk to just as many different people as I can to understand where things are going. And I spend a great deal of time immersed in listening to the ChangeLog podcast and other podcasts about math, programming, technology, computer science.
Yeah, so I listen to the community and what people in the community seem to think is important. And then I also talk to employers. I talk to professors. I talk to just as many different people as I can to understand where things are going. And I spend a great deal of time immersed in listening to the ChangeLog podcast and other podcasts about math, programming, technology, computer science.
I spend a lot of time studying history. I spend a lot of time talking to people in India, China, Brazil, countries where we're there are huge numbers of developers that dwarf even like what's in the United States in a lot of cases. And I'm, I'm trying to figure out what are they doing? What is important to them? Where, you know, you want to skate where the puck is going as Wayne Gretzky says.
I spend a lot of time studying history. I spend a lot of time talking to people in India, China, Brazil, countries where we're there are huge numbers of developers that dwarf even like what's in the United States in a lot of cases. And I'm, I'm trying to figure out what are they doing? What is important to them? Where, you know, you want to skate where the puck is going as Wayne Gretzky says.
And then like, we very much want to be out ahead of things. And that doesn't necessarily mean like, Oh, AI is a big thing. We're going to implement like, an AI chat bot that you can talk to when like, you know, I mean that, that might be useful, but it's just a wrapper around a big foundation model, right?
And then like, we very much want to be out ahead of things. And that doesn't necessarily mean like, Oh, AI is a big thing. We're going to implement like, an AI chat bot that you can talk to when like, you know, I mean that, that might be useful, but it's just a wrapper around a big foundation model, right?
Like I don't know that the world needs another wrapper around a foundation model, but what we can do is we can take AI and we can use it to speed up our localization process, right? Before we hand translated and then we use machine translation to translate the free cocaine core curriculum.
Like I don't know that the world needs another wrapper around a foundation model, but what we can do is we can take AI and we can use it to speed up our localization process, right? Before we hand translated and then we use machine translation to translate the free cocaine core curriculum.
And now what we're doing is we're feeding, you know, as much into the context window as we can into GPT four Oh, and we're saying, Hey, Translate this to the best of your ability into accessible Brazilian Portuguese, right? Sixth grade reading level. And then we're having native Portuguese speakers review and correct that.
And now what we're doing is we're feeding, you know, as much into the context window as we can into GPT four Oh, and we're saying, Hey, Translate this to the best of your ability into accessible Brazilian Portuguese, right? Sixth grade reading level. And then we're having native Portuguese speakers review and correct that.
And that is much faster to correct machine translation than it is to translate yourself.
And that is much faster to correct machine translation than it is to translate yourself.
So it dramatically speeds up the rate at which we can
So it dramatically speeds up the rate at which we can
get localization done and we're not going to be one of those pages where you just go and there's like some google you know plug-in thing that like basically auto translates everything and there's like 70 languages immediately available one click no these are all translated by native speakers there's always a native speaker in the loop because we value clear communication.
get localization done and we're not going to be one of those pages where you just go and there's like some google you know plug-in thing that like basically auto translates everything and there's like 70 languages immediately available one click no these are all translated by native speakers there's always a native speaker in the loop because we value clear communication.
We want to make sure the curriculum is clear. Uh, and there are like, I think 10, 12 novels worth of content in the core curriculum in terms of text. I mean, it's like hundreds of thousands of words that need to be translated and we want to make sure those are translated correctly. So we use, we use LLMs for that.
We want to make sure the curriculum is clear. Uh, and there are like, I think 10, 12 novels worth of content in the core curriculum in terms of text. I mean, it's like hundreds of thousands of words that need to be translated and we want to make sure those are translated correctly. So we use, we use LLMs for that.
So we're always adopting new tools, but we're not just following the leader because we don't have to follow the leader because we're a charity and we don't have investors and we don't have owners and we can just basically kind of follow the will of the community.
So we're always adopting new tools, but we're not just following the leader because we don't have to follow the leader because we're a charity and we don't have investors and we don't have owners and we can just basically kind of follow the will of the community.