The PodKnowledge|← All Episodes
Based on Lenny's Podcast data
Lenny's Knowledge Sketch

700 Million Weekly,
and Rising

Nick Turley
Head of ChatGPT, OpenAI · Fmr. Instacart, Dropbox · Helped originate ChatGPT and take it from 0 to 700M weekly users
AUG 9 2025
The Scale

Ten Percent
of the World, Weekly

CHATGPT WEEKLY ACTIVES
"About 10% of the world population uses every week. With scale comes responsibility. This model has taste."
  • 700 million weekly active users, and Nick expects to cross a billion "soon"
  • Roughly one in ten people on Earth touches ChatGPT in any given week
  • Nick joined OpenAI three years ago; he helped originate ChatGPT and take it from zero to here
  • GPT-5 launched the week this episode was recorded — his phrase: "the vibes are good"
Framework

Three Businesses,
One Product

CONSUMERDEVELOPER APIENTERPRISE
700M
weekly active users
5M
business subscribers (from 3M a month or two ago)
4M
developers on the API platform
  • Enterprise was launched because ChatGPT was starting to get banned in companies over privacy and deployment
  • Debate at the time: launch Enterprise, or launch an iOS app — that's how small the team still was
  • Organically hit 90% of the Fortune 500 shortly after enterprise shipped
  • Cross-surface primitives (projects, history search, sharing) largely serve all three businesses at once
Enterprise-only work is real workHIPAA, SOC 2, and the rest are non-negotiable if you want to be a serious enterprise player. Nick treats it as the tax of taking the business seriously.
The Retention

The Smiling Curve

  • 1-month retention: around 90%. Extremely rare for a consumer product
  • 6-month retention: around 80%. Also extremely rare
  • The shape: people try ChatGPT, drift, then come back months later and use it MORE
  • Why it matters: a smile curve means every early user is a compounding asset, not a one-time try
15-minute user interviews, all week

The weeks after launch Nick booked back-to-back 15-minute user interviews. He kept going until he could predict what the next person would say. Only then did he stop.

Users taught the team the product

Every bespoke idea the team prototyped, users bent to something else. The pattern kept saying "generic assistant, please" — which is what ChatGPT became.

Playbook

How Nick Actually Operates

  • Set the culture at pace and urgency, even when you are already the fastest-growing product ever
  • Nick learned that at Instacart, all-hands-on-deck during the pandemic; brought it to OpenAI
  • Interview users back-to-back until you can predict what the next one will say
  • When users bend every bespoke prototype toward one general behavior, follow them; don't hold onto the bespoke
The SA Server originThe code base still called "SA Server" — short for Super Assistant. It started as a hackathon around GPT-4, testing whether OpenAI should ship a consumer product to close the developer feedback gap.
Contrarian

Nick's Contrarian Reads

✗Call it a "super assistant"INSTEAD →✓ Nick avoids the label — "assistant" is limiting, personified, and not relatable outside of Silicon Valley managers with actual assistants.
✗Follow the model, users will catch upINSTEAD →✓ Users notice the vibes, not the benchmarks. Most ChatGPT users don't know which model powers the app.
✗Pick one business: consumer OR enterprise OR APIINSTEAD →✓ Do all three when the primitives overlap; only build enterprise-specific work (HIPAA, SOC 2) where it's non-negotiable.
✗Take your sweet time, think hardINSTEAD →✓ Thinking matters, but hustle beats hesitation at this stage. Nick picked up the pace at Instacart, and it now runs his OpenAI team.
Based on Nick Turley's episode on Lenny's Podcast. All ideas on this page are from the episode.Watch on YouTubeFollow @nickaturley on X
GO DEEPER IN THE EPISODE
04:52 GPT-5 launch09:13 The vision for ChatGPT and AI assistants13:52 The early days of ChatGPT17:14 The success and impact of ChatGPT20:44 Product development and iteration23:11 Maximally accelerated: the OpenAI approach26:17 Retention and user engagement33:42 The future of chat interfaces36:31 The evolution of ChatGPT38:52 Subscription model and pricing strategies42:10 Enterprise adoption and challenges44:10 Balancing multiple product lines52:13 Emergent use cases and user feedback01:02:15 OpenAI’s unique product development approach01:05:07 The importance of team composition01:08:50 Balancing speed and quality in AI development01:14:23 The role of evals in product development01:16:13 The future of AI-driven content and GPTs01:21:51 Philosophy and product leadership01:23:47 Career journey and advice01:27:49 Lightning round and final thoughts
Every sketch here began as a transcript.

Send us yours and we’ll draw it. $1 a transcript.

Sketch mine →
𝕏︎ X / Twitterin LinkedIn📸 Instagram🔗 Copy link
0:00