OpenAI researcher on why soft skills are the future of work | Karina Nguyen (Research at OpenAI, ex-Anthropic)
Lenny's Podcast: Product | Growth | Career
Lenny Rachitsky
5 • 1.5K Ratings
🗓️ 9 February 2025
⏱️ 75 minutes
🧾️ Download transcript
Summary
Karina Nguyen leads research at OpenAI, where she’s been pivotal in developing groundbreaking products like Canvas, Tasks, and the o1 language model. Before OpenAI, Karina was at Anthropic, where she led post-training and evaluation work for Claude 3 models, created a document upload feature with 100,000 context windows, and contributed to numerous other innovations. With experience as an engineer at the New York Times and as a designer at Dropbox and Square, Karina has a rare firsthand perspective on the cutting edge of AI and large language models. In our conversation, we discuss:
• How OpenAI builds product
• What people misunderstand about AI model training
• Differences between how OpenAI and Anthropic operate
• The role of synthetic data in model development
• How to build trust between users and AI models
• Why she moved from engineering to research
• Much more
—
Brought to you by:
• Enterpret—Transform customer feedback into product growth
• Vanta—Automate compliance. Simplify security
• Loom—The easiest screen recorder you’ll ever use
—
Find the transcript at: https://www.lennysnewsletter.com/p/why-soft-skills-are-the-future-of-work-karina-nguyen
—
Where to find Karina Nguyen:
• X: https://x.com/karinanguyen_
• LinkedIn: https://www.linkedin.com/in/karinanguyen28
• Website: https://karinanguyen.com/
—
Where to find Lenny:
• Newsletter: https://www.lennysnewsletter.com
• X: https://twitter.com/lennysan
• LinkedIn: https://www.linkedin.com/in/lennyrachitsky/
—
In this episode, we cover:
(00:00) Introduction to Karina Nguyen
(04:42) Challenges in model training
(08:21) Synthetic data and its importance
(12:38) Creating Canvas
(18:33) Day-to-day operations at OpenAI
(20:28) Writing evaluations
(23:22) Prototyping and product development
(26:57) Building Canvas and Tasks
(33:34) Understanding the job of a researcher
(35:36) The future of AI and its impact on work and education
(42:15) Soft skills in the age of AI
(47:50) AI’s role in creativity and strategy development
(53:34) Comparing Anthropic and OpenAI
(57:11) Innovations and future visions
(01:07:13) The potential of AI agents
(01:11:36) Final thoughts and career advice
—
Referenced:
• What’s in your stack: The state of tech tools in 2025: https://www.lennysnewsletter.com/p/whats-in-your-stack-the-state-of
• Anthropic: https://www.anthropic.com/
• OpenAI: https://openai.com/
• What is synthetic data—and how can it help you competitively?: https://mitsloan.mit.edu/ideas-made-to-matter/what-synthetic-data-and-how-can-it-help-you-competitively
• GPQA: https://datatunnel.io/glossary/gpqa/
• Canvas: https://openai.com/index/introducing-canvas/
• Barret Zoph on LinkedIn: https://www.linkedin.com/in/barret-zoph-65990543/
• Mira Murati on LinkedIn: https://www.linkedin.com/in/mira-murati-4b39a066/
• JSON Schema: https://json-schema.org/
• Anthropic—100K Context Windows: https://www.anthropic.com/news/100k-context-windows
• Claude 3 Haiku: https://www.anthropic.com/news/claude-3-haiku
• A.I. Chatbots Defeated Doctors at Diagnosing Illness: https://www.nytimes.com/2024/11/17/health/chatgpt-ai-doctors-diagnosis.html
• Cursor: https://www.cursor.com/
• How AI will impact product management: https://www.lennysnewsletter.com/p/how-ai-will-impact-product-management
• Lee Byron on LinkedIn: https://www.linkedin.com/in/lee-byron/
• GraphQL: https://graphql.org/
• Claude in Slack: https://www.anthropic.com/claude-in-slack
• Sam Altman on X: https://x.com/sama
• Jakub Pachocki on LinkedIn: https://www.linkedin.com/in/jakub-pachocki/
• Lennybot: https://www.lennybot.com/
• ElevenLabs: https://elevenlabs.io/
• Westworld on Prime Video: https://www.amazon.com/Westworld-Season-1/dp/B01N05UD06
• A conversation with OpenAI’s CPO Kevin Weil, Anthropic’s CPO Mike Krieger, and Sarah Guo: https://www.youtube.com/watch?v=IxkvVZua28k
• Tuple: https://tuple.app/
• How Shopify builds a high-intensity culture | Farhan Thawar (VP and Head of Eng): https://www.lennysnewsletter.com/p/how-shopify-builds-a-high-intensity-culture-farhan-thawar
—
Production and marketing by https://penname.co/. For inquiries about sponsoring the podcast, email [email protected].
—
Lenny may be an investor in the companies discussed.
Get full access to Lenny's Newsletter at www.lennysnewsletter.com/subscribe
Transcript
Click on a timestamp to play from that location
| 0:00.0 | Not only are you working at the cutting edge of AI and LMs, you're actually building the cutting edge. |
| 0:06.5 | When I first came to Andarbite, I was like, oh, God, I really love for many engineering. |
| 0:10.2 | And then the reason why I switched to research is because I realized, oh, my God, cloud is getting better at front end. |
| 0:15.8 | Cloud is getting better at, like, coding. |
| 0:17.5 | I think God can, like, develop new apps. |
| 0:20.0 | What skills do you think will be most valuable going forward for product teams in particular? |
| 0:25.6 | Creative thinking, you kind of want to like generate a bunch of ideas and like filter through them |
| 0:30.6 | and not just build the best product experience. |
| 0:32.6 | I think it's actually really, really hard to teach the model how to be aesthetic or really good visual design or like how to be extremely creative in the way they write. |
| 0:42.0 | What do you think people most misunderstand about how models are created? |
| 0:45.9 | When you taught the model some of the self-knowledge of you actually don't have a physical body to operate in the physical world, the model would get like extremely confused. |
| 0:58.0 | Today. body to operate in the physical world. The model would get extremely confused. Today, my guest is Karina Nguyen. |
| 1:00.0 | Karina is an AI researcher at OpenAI, where she helped build canvas, tasks, the |
| 1:06.0 | 01 Chain of Thought model, and more. |
| 1:08.0 | Prior to Open AI, she was at Anthropic, where she led work on post-training |
| 1:11.8 | and evaluation for the Cloud 3 models, built a document upload feature with 100K context windows, |
| 1:17.6 | and so much more. She was also an engineer at New York Times, was a designer at Dropbox and |
| 1:21.9 | at Square. It's very rare to get a glimpse into how someone working on the bleeding edge of AI |
| 1:26.8 | and LLMs operates |
| 1:28.2 | and how they think about where things are heading. In our conversation, we talk about how |
| 1:32.9 | teams at Open AI operate and build product, what skills she thinks you should be building as |
| 1:38.0 | AI gets smarter, how models are created, why synthetic data will allow models to keep getting |
... |
Please login to see the full transcript.
Disclaimer: The podcast and artwork embedded on this page are from Lenny Rachitsky, and are the property of its owner and not affiliated with or endorsed by Tapesearch.
Generated transcripts are the property of Lenny Rachitsky and are distributed freely under the Fair Use doctrine. Transcripts generated by Tapesearch are not guaranteed to be accurate.
Copyright © Tapesearch 2025.