YouSaid · the spoken record

Karina Nguyen

lines on the record
72
first
2025-02-09
most recent
2025-02-09
sittings or episodes
1
sources
podcast

Every line below is reproduced as it was said and linked to the record it came from. Nothing here is summarised or generated. Directory · Search · Corrections

  1. You can find me, I'm on Twitter, it's premium. You can also shoot me an email on my website. And my team is hiring. And so I'm looking for research engineers, research scientists, as well as machine learning engineers, like people who come from product engineers who want to learn model training. It's really hiring for my team. My team is called Frontier Product Research. And the training models, we develop new methods, but for product oriented outcomes.

    2025-02-09 · Lenny's Podcast · OpenAI researcher on why soft skills are the future of work | Karina Nguyen (Research at OpenAI, ex-Anthropic) · IDENTIFIED FROM THE TRANSCRIPT · source

  2. I've been thinking about this. Of times, I feel like I have a lot of jobs options. I would love to be a writer, I think. I think that would be super cool. You should write short stories, like sci-fi stories, novels. I really like art history. So, you know, it's like... Conservation as to like in the museums, which is like try to preserve art paintings, but just like painting through a lot of things. I think that would be really cool. To do.

    2025-02-09 · Lenny's Podcast · OpenAI researcher on why soft skills are the future of work | Karina Nguyen (Research at OpenAI, ex-Anthropic) · IDENTIFIED FROM THE TRANSCRIPT · source

  3. What do people like? Like, kind of like creating the mental model of the user and care about the user in order to ask certain questions, like actually that part is hard for the models.

    2025-02-09 · Lenny's Podcast · OpenAI researcher on why soft skills are the future of work | Karina Nguyen (Research at OpenAI, ex-Anthropic) · IDENTIFIED FROM THE TRANSCRIPT · source

  4. Much of it is like because right now the models Operating on pixels instead of language or whatnot, like pixels is actually really, really hard for the models because perceptual visual perception. I think there's still a lot of multimodal research that's going on. But I think language scaled so much easier compared to multimodals because of that. Another thing that I guess my team is working on is like how do you derive human intent very correctly. Sometimes does the model know enough information to ask a follow-up question or to complete the task? You kind of don't want an agent to go off for like 10 minutes and then come up with an answer that you didn't even want that actually creates like much more worse user experience. And this comes with like teaching the model like people skills. Like, you know,

    2025-02-09 · Lenny's Podcast · OpenAI researcher on why soft skills are the future of work | Karina Nguyen (Research at OpenAI, ex-Anthropic) · IDENTIFIED FROM THE TRANSCRIPT · source

  5. Oh, nice. Yeah, so it's a very cool product where you can just call anyone at any time and then share screen and the other person can have access to the screen and start literally operating your computer. And it's very real time. The allegiance is very Very high quality And it's just like, I kind of want the same. It's like, I want to pair a program with my model. And the model should even talk to me, like, draw very specific section in my code in US code and tell me, I will teach me and you can have different modes. It's like right here. This is like a product right here for you. I don't know. Some people should build up.

    2025-02-09 · Lenny's Podcast · OpenAI researcher on why soft skills are the future of work | Karina Nguyen (Research at OpenAI, ex-Anthropic) · IDENTIFIED FROM THE TRANSCRIPT · source

  6. It's actually really hard. And I think. They're still cracking this way. I feel like, I don't know if you was like Tuple, it's like a pair programming product. I don't know if you love pair programming. So if

    2025-02-09 · Lenny's Podcast · OpenAI researcher on why soft skills are the future of work | Karina Nguyen (Research at OpenAI, ex-Anthropic) · IDENTIFIED FROM THE TRANSCRIPT · source

  7. Yeah, so I unfortunately did not work enough, but I'm really, really excited about this launch. It's basically a magn that can complete the task in its own virtual computer, like in its own virtual environment. You can do any literally task, like ordinary a book on Amazon, and then ideally the model will either like fall off. Hit like, yeah, help me by, and then the model goes off into its own virtual little browser and complete the task and buy the book on the Amazon. And then if you give the model credentials, credit cards, obviously it comes with a lot of trust and safety, then it will just complete the thing for youth. It's a virtual assistant.

    2025-02-09 · Lenny's Podcast · OpenAI researcher on why soft skills are the future of work | Karina Nguyen (Research at OpenAI, ex-Anthropic) · IDENTIFIED FROM THE TRANSCRIPT · source

  8. Yeah, that's really cool. I feel like Chad also has a social element, which is very humane. It's like, you know, you sometimes want to get into a group chat and have a conversation with AI is kind of like a group chat in itself. It's like messaging. I saw this idea of how do you build features like this? I see tasks as this general kind of feature that will scale very nicely as the models would develop new capabilities or so it's like the models will be able to like do better like searches and like yeah create new like come up with like more creative like writing or like render you know react apps and like HTML like apps and like you can have like everyday new puzzle for you like everyday like continue the story from the previous day it's like it it scales very nicely

    2025-02-09 · Lenny's Podcast · OpenAI researcher on why soft skills are the future of work | Karina Nguyen (Research at OpenAI, ex-Anthropic) · IDENTIFIED FROM THE TRANSCRIPT · source

  9. Yeah, I think content transformation is like, you know, like, I would imagine sometime when you generate the sci-fi story in Canvas, like you can transform this into audiobook or like where you have very natural content transformation from one media to another media. I think one of my earliest inspiration is one of the last episodes of West World where I want to Lores comes to her work at the time and she comes to like this like new workspace and she starts like writing a story and then as she writes a story like a 3D like virtual reality starts like creating on the fly so I kind of want to create that

    2025-02-09 · Lenny's Podcast · OpenAI researcher on why soft skills are the future of work | Karina Nguyen (Research at OpenAI, ex-Anthropic) · IDENTIFIED FROM THE TRANSCRIPT · source

  10. Yeah, I think the computer use agents like the model operating, your desktop. And you can essentially think of new kind of like Experience where the model can learn the way you browse and from that preference it can just like browse as just like you. It's kind of like simulated persona. And it's actually very similar to the idea of like, okay, maybe some album doesn't have a lot of time. Maybe I want to talk to his simulated simulation and ask, or like for example, like, yeah, like I really appreciate some of the technical mentorship from like Jacob. But he doesn't have a lot of time. So it's like I really want to ask him these questions like how did we respond like simulated environments like this would be really cool. That's a great.

    2025-02-09 · Lenny's Podcast · OpenAI researcher on why soft skills are the future of work | Karina Nguyen (Research at OpenAI, ex-Anthropic) · IDENTIFIED FROM THE TRANSCRIPT · source

  11. Thinking about there was like one part of clog workspaces and it's like kind of the same like idea of like Claude and I would have this shared workspace and that shared workspace is like a document and we can like tweet in the document. I feel like sometimes the ideas private ideas lag and they lag for like two years just like in this case

    2025-02-09 · Lenny's Podcast · OpenAI researcher on why soft skills are the future of work | Karina Nguyen (Research at OpenAI, ex-Anthropic) · IDENTIFIED FROM THE TRANSCRIPT · source

  12. The very first launch when we felt like When clicked from you, the game was like 100k context launch is like when the models could input the entire book and give you a whole summary of the book or something or the entire financial or have multi-files, financial reports, and then give you an answer to the question, to very specific question. I think there was something in there that kind of like, oh my God, this is like a really cool new capability, not like model capability, but more like the capabilities that came from the product form factor itself rather than the model capability as much. Think like other prototypes that we were.

    2025-02-09 · Lenny's Podcast · OpenAI researcher on why soft skills are the future of work | Karina Nguyen (Research at OpenAI, ex-Anthropic) · IDENTIFIED FROM THE TRANSCRIPT · source

  13. I know that Claude and Slug were sunsetted in like 2023 or something, but that's because I think it was after ChatGPT it was mostly like the focus on consumer use cases or like enterprise use cases. I think the form factor of Claude and Slack is kind of constrained a little bit when you wanted to develop new features.

    2025-02-09 · Lenny's Podcast · OpenAI researcher on why soft skills are the future of work | Karina Nguyen (Research at OpenAI, ex-Anthropic) · IDENTIFIED FROM THE TRANSCRIPT · source

  14. Is like so important because you both trust and the model learns from your preferences so that it can become more personalized. And it will start predicting the next action that you want to take on the computer or something. And it's kind of more predictive, much more, we went from personal computers to like personal model basically here

    2025-02-09 · Lenny's Podcast · OpenAI researcher on why soft skills are the future of work | Karina Nguyen (Research at OpenAI, ex-Anthropic) · IDENTIFIED FROM THE TRANSCRIPT · source

  15. Or, like every Friday, we just summarize a bunch of channels and give the news about the organization or something. It's kind of like really cool factor. I think I'm thinking about form factors a really important question in AI, especially we haven't even figure out how do we create an awesome product experience with like a serious models. It's like the paradigm between synchronous, real-time, give an answer, paradigm into more asynchronous paradigm of agents working on the background, but then now the question is like the agents should both trust with you, right? And trust both over time, which is like with humans. And, you know, you saw this collaboration, which is why this collaboration model was like you and a model.

    2025-02-09 · Lenny's Podcast · OpenAI researcher on why soft skills are the future of work | Karina Nguyen (Research at OpenAI, ex-Anthropic) · IDENTIFIED FROM THE TRANSCRIPT · source

  16. And I think we did a lot of cool ideas like Claude and Slack was actually one of the first tool usy products. Cloud could operate in your workplace now. It's kind of cool. But then you add Clot to summarize the thread. So maybe you have an entire conversation with someone and then you want to like a summary of what happened. You can say at Cloud summarizes also it was really fun to even iterate on the model itself is like when you just like talk to the model in like Slack forever. It creates some social element. It's kind of cool. It's kind of like me journey and like this Discord like people learned so much about prompting and like how to work with like Claude. Actually one of the features that was like early tasks prototype was like you know every Monday clog would just like summarize the entire channel

    2025-02-09 · Lenny's Podcast · OpenAI researcher on why soft skills are the future of work | Karina Nguyen (Research at OpenAI, ex-Anthropic) · IDENTIFIED FROM THE TRANSCRIPT · source

  17. One of my regrets, I guess, when I was early days at Antopric was that like, I think there was like some luxury of the time pre-Chat GPT to actually come in with a bunch of ideas on prototype almost every day.

    2025-02-09 · Lenny's Podcast · OpenAI researcher on why soft skills are the future of work | Karina Nguyen (Research at OpenAI, ex-Anthropic) · IDENTIFIED FROM THE TRANSCRIPT · source

  18. Time job can be just like teaching them how to be like creator writers. And it's like there's some luxury in this research freedom that comes to scale maybe. I don't know, but it gives you, I feel like I have much more creative product freedom to do almost anything, I guess, within OpenAI, like Evlos, ChatGPT into the vision. It's more like probably bottoms up, I guess.

    2025-02-09 · Lenny's Podcast · OpenAI researcher on why soft skills are the future of work | Karina Nguyen (Research at OpenAI, ex-Anthropic) · IDENTIFIED FROM THE TRANSCRIPT · source

  19. A lot of like craft and like to read it like datasets and this is where I learned that part of like art I guess I don't argue Azila Antarctic is like much smaller like when I joined it was like what like 70 people when I left it was 70 people and like obviously the culture changed so much I really enjoyed the early days startup like vibes and like people knew each other as a family but like the culture shifted I would say like I learned from Antarctic that like they're much better at like focusing and like prioritization and like very very hard like very hardcore prioritization I guess and they need to do it like but I think like open AI is like much more innovative and much more like risk takers in terms of like product or like research actually you know like at a can like your full

    2025-02-09 · Lenny's Podcast · OpenAI researcher on why soft skills are the future of work | Karina Nguyen (Research at OpenAI, ex-Anthropic) · IDENTIFIED FROM THE TRANSCRIPT · source

  20. Behavior, model cost, model tainting. And I've been thinking a lot about like, okay, like what makes claw cloth and what makes ChatGPTPTs like actually comes down to like operational processes that kind of lead to the outputs to the model is the outputed model. And it's like the reason why Collade has so much more personality and is more like a librarian. I don't know. I don't know. I'm like visualizing a cloud being like a librarian like a very like nerdy or something. It's because I feel like it's a reflection of the creators who are like making this model and like a lot of details around like the character and the personality and like whether the model should follow up on this question or like not like was the correct like ethical behavior for the model to like in this scenario is like

    2025-02-09 · Lenny's Podcast · OpenAI researcher on why soft skills are the future of work | Karina Nguyen (Research at OpenAI, ex-Anthropic) · IDENTIFIED FROM THE TRANSCRIPT · source

  21. It's more similar or than different. Obviously, there is a lot of differences also when it comes to nuances culture. I really love Anthropic and I have a lot of friends there and I also love OpenAI and I still have a lot of friends though. So it's like it's not about enemy. I feel like there's like in the eyes all like the competitors and there's like enemies. This is actually like one big community and like of people doing the same thing I would say would have learned from Antarctic is this real care and craft towards like moral

    2025-02-09 · Lenny's Podcast · OpenAI researcher on why soft skills are the future of work | Karina Nguyen (Research at OpenAI, ex-Anthropic) · IDENTIFIED FROM THE TRANSCRIPT · source

  22. Yeah, I'm thinking a lot about how do we make organizations more effectively? And I think this is most of like management, I guess. It's like how do you organize research teams or generally teams combine, compose teams such that they will be at the maximally succeed or at the maximum performance of Organizations like scaling product research, I guess.

    2025-02-09 · Lenny's Podcast · OpenAI researcher on why soft skills are the future of work | Karina Nguyen (Research at OpenAI, ex-Anthropic) · IDENTIFIED FROM THE TRANSCRIPT · source

  23. Yeah, same as like scientific research. It's because ideally the model will be able to suggest ideas, like new ideas or iterate on the experiment or given the empirical results of the previous experiments, like how do you come up with new ideas or methods

    2025-02-09 · Lenny's Podcast · OpenAI researcher on why soft skills are the future of work | Karina Nguyen (Research at OpenAI, ex-Anthropic) · IDENTIFIED FROM THE TRANSCRIPT · source

  24. Connecting the dots, I think it's like okay if you have users feedback from this source but you also have internal like dashboard with metrics and then you have you know like other kind of like feedback or like input and then like it can co-create like a plan for you like recommendations even I think this is like one of the most common use cases for chat2 is like coming up with like these sorts of things

    2025-02-09 · Lenny's Podcast · OpenAI researcher on why soft skills are the future of work | Karina Nguyen (Research at OpenAI, ex-Anthropic) · IDENTIFIED FROM THE TRANSCRIPT · source

  25. I think so too. I think, again, you teach the model all sorts of tools and capabilities and reasoning, right? And when it comes down to for Canvas right now, it would be very cool to for the models to just aggregate all the feedback from users, like summarize me the top five most painful flows of user experiences, and then the model itself is very capable of thinking of like knowing how it's being made, figure out how to create the data sets for itself to train on it. And I don't think we are far away from that kind of self-improvement models becoming self-improved via the product development is basically kind of like self-improving it's kind of like its own like organism or something. Yeah, like a strategy. Is like it's more like data analysis and coming up with like I think what models are really good at is like

    2025-02-09 · Lenny's Podcast · OpenAI researcher on why soft skills are the future of work | Karina Nguyen (Research at OpenAI, ex-Anthropic) · IDENTIFIED FROM THE TRANSCRIPT · source

  26. I think currently it's difficult For many reasons, I think it's still like an active research area, young is something that I think Mike King is working on is like, okay, how do we teach the model to be more creative in the writing? And actually, this new paradigm of the models think more should actually lead to better writing in itself. But when it comes down to idea generation or discriminating of like what is a good like visual design or not, I feel like it hasn't had learned examples from people to discriminate it very well. I do think it's because You know, there are another many people who are like actually really it's not like accessible to like models to learn from these people I guess so I guess like that's why it sucks yeah that makes sense

    2025-02-09 · Lenny's Podcast · OpenAI researcher on why soft skills are the future of work | Karina Nguyen (Research at OpenAI, ex-Anthropic) · IDENTIFIED FROM THE TRANSCRIPT · source

  27. Apple designers, and it's like so cool to see and like how do you create this like collaboration between people. It's just like something that's still humane, I think

    2025-02-09 · Lenny's Podcast · OpenAI researcher on why soft skills are the future of work | Karina Nguyen (Research at OpenAI, ex-Anthropic) · IDENTIFIED FROM THE TRANSCRIPT · source

  28. About was like, you need to have a really high conviction in the research bats to put the compute. And like, it's more like return on investment kind of situation. It's like, okay, yeah, like I'm thinking a lot about, okay, like how do I across all my projects which projects a higher priority is like prioritization and also like on the lower level, like which experiments are really important to run right now and which are not and like cut through the line. So I think like prioritization and communication management people's skills like empathy like understanding people like kind of like collaboration like I think like Canvas wouldn't be like an amazing launch if it wasn't like about like people and I think it's a wonderful group of people and like I get a chance to like work with like people like Lee Byron who's like a co-creator like Raptio and like some of the best like

    2025-02-09 · Lenny's Podcast · OpenAI researcher on why soft skills are the future of work | Karina Nguyen (Research at OpenAI, ex-Anthropic) · IDENTIFIED FROM THE TRANSCRIPT · source

  29. I think it's actually really, really hard to teach the model how to be aesthetic or do really good visual design or how to be extremely creative in the way they write. I think I still think ChatGP kind of sucks at writing. And that's because it's bottlenecked by this creative reasoning. I think prioritization is one of the most important, like I think for a manager, I feel like actually like AI research progress is all knocked by like management like research management is because you have like constrained set of compute and you need to like allocate the compute to the research path that you feel the most

    2025-02-09 · Lenny's Podcast · OpenAI researcher on why soft skills are the future of work | Karina Nguyen (Research at OpenAI, ex-Anthropic) · IDENTIFIED FROM THE TRANSCRIPT · source

  30. But I feel like Claude 1.3 model itself was like not there to make really extreme good high quality edits, for example, like coding. And I feel like I see startups like cursor and is like doing super well. And that's because he like it rates so fast. They like invents new ways of training models. They move really fast. They listen to users, massive distributions.

    2025-02-09 · Lenny's Podcast · OpenAI researcher on why soft skills are the future of work | Karina Nguyen (Research at OpenAI, ex-Anthropic) · IDENTIFIED FROM THE TRANSCRIPT · source

  31. NAI fields are like, I wish they were a little more creative and connecting the dots across linked fields or something like that to develop really cool new generation and new paradigms of interactions with this AI. I don't think we've practiced problem at all. A couple of years ago, I was like telling some people I was like, you know, you kind of want to build for the future so it's like It doesn't necessarily matter whether the model is good or not good right now, but you can build product ideas such that by the time the models will be really good it will work really well and I think it just happened naturally like for example like anthropic like right like the clot artifacts and I feel like early ideas of canvas was like back in like 2022 like before

    2025-02-09 · Lenny's Podcast · OpenAI researcher on why soft skills are the future of work | Karina Nguyen (Research at OpenAI, ex-Anthropic) · IDENTIFIED FROM THE TRANSCRIPT · source

  32. Yeah, I think creative thinking, like you kind of want to come up, like generate a bunch of ideas and like filter through them and not just build the best product experience. Listening, you know, you want to build something that the most general model will not replace you. And oftentimes you build something and you make it really, really good for like specific set of users and actually the mode is now in your user feedback. The mode is more in like whether you listen to them, like whether you can rapidly iterate. The mode is in here. I don't think we are yet to like there are so many ideas. I think there's an abundance of ideas that you can look at. It's like I wouldn't be worried. I feel like in fact I do think like people

    2025-02-09 · Lenny's Podcast · OpenAI researcher on why soft skills are the future of work | Karina Nguyen (Research at OpenAI, ex-Anthropic) · IDENTIFIED FROM THE TRANSCRIPT · source

  33. Would say we still have like, I think Canvas team has still have really cool front engineers that really like, you know, people who really care about interaction, design, like interacting spirits, like I don't think models are there yet, but I think, but we can get the models to this top 1% of front end or something.

    2025-02-09 · Lenny's Podcast · OpenAI researcher on why soft skills are the future of work | Karina Nguyen (Research at OpenAI, ex-Anthropic) · IDENTIFIED FROM THE TRANSCRIPT · source

  34. Like, it's humbling to have launch Canvas and bring that thing to the people, enable them to do something else that they couldn't have ever before. I think this is, there's something like magical around this experience. So education will have massive implications. I guess like scientific research, right? Like, I think it's like the dream of any AI research is augment AI research. It's kind of scary, I'd say, which makes me think that like people management will stay, you know, it's like One of the hardest things to like emotional intelligence for the models or like creativity in itself is like one of the hardest things. So writers, I don't think people should be worried as much. I think it's like, I think it alleviates a lot of redundant tasks for people.

    2025-02-09 · Lenny's Podcast · OpenAI researcher on why soft skills are the future of work | Karina Nguyen (Research at OpenAI, ex-Anthropic) · IDENTIFIED FROM THE TRANSCRIPT · source

  35. Made it worse, yeah. Yeah, that's crazy, like, right? Like education, I think I would have dreamt if I had the tool like ChatGPT when I was young and would learn so much. But people can now learn almost anything from these models so they can learn new language, they can learn how to build new look ups like, all right, anything that you want in like I'm so

    2025-02-09 · Lenny's Podcast · OpenAI researcher on why soft skills are the future of work | Karina Nguyen (Research at OpenAI, ex-Anthropic) · IDENTIFIED FROM THE TRANSCRIPT · source

  36. Bigger lesson like odds, but like the power of small models become very intelligent and fast and cheap, we are moving towards the world that has multiple impatience, but that means that like people will have more access to AI and that's really good. Builders and developers will have much better access to AI, but also it means like all the work that has been bottlenecked by intelligence will be kind of unblocked. So anyone, I'm thinking about healthcare, right? Like if I have, instead of going to the doctor, I can ask ChatGPT, like give chat GPT a list of symptoms and ask me which, like, would I have like a cold flu, like something else? Like I can literally get the access to. Doctor almost, and there's like been some research studies around that.

    2025-02-09 · Lenny's Podcast · OpenAI researcher on why soft skills are the future of work | Karina Nguyen (Research at OpenAI, ex-Anthropic) · IDENTIFIED FROM THE TRANSCRIPT · source

  37. Yes, so basically there are trends that are going on. The first trend is the cost of reasoning and intelligence is drastically going down. I had a blog post about this. Maybe I should update them latest benchmarks because at that time everybody was doing one benchmark and they'd be like quickly saturated the benchmarks. I'm like now you need to like do the same plot but with another like frontier eval. But the cost of intelligence is like going down because it becomes like much cheaper smart small models are becoming even smarter than like large models and that's because up like the distillation research. This happened with a lot of high school. I was working on like post-chaining like clattery high school and I realized it was much smarter than like claw which was like way, you know.

    2025-02-09 · Lenny's Podcast · OpenAI researcher on why soft skills are the future of work | Karina Nguyen (Research at OpenAI, ex-Anthropic) · IDENTIFIED FROM THE TRANSCRIPT · source

  38. Hanukkah context, but because like file uploads is like form follows function. It's like The front five of the phyloplow is kind of enable people to just like literally upload anything, the books, like any reports financial and ask any task to the model. And then I remember it was like, you know, enterprise customers, like financial customers are really interested in that. It's like, oh, wow, like actually it's actually one of the very common tasks that people do in that setting is like kind of crazy to see how some of the redundant tasks are getting automated basically by this like smart models. And we're entering the era where I actually don't know for example sometimes like if L1 gives me the correct answer or not because I'm not an expert in that field.

    2025-02-09 · Lenny's Podcast · OpenAI researcher on why soft skills are the future of work | Karina Nguyen (Research at OpenAI, ex-Anthropic) · IDENTIFIED FROM THE TRANSCRIPT · source

  39. Very humbling experience to be in both labs, I guess. To me, when I first came to Andarbane, I was like, oh no, I really love format engineering. And then the reason why I switched to research is because I realized at that time it's like, oh my God, like cloud is getting better. Cloud is getting better at coding. I think cloud can develop new apps or something. And so it can develop new features for the thing that I'm working. So it's like It was kind of like this meta realization where it's like, oh my God, like the world is actually changing and they like when we first launched 100k context at that time obviously you know I'm thinking about like form factors that's like yeah like file uploads were like very natural very familiar to people but you can imagine we could just like make like infinite chats in the cloud.ai app right like as if like it's like

    2025-02-09 · Lenny's Podcast · OpenAI researcher on why soft skills are the future of work | Karina Nguyen (Research at OpenAI, ex-Anthropic) · IDENTIFIED FROM THE TRANSCRIPT · source

  40. Ways doing chuck, like diversity as a general method that will work for all is like one of the research explorations. Other ones is more like developing new capabilities. I feel like it's all always about like, you know, like you work on this new method and you have signs of life that it's working. Either you think of like how do you make it more general or you think of like how do you make it very useful or like and this is how like longer term projects become more like medium like short-term projects?

    2025-02-09 · Lenny's Podcast · OpenAI researcher on why soft skills are the future of work | Karina Nguyen (Research at OpenAI, ex-Anthropic) · IDENTIFIED FROM THE TRANSCRIPT · source

  41. So, the project that I described are mostly like product oriented research, it's mostly product research. Another part component of my team is actually more longer-term exploratory projects. And it's more about developing new methods, understanding those methods and a variety of circumstances. So basically develop your methods, you kind of need to follow very similar kind of recipe of building evals, but it's much more sophisticated evals. Like you kind of want to have other distribution of like if you want to measure generalization, if I don't need to capture that, but it's basically more science-y in a way where, you know, if we talk about synthetic data, like one of the hardest things about synthetic data is like, how do you make it more diverse in synthetic data is like one of the most important questions right now? And it's like exploring like.

    2025-02-09 · Lenny's Podcast · OpenAI researcher on why soft skills are the future of work | Karina Nguyen (Research at OpenAI, ex-Anthropic) · IDENTIFIED FROM THE TRANSCRIPT · source

  42. Use. And the reason why I really love synthetic, like relying purely synthetic data instead of collecting data from humans is because it's much more scalable. It's cheap, doesn't have like, you literally sample from the model and you teach the core behaviors of the models and that will generalize to all sorts of diverse coverage. And when you launch the beta feature, you learn so much from the users that you can all use synthetic sets can be shifted in the distribution of how the users behave in the product behavior and this is how you improve. And this is what happens with Canvas too when you learn from beta to GA.

    2025-02-09 · Lenny's Podcast · OpenAI researcher on why soft skills are the future of work | Karina Nguyen (Research at OpenAI, ex-Anthropic) · IDENTIFIED FROM THE TRANSCRIPT · source

  43. Canvases was like four five months, I guess, to go from a zero to one. But yeah, and I'm like, you know, you teach private managers how to build evals and like maybe, you know, how do we not only ship the better feature, but how do we think longer term? What kind of cool features did you want tasks to have? I think it would be nice for tasks to be extreme, a little bit more personalized. It would be nice to have to create tasks via voice on a mobile, right? So you kind of need to, this is how you get research roadmap right here. It's like thinking how the feature will be developed in the future. And then from there, it's like you start creating data sets with evas, you want to make sure that goes well. And then you need to have a trade-off between what methods you want.

    2025-02-09 · Lenny's Podcast · OpenAI researcher on why soft skills are the future of work | Karina Nguyen (Research at OpenAI, ex-Anthropic) · IDENTIFIED FROM THE TRANSCRIPT · source

  44. It's like through conversations, I don't like people ask me to join the Ibn team and they're like, oh my God, we need to research artists, like we need to train the models. Or sometimes, like, it was Canvas is mostly like, I just pitched the idea, like it got stuffed quite immediately in the Europe. Actual product designer, a couple of researchers, and a bunch of applied engineers depends on the complexity of the project. And then for tasks, it took like two months or so to go from zero to one basically.

    2025-02-09 · Lenny's Podcast · OpenAI researcher on why soft skills are the future of work | Karina Nguyen (Research at OpenAI, ex-Anthropic) · IDENTIFIED FROM THE TRANSCRIPT · source

  45. How do you want the model to notify you is basically the user should give instruction to the model. And then this instruction would fire off every day or something at that particular time. So for example, if you say every day I want to learn know about the latest AI news, the model should rewrite into like, okay, like search for the latest AI news and this will this task will get fired at that particular type that the model that the user requested. And then your design is like tool spec and then actually I don't know if I feel like sometimes like

    2025-02-09 · Lenny's Podcast · OpenAI researcher on why soft skills are the future of work | Karina Nguyen (Research at OpenAI, ex-Anthropic) · IDENTIFIED FROM THE TRANSCRIPT · source

  46. Prompted prototype of how you would want the model to behave. For like tasks, for example, like you kind of need to design a little bit like design systems design thinking is like, okay, like, well, if the user says, remind me to go to lunch at 8 a.m. tomorrow. Okay, what kind of information does a model need to extract from that prompt in order to create a reminder? And so this is how you design a spec for a new feature, like a tool, canvas and tasks, all tools. So it's like how do you create the tool stack? And then it's like basically mostly like developing JSON schema was like, okay, like from this problem, maybe the model should extract the time that the user requested. And then you think about which format you want the time to be. And then like.

    2025-02-09 · Lenny's Podcast · OpenAI researcher on why soft skills are the future of work | Karina Nguyen (Research at OpenAI, ex-Anthropic) · IDENTIFIED FROM THE TRANSCRIPT · source

  47. It's kind of like PRD. It's like creating a spec of the behavior of the model. I don't think tasks is extremely groundbreaking feature necessarily. What makes it really cool is because the models are so general, model can now search. They can write sci-fi stories. They can search for stocks. They can summarize your news every day because the models are so general, like giving something familiar to people that notifications is very familiar. Having reminders is very familiar. So creating a form factor for the people who are very familiar symbols like Canvas is like Google Docs is very familiar. But then you add magical AI moment and it becomes very powerful. But the way it comes usually operationally, like, yeah, it starts as like a prototype, like literally.

    2025-02-09 · Lenny's Podcast · OpenAI researcher on why soft skills are the future of work | Karina Nguyen (Research at OpenAI, ex-Anthropic) · IDENTIFIED FROM THE TRANSCRIPT · source

  48. Having canvas and tasks are going into the bucket of projects where it's like more short or medium terms. And actually the way canvas and tasks came about to be was like it started with one person prototyping and creating like

    2025-02-09 · Lenny's Podcast · OpenAI researcher on why soft skills are the future of work | Karina Nguyen (Research at OpenAI, ex-Anthropic) · IDENTIFIED FROM THE TRANSCRIPT · source

  49. Experiment with that. Another feature was like generating titles for the conversations. It's a very small microwave spirit, but I'm really proud of the way we did that was we took five latest conversations from the user and asked the model, like, what's the style of the user? And then for the next new conversation, the generated title will be of the same style. It's just like really little micro experiences like this. That's so cool.

    2025-02-09 · Lenny's Podcast · OpenAI researcher on why soft skills are the future of work | Karina Nguyen (Research at OpenAI, ex-Anthropic) · IDENTIFIED FROM THE TRANSCRIPT · source

  50. Early days at Andarag Renay was working on file uploads feature. I remember it was just like prompting the model to just like I think we were like launching like Handi K context. I was just like prototyping this in a local person which I did the demo like people really really loved it and they just wanted like API for like file uploads or something and then that's when it clicked to me like I also like wrote a blog post a long time ago like it clicked to me like prompting is like a new way of like product development or like prototyping for designers and for like product managers for example one of the features that I wanted to do is like have like personalized recommendalized starter prompts so whenever you come to like claud like it should like recommend you like starter prompts based on what your interests are

    2025-02-09 · Lenny's Podcast · OpenAI researcher on why soft skills are the future of work | Karina Nguyen (Research at OpenAI, ex-Anthropic) · IDENTIFIED FROM THE TRANSCRIPT · source