Episode

Codex from 0 to 10M Users: Building ChatGPT Work — Akshay Nathan, OpenAI

Podcast: Latent Space: The AI Engineer Podcast
Published: Jul 28, 2026
Duration seconds: 4168
Processing state: processed
Canonical source: https://www.latent.space/p/chatgpt-work
Audio: https://api.substack.com/feed/podcast/208716574/1644ff34767eacc3e04d93b5c8621905.mp3
JSON: /v1/public/podcasts/latent-space-ai-engineer/episodes/codex-from-0-to-10m-users-building-chatgpt-work-akshay-nathan-openai
Markdown: /podcast/latent-space-ai-engineer/codex-from-0-to-10m-users-building-chatgpt-work-akshay-nathan-openai.md

Actions

POST https://stenobird.com/v1/public/podcasts/latent-space-ai-engineer/episodes/codex-from-0-to-10m-users-building-chatgpt-work-akshay-nathan-openai/transcription-requests
Idempotently request low-priority transcript generation for this episode.
GET https://stenobird.com/podcast/latent-space-ai-engineer/codex-from-0-to-10m-users-building-chatgpt-work-akshay-nathan-openai.md
Read the agent-friendly Markdown representation of this episode resource.

Summary

OpenAI's Codex is evolving from a developer tool into a 'superapp' for general knowledge work. Akshay Nathan explains how the transition from coding agents to productivity agents is driving massive user growth.

Topics

OpenAI
ChatGPT Work
Codex
AI Agents
Productivity Engineering
Knowledge Work
Software Development
No-code

Highlights

Main idea: The expansion from coding agents to knowledge work agents is the next frontier for AI utility
Practical takeaway: Focus on 'at bats'—the ability to rapidly iterate from idea to validation—rather than just output volume
Failure mode: Avoid the trap of conflating 'motion' (activity enabled by AI) with actual 'progress' (achieving deliberate goals)
Trend: Non-developers are adopting Codex-powered tools at a rate 3x faster than traditional engineers
Strategic shift: Moving from 'telling' users how to use tools to 'showing' them through agentic capabilities like artifacts and computer use

Chapters

1:00 The Thesis of No-Code: Reflecting on the transition from Airtable to OpenAI and the dream of making database primitives accessible via LLMs.
6:00 The Codex Inflection Point: Discussing the unexpected surge in adoption among non-developers and the move toward a 'superapp' model.
17:00 The Role of Model Intelligence: How improvements in model quality and cost allow for more complex, automated productivity workflows.
27:00 Scaling Infrastructure and Product: The engineering challenges of building the infrastructure necessary to support massive user growth and agentic features.
38:00 The Future of Agentic Work: How agents will eventually have access to a user's entire digital environment to create and manage artifacts.
1:04:00 Motion vs. Progress: A critical distinction for managers: using AI to drive meaningful outcomes rather than just increasing the volume of tasks.