Why AI is going vertical (again) | Dianne Penn (Anthropic)
Dianne Penn is Head of Product for Anthropic’s AI Research and Labs teams. She joined in 2023 as Anthropic’s first technical product manager, when the entire product team was five engineers, and has since helped ship every model from Claude 2 through Fable, and helped incubate Claude Code, MCP, Skills, computer use, tool use, and reasoning. Before Anthropic, she helped build Alexa’s AI at Amazon and, before that, traded high-yield bonds at JP Morgan Chase. *In our in-depth conversation, we discuss:* 1. What Anthropic’s early days were like 2. The inflection points that turned Anthropic from an underdog into the fastest-growing company in history 3. How exactly Claude got so good at coding 4. The eval-driven development loop her team is pioneering 5. How to find joy in AI when everything is moving this fast 6. Why Claude’s willingness to push back is key to its success 7. Where human judgment remains irreplaceable *Brought to you by:* WorkOS—Make your app enterprise-ready, with SSO, SCIM, RBAC, and more: https://workos.com/lenny Mercury—Radically different banking, now with Command: https://mercury.com/ *Episode transcript:* https://www.lennysnewsletter.com/p/anthropics-first-technical-pm-on *Archive of all Lenny's Podcast transcripts:* https://www.dropbox.com/scl/fo/yxi4s2w998p1gvtpu4193/AMdNPR8AOw0lMklwtnC0TrQ?rlkey=j06x0nipoti519e0xgm23zsn9&st=ahz0fj11&dl=0 *Where to find Dianne Penn:* • LinkedIn: https://www.linkedin.com/in/dianne-na-penn 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 (02:31) Early Anthropic days (08:55) Big milestones (13:50) Inside the exponential (20:02) Token maxing (23:30) Anthropic Labs and the incubation model (27:30) How the research role works (31:35) How to become a top researcher (35:18) Frontier model safeguards (39:38) Hiring in the AI era (44:16) Building an eval set (47:48) Evals vs PRDs (49:55) The importance of hands-on leadership (52:46) Finding joy in AI (58:10) How Dianne uses Claude (01:01:05) Avoiding overreliance on AI (01:03:50) The constitution that makes Claude better (01:07:11) AI writing and verification (01:11:40) Where human brains will continue to be valuable (01:14:10) Navigating AI with kids (01:16:26) Alignment, the future of the PM role, and burnout (01:21:54) Lightning round and final thoughts *Referenced:* • Anthropic: https://www.anthropic.com • Golden Gate Claude: https://www.anthropic.com/news/golden-gate-claude • Dario Amodei’s website: https://darioamodei.com • Scaling Laws and Interpretability of Learning from Repeated Data: https://www.anthropic.com/research/scaling-laws-and-interpretability-of-learning-from-repeated-data • Tokenmaxxing: How Top Builders Use AI To Do The Work Of 400 Engineers: https://www.ycombinator.com/library/Pa-tokenmaxxing-how-top-builders-use-ai-to-do-the-work-of-400-engineers • Garry Tan on X: https://x.com/garrytan • Anthropic co-founder on quitting OpenAI, AGI predictions, $100M talent wars, 20% unemployment, and the nightmare scenarios keeping him up at night | Ben Mann: https://www.lennysnewsletter.com/p/anthropic-co-founder-benjamin-mann • Anthropic’s CPO on what comes next | Mike Krieger (co-founder of Instagram): https://www.lennysnewsletter.com/p/anthropics-cpo-heres-what-comes-next • Introducing Labs: https://www.anthropic.com/news/introducing-anthropic-labs • Louis CK | about airplane Wi Fi: https://www.youtube.com/watch?v=me4BZBsHwZs • What happens after coding is solved? | Fiona Fung (Manager of the Claude Code and Cowork Teams): https://www.lennysnewsletter.com/p/building-the-most-ai-pilled-engineering • The Anthropic Hive Mind: https://steve-yegge.medium.com/the-anthropic-hive-mind-d01f768f3d7b • How to build a company that withstands any era | Eric Ries, Lean Startup author: https://www.lennysnewsletter.com/p/how-to-build-a-company-that-withstands • Fallout on Prime Video: https://www.amazon.com/dp/B0CN4GGGQ2 • Fallout (video game): https://fallout.bethesda.net • Claude Tag: https://www.anthropic.com/news/introducing-claude-tag *Recommended books:* • Crucial Conversations: Tools for Talking When Stakes Are High: https://www.amazon.com/dp/0071771328 • How to Raise an Adult: Break Free of the Overparenting Trap and Prepare Your Kid for Success: https://www.amazon.com/How-Raise-Adult-Overparenting-Prepare/dp/1627791779 • Incorruptible: Why Good Companies Go Bad... and How Great Companies Stay Great: https://www.amazon.com/dp/B0FWZZBPZB _Production and marketing by https://penname.co/._ _For inquiries about sponsoring the podcast, email podcast@lennyrachitsky.com._ Lenny may be an investor in the companies discussed.
Read Video · Transcript & Insight
This episode has a full transcript + AI insights
Free account · no card needed · 150 credits on signup, enough to unlock this episode
- 📄 Full transcript with timestamps
- ✨ AI summary, keywords & mind map
- 💡 Key takeaways & quotes
- Speaker 1
- Speaker 2
- Speaker 1
- Speaker 2
- Speaker 1
- Speaker 2
- Speaker 1
- Speaker 2
Episode Timeline
Dianne Penn's early days at Anthropic and the company's culture
- Dianne joined Anthropic in 2023 as the first technical PM when the product team had only five engineers, and the API business was run by a single engineer.
- The early culture was very bottoms-up and mission-driven, exemplified by the rapid 24-hour launch of Golden Gate Claude, a quirky feature that dialed up the model's obsession with the Golden Gate Bridge.
- Golden Gate Claude, though reaching only about 2,000 users, was a hidden inflection point that showed the team could build unique, research-driven product experiences at startup speed.
Key inflection points: Opus 3 and the focus on coding
- Training and testing Opus 3 in late 2023 was a major milestone when Anthropic was still under 200 people; it built foundational trust across research and product teams that persists today.
- In 2023, nobody associated Anthropic with coding, but Dianne noticed users starting to write long-form code with models, not just autocomplete, which led to a small training change that made Opus 3 differentiate on coding.
- Opus 3's coding capability brought early Claude enthusiasts and developers, proving that a relatively small training adjustment could create competitive advantage.
Opus 4.5 and the synergy between frontier models and products
- Opus 4.5, released a year later during winter break, was magical because it combined a frontier model with a great product experience like Claude Code.
- The team's saying 'you need frontier products to have frontier models' captures how Claude Code and Opus 4.5 amplified each other's adoption and impact.
- Opus 4.5 reached a level of intelligence where users could run tasks end-to-end in an agentic manner, marking a clear inflection point.
Key Concepts
- Anthropic— The AI company where Dianne Penn works, central to the episode.
- Claude— Anthropic's AI model, discussed extensively.
- product management— The role and how it's evolving in the AI era.
Notable Quotes
We actually have a saying on the team of evals are the new PRDs.
💡— Overturns the traditional product management approach where PRDs were central, showing how AI product work is fundamentally different.
If you are willing to spend $100,000 a year right now in tokens, you are living the way somebody in 2028 is going to live.
🤯— Reveals a surprising alpha opportunity: heavy token spend now gives a glimpse of the future, challenging the notion of cost efficiency.
Actionable Takeaways
📋Product Management in AI
Evals are the new PRDs; product managers must define user pain points through evaluative datasets.
This week, identify one recurring user complaint about your AI product and create a small eval set (10-20 examples) to measure it.
Sweat tokens as much as pixels; deep analysis of model failures (hallucinations, overconfidence) is key.
Review 5-10 failed user interactions from logs and categorize the failure types (e.g., schema, hallucination).
💡Innovation Culture
Labs teams thrive on small, autonomous pods with strong opinions loosely held.
Start a 'labs' style project: pick one ambiguous bet, assign one engineer, and set a 2-week deadline for a prototype.
Work in public internally; share experiments and learnings across teams.
Create a Slack channel where team members post daily AI experiments and insights.
Transcript and insights are AI-generated and may contain errors. Accuracy depends on audio quality and speaker clarity — if something looks off, the original audio is always the source of truth.
Episodes & Videos Ready to Read
Podcast episodes

Raising a Dog & Mastering Calm Assertive Energy | Cesar Millan
Huberman Lab
Jul 6, 20262:38:23EN
The Total Christian Man - 1 - Major Thomas
Torchbearers Archives
Sep 3, 202555:23EN
How to get past the "messy middle" of a big change | Mary Martin
TED Talks Daily
Jul 24, 202614:38EN
EP79 你不是不自律,你只是被劫持了注意力
纵横四海
Mar 30, 20264:28:52ZH
Zweifel-Los! 15 - Das Leben wie ein Profi leben - Peter Ballmer
Zweifel-Los!
Jul 24, 202424:33DE
Café com Deus Pai | 24 de maio
Café Com Deus Pai | Podcast oficial
May 24, 20264:44PT
Videos

Gemini Robotics – AI for the Physical World, with Keerthana and Ted of Google DeepMind
Cognitive Revolution "How AI Changes Everything"
May 17, 20251:48:11EN
Staphylococcus: Aureus, Epidermidis, Saprophyticus
Ninja Nerd
Oct 21, 20211:01:18EN
GPUs, TPUs, & The Economics of AI Explained | Gavin Baker Interview
Invest Like The Best
Dec 9, 20251:28:22EN
Firma bez šéfov: Funguje to? - Money Talk 113 s Ferom Baníkom
Milan Dubec
Aug 4, 202654:10SK
2026/08/24(一) 輝達伺服器傳漲價15%:AI成本暴增,成本誰吸收?
財女珍妮
Aug 24, 202630:06ZH-Hant