Fetching from the wire…
Top 5 · 2026-06-23 · source-backed
In 2023 Samsung banned generative AI internally after engineers leaked source code into ChatGPT. On June 21, 2026, OpenAI confirmed Samsung is deploying ChatGPT Enterprise and Codex to all employees in South Korea and across its worldwide Device eXperience division. That's a full reversal, and it's one of OpenAI's largest rollouts to date. (OpenAI)
The detail that matters most: Codex is going to non-developers. R&D, manufacturing, corporate functions. Not just the software teams. OpenAI cited ~5M weekly Codex users and roughly 800% growth in Korea since February. Whatever you think of those figures coming from the vendor, the direction is clear. A coding agent is being handed to people whose job title doesn't say "engineer," at a company paranoid enough about IP that it banned the category two years ago.
Connect this to the Anthropic study above and a pattern shows up. If domain expertise, not coding background, is what decides success, then handing Codex to a manufacturing engineer who deeply understands a process line isn't reckless. It might be the highest-leverage move Samsung makes this year. The manufacturing expert owns the "what." Codex handles the "how." That's the same thesis, now playing out as a 270,000-person enterprise bet instead of a research finding.
What changed Samsung's mind isn't in the announcement, but I'd guess it's the same thing that changes every enterprise's mind: the leak risk got contained by enterprise tenancy and data controls, and the productivity delta got too large to ignore. The 2023 ban was about a consumer product with no guardrails. ChatGPT Enterprise is a different deployment surface.
For builders, the signal is that the enterprise "we don't allow AI" position is collapsing faster than the policy teams can update their handbooks. If a chaebol that banned the tech over a source-code leak is now pushing Codex to non-engineers worldwide, the "is it allowed" conversation at smaller companies is basically over. The new conversation is "which agent, scoped how, billed by what." Plan your tooling and your security story for a world where the answer to "can I use a coding agent at work" is yes by default.
Each link below shares sources, entities, or timing with this story.
Announced around June 18–20, the Global Admin Console now breaks down ChatGPT and Codex credit consumption by user, product, and model, with workspace defaults, per-group limits, and individual overrides. It targets CFOs staring at escalating per-seat AI bills from power users...
usehax.dev (104 points on HN) is a single native MIT-licensed binary that starts instantly on a few megabytes, auto-discovers model and runtime capabilities, and talks to OpenAI-compatible endpoints, Anthropic-compatible endpoints, Codex via ChatGPT subscription, OpenRouter, a...
The IDE market is fragmenting, and this week drew the sharpest lines yet. Cursor 3 launched as a rebuilt agent-orchestration platform in Rust and TypeScript, replacing the VS Code fork with an Agents Window for dispatching and monitoring multiple AI coding agents. Anysphere hi...
On Latent Space July 28, OpenAI core product engineering lead Akshay Nathan said Codex and ChatGPT Work combined reached 10 million users within two weeks of the July 9 launch, with monthly actives up more than 10x since January 2026. The number that should reframe your produc...
On July 9 the standalone Codex app merged into ChatGPT desktop for macOS and Windows, keeping a dedicated coding surface alongside Chat and Work modes, plus inline diff editing, side-panel PR review, faster Computer Use on GPT-5.6, and multi-repository projects. The tiering: S...
OpenAI announced March 19 it's acquiring Astral, the company behind uv (126M downloads/month), ruff (the Python linter/formatter), and ty (type checker). The team integrates into Codex. The strategic logic is explicit: move Codex beyond code generation into full-lifecycle deve...
MindPattern daily
One email a day at 7 AM. Sources and a take on every story. Unsubscribe anytime.