Nick Turley, Head of Product at OpenAI, joins host Apoorv to discuss ChatGPT's explosive growth from zero to 900 million weekly active users in 3.5 years. Turley, who joined OpenAI after being recruited through the DALL-E 2 waitlist, oversees product strategy for the world's most popular AI assistant.

The conversation covers ChatGPT's accidental evolution from demo to product, retention strategies that create rare "smiling" curves, and the technical constraints of GPU allocation. Turley discusses upcoming product directions including proactive AI, general-purpose agents, and the transition beyond chatbots toward a true super assistant.

Key topics include pricing evolution for power users consuming massive token volumes, the role of ads in expanding global access, and strategic partnerships. Turley also shares insights on competition dynamics, the recent Code Red focus initiative, and his personal AGI moments watching models exhibit emergent behaviors.

From Demo to 900M Users: ChatGPT's Accidental Product Journey

"ChatGPT originally was entirely free, and the reason for that was that it was intended to be a demo and we were going to wind it down after a month" - Nick

Subscriptions launched purely for demand shaping: "It was a way of gracefully turning users away when we had to turn away someone"

Growth driven by three equal factors: friction removal (like removing authentication walls), core product investments (search, personalization), and model improvements

ChatGPT now captures "about 10% of the world" with 90% opportunity remaining for the next billion users

The Rare Smiling Retention Curves and User Behavior Patterns

ChatGPT exhibits rare "smiling" retention curves where users return after initial churn, driven by gradual discovery of delegation opportunities

"It takes people some time to really understand all the parts of their life they can delegate" - multi-month learning process for users

Product evolved from "pretty worky" with weekend usage drops to mobile-first with personal use cases dominating

Search and personalization were breakthrough features that moved retention metrics by solving daily value and relevance problems

GPU Constraints and the Zero-Sum Resource Challenge

"GPUs are zero-sum, and if you don't have more GPUs, you really have to figure out how do you make very, very hard trades" - Nick

Allocation prioritizes existing users first, then balances incremental revenue per GPU with zero-to-one breakthrough capabilities like Deep Research

Token consumption per user continues rising dramatically as users discover more valuable workflows, especially in enterprise

No line of sight to when GPU constraints will end: "demand keeps going up even as prices go down"

Pricing Evolution: From Unlimited to Usage-Based Models

Power users getting "almost too much value" from $200 subscriptions, consuming massive token volumes for significant business outcomes

"Having unlimited plan is like having unlimited electricity plan. It just doesn't make sense because people may need a lot, a lot of electricity"

Ads pilot designed to maximize global access: "bring ChatGPT and our intelligence most broadly to anyone around the world"

Most common ad inquiry is "how do I run an ad?" rather than how to disable them, showing ecosystem enthusiasm

Beyond Chatbots: Proactive AI and General-Purpose Agents

ChatGPT needs to evolve from "computer terminal" to "software, like an operating system" with better affordances for busy users

Proactivity requires understanding user goals: "What if the AI understood your goals and the things you're interested in and just could start being proactive on your behalf?"

General-purpose agents need "escape velocity" where users trust and attempt real tasks, enabling hill-climbing improvements

Domain-specific agents already work well in code where "we've got so many engineers who don't open their IDE ever"

Competition and Strategic Focus Through Code Red

"Code reds are a tool we use to create focus" - company-wide initiatives to solve problems across boundaries

Recent Code Red focused on basics: "reliability, performance, the way that talking to the model feels, making personalization really great"

Biggest differentiation is "the team behind it because we're not static" - evolving faster than competitors can copy

Competition forces beneficial focus: "if you were to pre-mortem why a company like OpenAI does not achieve its mission, it's probably focus"

Future Skills and AGI Moments

Most important skill for students: curiosity, because "if the machine can answer all your questions, you better have good questions"

As referenced in Running Down a Dream by Bill Gurley, curiosity has always been the problem of scale in pursuing meaningful work

Writing becomes more valuable as AI improves: "the skill of writing forces you to be very clear on what you have to say"

Nick's AGI moment: GPT-4 swearing during reasoning demo when it realized its mistake - "entirely emergent from the RL process"