LLenny's Podcast
← All episodes
Karina Nguyen (Research at OpenAI, ex-Anthropic)09 February 2025

OpenAI researcher on why soft skills are the future of work

4Frameworks
15Insights

Episode overview

OpenAI researcher Karina Nguyen explains how model training combines high-quality data, synthetic task generation, reinforcement learning, and robust evaluations. Drawing on her work on Canvas and Tasks and her earlier experience at Anthropic, she describes product development as a tight collaboration among researchers, engineers, product managers, and designers. She argues that as reasoning becomes cheaper and technical work becomes easier to automate, creativity, prioritization, communication, empathy, and people management will become increasingly valuable.

Key ideas

  • Model training is closer to an art than a settled science, with data quality and careful debugging central to shaping reliable behavior.
  • Synthetic data enables rapid, scalable iteration on product-specific behaviors, while expert human data remains important for specialized knowledge.
  • Product teams building AI features need deterministic and human evaluations to define ideal behavior, measure progress, and catch regressions.
  • Canvas and Tasks began as prototypes and behavioral specifications before multidisciplinary teams trained models and iterated from user feedback.
  • Cheaper reasoning and stronger small models will broaden access to intelligence and automate work previously constrained by scarce expertise.
  • Creative thinking, listening, prioritization, communication, empathy, collaboration, and management are likely to become more valuable as hard skills are automated.
  • Useful agents must infer human intent, ask timely follow-up questions, earn trust over time, and learn individual preferences.
  • OpenAI is characterized as more bottom-up and risk-taking, while Anthropic is described as especially focused on craft, model character, and prioritization.

Transcript available · source text is retained privately and is not published

Frameworks in this episode

Strategy3 steps

Build for the Future Model

Design product ideas for the model capability that is coming, not the one you have today

Read →
Innovation5 steps

Core-Behavior Synthetic Training

Ship an AI feature by naming its 3-4 core behaviors and teaching them with model-generated data

Read →
Productivity4 steps

Eval-First AI Product Development

Define what 'correct' looks like as evals first; the spec becomes the scorecard, not the build instructions

Read →
Innovation3 steps

Prompting-as-Prototyping

Validate a product idea by prompting a model in a local browser before building anything

Read →

People & resources mentioned

Attributed to the moment in the episode. Timestamps are approximate.

People · 5

  • Karina NguyenMentionsKarina is an AI researcher at OpenAI and previously worked at Anthropic, The New York Times, Dropbox, and Square.

    “Today, my guest is Karina Nguyen. Karina is an AI researcher at OpenAI”

    ▶ Watch at ~1:00approx.
  • Christina CacioppoMentionsI am very excited to have Christina Copo CEO and co-founder vanta joining me

    “I am very excited to have Christina Cassiopo, CEO and co-founder of Vanta, joining me”

    ▶ Watch at ~3:00approx.
  • Lenny RachitskyMentionstoday we've got another very special compilation episode something I've been pulling on more and more with the podcast and the newsletter

    “Thank you so much, Lenny, for inviting me.”

    ▶ Watch at ~4:30approx.
  • Lee ByronMentionsmentioned

    “I got a chance to like work with like people like Lee Byron”

    ▶ Watch at ~47:00approx.
  • Albert EinsteinMentionsThere's that great Einstein quote about insanity being repeating things that you're doing, expecting a different outcome.

    “You could talk to Albert Einstein, you could talk to someone not very smart”

    ▶ Watch at ~1:06:00approx.

Resources · 37

  • DropboxMentionscompany · Drew Houston and Arash Ferdowsi

    “was a designer at Dropbox and at Square.”

    ▶ Watch at ~1:00approx.
  • OpenAI o1Coinedsoftware · OpenAI

    “where she helped build Canvas, Tasks, the 01 chain of thought model, and more.”

    ▶ Watch at ~1:00approx.
  • The New York TimesMentionscompany · The New York Times Company

    “She was also an engineer at New York Times”

    ▶ Watch at ~1:00approx.
  • SquareMentionscompany · Block, Inc.

    “was a designer at Dropbox and at Square.”

    ▶ Watch at ~1:00approx.
  • CanvasCoinedsoftware · OpenAI

    “Karina is an AI researcher at OpenAI, where she helped build Canvas, Tasks”

    ▶ Watch at ~1:00approx.
  • OpenAIMentionscompany

    “Karina is an AI researcher at OpenAI, where she helped build Canvas, Tasks, the 01 chain of thought model”

    ▶ Watch at ~1:00approx.
  • AnthropicMentionscompany

    “Prior to OpenAI, she was at Anthropic, where she led work on post-training and evaluation for the Claude 3 models”

    ▶ Watch at ~1:00approx.
  • WisdomMentionssoftware · Enterpret

    “identifying drivers of churn with Interpret's AI assistant Wisdom.”

    ▶ Watch at ~2:30approx.
  • EnterpretMentionssoftware

    “Visit enterpret.com/lenny to connect with the team and to get two free months”

    ▶ Watch at ~2:30approx.
  • VantaCoinedsoftware · Christina Cacioppo

    “we started Vanta in 2018 focused on founders”

    ▶ Watch at ~3:30approx.
  • ChatGPTUsessoftware · OpenAI

    “90% of people said that they use ChatGPT regularly.”

    ▶ Watch at ~5:00approx.
  • StargateMentionsother

    “OpenAI announced Stargate which is this half trillion dollar investment in AI infrastructure.”

    ▶ Watch at ~5:00approx.
  • Claude 3Mentionssoftware · Anthropic

    “especially with like Claude 3 training”

    ▶ Watch at ~7:00approx.
  • GPQAMentionspaper

    “GPQA, which is like a Google proof question answering like PhD level.”

    ▶ Watch at ~10:30approx.
  • TasksCoinedsoftware · OpenAI

    “the way we made canvas and tasks and like new like product features”

    ▶ Watch at ~12:30approx.
  • Google SheetsUsessoftware · Google

    “sometimes I ask product managers to just like go create like a Google Sheet”

    ▶ Watch at ~22:00approx.
  • ClaudeUsessoftware · Anthropic

    “I love the file upload feature that Claude has by the way.”

    ▶ Watch at ~26:00approx.
  • LoomMentionssoftware · Loom

    “Loom lets you record your screen, your camera, and your voice to share video messages easily.”

    ▶ Watch at ~32:30approx.
  • AtlassianMentionscompany

    “Now part of Atlassian, the makers of Jira.”

    ▶ Watch at ~33:00approx.
  • JiraMentionssoftware · Atlassian

    “Now part of Atlassian, the makers of Jira.”

    ▶ Watch at ~33:00approx.
  • Claude 3 HaikuMentionssoftware · Anthropic

    “This happened with like Claude 3 Haiku.”

    ▶ Watch at ~38:30approx.
  • Claude 2Mentionssoftware · Anthropic

    “it was much smarter than like Claude 2”

    ▶ Watch at ~39:00approx.
  • Claude ArtifactsMentionssoftware · Anthropic

    “like the Claude artifacts.”

    ▶ Watch at ~44:00approx.
  • Character.AIMentionscompany

    “I feel like startups like Character is like doing super well.”

    ▶ Watch at ~44:30approx.
  • SlackUsessoftware · Slack Technologies

    “Claude and Slack was actually one of the first like tool using like products.”

    ▶ Watch at ~57:30approx.
  • MidjourneyMentionssoftware · Midjourney, Inc.

    “It's kind of like Midjourney and like this Discord.”

    ▶ Watch at ~58:00approx.
  • DiscordMentionssoftware · Discord Inc.

    “It's kind of like Midjourney and like this Discord.”

    ▶ Watch at ~58:00approx.
  • LennybotCoinedwebsite · Lenny Rachitsky

    “That's a great place to plug Lenny bot. I have one of those.”

    ▶ Watch at ~1:03:30approx.
  • ElevenLabsUsescompany · ElevenLabs

    “There's an 11 Labs voice version that's trained on my voice”

    ▶ Watch at ~1:04:00approx.
  • WestworldMentionstv show · Jonathan Nolan and Lisa Joy

    “one of the last episodes of like Westworld”

    ▶ Watch at ~1:04:30approx.
  • OperatorMentionssoftware · OpenAI

    “which will be out by the time this comes out called operator.”

    ▶ Watch at ~1:07:00approx.
  • AmazonMentionscompany · Jeff Bezos

    “order me a book on Amazon”

    ▶ Watch at ~1:07:30approx.
  • TupleMentionssoftware

    “I don't know if you heard of this like topol it's like a pair programming product.”

    ▶ Watch at ~1:08:30approx.
  • Visual Studio CodeUsessoftware · Microsoft

    “draw like very specific like section in my code and VS code”

    ▶ Watch at ~1:09:30approx.
  • Apple PodcastsRecommendssoftware · Apple Inc.

    “you can subscribe to the show on Apple Podcasts, Spotify, or your favorite podcast app.”

    ▶ Watch at ~1:14:00approx.
  • SpotifyRecommendssoftware · Spotify

    “you can subscribe to the show on Apple Podcasts, Spotify, or your favorite podcast app.”

    ▶ Watch at ~1:14:00approx.
  • Lenny's Podcast websiteMentionswebsite · Lenny Rachitsky

    “You can find all past episodes or learn more about the show at lennyspodcast.com.”

    ▶ Watch at ~1:14:00approx.

Spot an error or want something removed? Request a correction or removal.

Insights & moments

The myth-busts, hot takes, explainers, and tools worth keeping.

Myth Buster· 1

Myth Buster08:30

Why There Is No 'Data Wall'

The popular fear is that models will stop improving because they've exhausted the internet. Karina argues pre-training teaches a model to compress knowledge and model the world, but the real scaling now happens in post-training via reinforcement learning. Because you can teach a model an effectively infinite number of tasks, there is no data wall.

  • Pre-training is really about learning to compress knowledge and model the world
  • Post-training via reinforcement learning is not hitting a wall
  • There is an effectively infinite supply of tasks to teach the model

“we we went from like raw data sets from from pre-trained models to infinite amount of tasks that you can teach the model in the…”

Karina Nguyen · 10:00

“there's no data wall or whatever because there will be infinite amount of tasks. And that's how the model becomes extremely super intelligent.”

Karina Nguyen · 10:30
#synthetic-data#reinforcement-learning#post-training#scaling

Hot Take· 3

Hot Take10:30

The Real Bottleneck Now Is Evals, Not Data

Karina points out that models are saturating existing benchmarks — PhD-level tests like GPQA are being solved at human-expert levels. The limiting factor for progress is no longer data or intelligence, but the lack of hard enough evaluations to measure the frontier.

  • Models are saturating nearly all existing benchmarks
  • GPQA (PhD-level Q&A) is reaching the ~60-70% humans score
  • The frontier bottleneck is building good evaluations

“we are actually getting saturated on all benchmarks. So, I think the bottom neck is actually in evaluations.”

Karina Nguyen · 10:30
#evals#benchmarks#ai-research
Hot Take42:00

Soft Skills, Not Hard Skills, Are the Future of Work

Karina's core thesis: models are getting great at the 'hard' skills — coding, synthesis, writing — but struggle with aesthetics, taste, and genuine creativity. The durable human advantages are creative thinking, listening, prioritization and management. She even argues AI research progress itself is bottlenecked by management and compute allocation.

  • Teaching a model aesthetics, visual design and real creativity is very hard
  • Creative thinking, listening and prioritization become more valuable
  • AI research progress is bottlenecked by management and compute allocation
  • Hard skills (coding, design, writing) are the ones AI is taking over

“I think it's actually a really, really hard to teach the model how to be aesthetic or like do like visual or really good like…”

Karina Nguyen · 46:00

“AI research progress is bottlenecked by like management like research managers because you have like constrained set of compute”

Karina Nguyen · 46:00
#soft-skills#future-of-work#management#creativity
Hot Take49:30

AI Will Be Incredibly Good at Strategy

Lenny argues strategy is really just taking all available inputs, understanding the world, and forming a plan to win — exactly what a smart LLM should excel at. Karina agrees: models are strongest at connecting the dots across user feedback, dashboards and other inputs to co-create plans and recommendations, and self-improvement isn't far off.

  • Strategy is synthesizing all inputs into a plan to win
  • Models excel at connecting the dots across disparate sources
  • AI can aggregate feedback and dashboards to co-create plans and recommendations
  • Self-improving product development loops aren't far away

“isn't strategy just take all the inputs, all the data you have available, understand the world around you and come up with a plan to…”

Lenny Rachitsky · 49:30

“what models are really good at is like um like connecting the dots”

Karina Nguyen · 51:00
#strategy#ai-capabilities#data-analysis#self-improvement

Explainer· 4

Explainer06:30

Why Model Training Is More Art Than Science

Karina explains that the biggest misconception about how models are built is that it's a precise science. In practice, ensuring the highest-quality data for a desired behavior is one of the most important and least deterministic parts of the job. Debugging a model, though, looks a lot like debugging software.

  • Data quality is one of the most important levers in model training
  • Training is closer to a craft than a repeatable science
  • Debugging model behavior resembles debugging software

“Model training is more an art than a science and in a lot of ways”

Karina Nguyen · 06:30

“the way you debug models is actually very similar the way you debug software.”

Karina Nguyen · 06:30
#model-training#data-quality#ai-research
Explainer38:00

The Cost of Intelligence Is Collapsing

Karina names a core trend: the cost of reasoning and intelligence is dropping fast, and small models are becoming smarter than older large ones thanks to distillation. She saw this first-hand with Claude 3 Haiku, which was smarter than the much larger Claude 2 — pointing to a world of cheap, fast, highly capable small models.

  • The cost of reasoning and intelligence is dropping sharply
  • Distillation makes small models smarter than older large models
  • Claude 3 Haiku was smarter than the much bigger Claude 2

“the cost of reasoning and intelligence is drastically going down.”

Karina Nguyen · 38:00

“Smart small models are becoming even smarter than like large models and that's because of like the distillation research.”

Karina Nguyen · 38:30
#distillation#small-models#cost-of-intelligence#claude
Explainer39:00

AI Will Unblock Work That Was Bottlenecked by Intelligence

As intelligence gets cheap and abundant, Karina argues work that was previously limited by access to expertise gets unblocked — healthcare, education, and scientific research. Lenny cites a New York Times study where ChatGPT alone outperformed both doctors and doctors using ChatGPT.

  • Cheaper AI gives builders and developers far more access to intelligence
  • Feeding symptoms to ChatGPT can approximate access to a doctor
  • A NYT study found ChatGPT alone beat both doctors and doctors-with-ChatGPT
  • Education and scientific research see massive implications

“all the work that has been like bottlenecked by the intelligence will be kind of like unblocked.”

Karina Nguyen · 39:00

“there's a New York Times story about that where they compared doctors to doctors using ChatGPT to just ChatGPT and just just ChatGPT was the…”

Lenny Rachitsky · 40:00
#healthcare#education#ai-access#automation
Explainer48:00

Why Models Still Can't Do Taste

Karina explains that idea generation and discriminating good visual design or art are areas where models haven't learned enough from people to judge well. The deeper reason: there aren't many people with exceptional taste whose judgment is accessible to models to learn from. Lenny sums it up: 'there's not enough of you yet.'

  • Models haven't learned enough human examples to discriminate good design/art
  • New reasoning paradigms may improve writing over time
  • The scarcity of accessible taste-makers to learn from is the core limit

“when it comes down to like idea generation or like um discriminating of like what is a good like visual design and art? I feel…”

Karina Nguyen · 48:30

“Basically, there's not enough of you yet.”

Lenny Rachitsky · 49:00
#taste#creativity#visual-design#model-limits

Story· 3

Story07:00

The Model That Got Confused About Having No Body

During Claude 3 training, the team taught the model self-knowledge (it has no physical body) while also teaching it function calls like setting an alarm. The conflicting signals confused the model about whether it could actually perform physical-world actions, sometimes causing it to over-refuse. It illustrates the constant trade-off between being helpful and being safe.

  • Self-knowledge data ('you have no body') collided with function-call data ('this is how you set an alarm')
  • The confusion sometimes made the model over-refuse requests
  • Training balances helpfulness against harm across diverse scenarios

“the model would get like extremely confused uh about like whether it it can set an alarm in the but it doesn't have a body…”

Karina Nguyen · 07:30
#claude-3#model-behavior#anthropic#refusals
Story13:00

How Canvas Broke the Wall Between Research and Engineering

Karina describes pitching Canvas at OpenAI and watching a cross-functional team form almost out of nothing. It was, in her telling, the first OpenAI project where researchers and applied engineers worked together from the very start of the product cycle — moving ChatGPT from a chatbot toward a genuine collaborator.

  • Canvas turned ChatGPT from a chatbot into a collaborative agent
  • A team of engineers, designers, PMs and researchers formed rapidly
  • It was OpenAI's first product built with research + engineering together from the beginning

“So, going from like being a chatbot to more of a collaborative agent and a collaborator”

Karina Nguyen · 13:00

“Canvas is like one of the I would say like the first project at OpenAI where researchers and applied engineers started working together from the…”

Karina Nguyen · 13:30
#canvas#openai#product-development#collaboration
Story57:00

Claude in Slack: One of the First Tool-Using Products

Karina recalls early Anthropic days when Claude in Slack was one of the first tool-using products — Claude could operate in your workplace, summarize threads on demand, and every Monday auto-summarize entire channels. It created a social, Midjourney-on-Discord-like element where people learned to prompt together, before being sunsetted post-ChatGPT.

  • Claude in Slack was one of the first tool-using products
  • It could summarize threads and, weekly, whole channels automatically
  • It created a social, community learning element around prompting
  • It was sunsetted around 2023 as focus shifted to consumer/enterprise

“Claude and Slack was actually one of the first like tool using like products.”

Karina Nguyen · 57:30

“every Monday Claude would just like summarize the entire channel.”

Karina Nguyen · 58:00
#claude#slack#anthropic#agents

Q&A· 1

Q&A53:30

Anthropic vs OpenAI: Craft vs Risk-Taking

Having worked at both labs, Karina says they're more similar than different but with real cultural nuances. Anthropic taught her deep care and craft around moral training and hardcore prioritization; OpenAI is more innovative, risk-taking and bottoms-up. She frames Claude's 'librarian' personality as a reflection of its creators.

  • Anthropic emphasizes moral craft and hardcore prioritization
  • OpenAI is more innovative, risk-taking and bottoms-up
  • A model's personality reflects its creators (Claude as a 'librarian')
  • The two labs are one big community, not enemies

“I think what I have learned from Anthropic is this like real care and craft towards like moral behavior, moral craft, moral training.”

Karina Nguyen · 54:00

“OpenAI is like much more innovative and much more like risk takers in terms of like product or like research”

Karina Nguyen · 56:00
#anthropic#openai#company-culture#claude

Tool· 1

Tool67:30

Operator: An Agent With Its Own Virtual Computer

Karina describes Operator, an OpenAI agent that completes tasks inside its own virtual computer — e.g. ordering a book on Amazon by driving a browser itself. She explains why this is genuinely hard: the models operate on pixels rather than language, and visual perception is much harder to scale than text.

  • Operator completes tasks inside its own virtual environment/browser
  • It can carry out real tasks like buying a book on Amazon
  • It's hard because agents act on pixels, not language
  • Deriving human intent and knowing when to ask follow-ups is a key challenge

“it's basically an agent that can complete the task in its own like virtual computer like in its own virtual environment.”

Karina Nguyen · 67:30

“Much of it is like because right now the models it operating on like pixels instead of like language or what not.”

Karina Nguyen · 70:00
#operator#computer-use#agents#openai

Takeaway· 2

Takeaway24:30

Prompting Is the New Way to Prototype Products

While building the 100K-context file-upload feature at Anthropic, Karina prototyped it just by prompting the model in a local browser — and people immediately wanted the API. That's when it clicked that prompting is a genuine product-development and prototyping tool for designers and PMs, not just a way to chat.

  • The 100K-context file upload was prototyped purely by prompting in a browser
  • Prompting lets designers and PMs prototype features without heavy engineering
  • Small 'micro experiences' (like personalized starter prompts and titles) can be built via prompting

“prompting is like a new way of like product development and like prototyping for designers and for like product managers.”

Karina Nguyen · 24:30
#prompting#prototyping#product-development#anthropic
Takeaway43:30

Build for the Model You'll Have, Not the One You Have

Karina's advice for builders: design product ideas for where the models are going, not where they are today. It doesn't matter if the model is good enough right now — build so that by the time models get really good, your product works really well. Claude artifacts and early Canvas are examples of ideas that landed as capability caught up.

  • Design for future model capability, not current limitations
  • A product built ahead of the curve pays off as models improve
  • Fast-iterating startups win by inventing new ways to train and listening to users

“you kind of want to like build for the future. So it's like it doesn't necessarily matter whether the model is good or not good…”

Karina Nguyen · 43:30
#product-strategy#building#future-planning