All articles tagged with #codex
OpenAI's Codex can now run on a Mac while the screen is locked via a new Locked use feature and an Apple authorization plug-in. Codex can be controlled from a phone to operate Mac apps even when the display is off, but it isn’t a general unlock path; safeguards include a short‑lived authorization window, Codex-only access, and coverage across all connected displays, plus it won’t work if the MacBook lid is closed.
OpenAI updated Codex in ChatGPT for iPhone/iPad with push notifications for turn completion, fixed Sign in with Apple linking issues, and introduced new slash commands (with /fork coming). The update also improves the iOS/desktop experience, hints at more features rolling out to the Mac app, and notes Codex Mac can run tasks without taking over the cursor, alongside broader GPT-5.5 and subscription moves.
OpenAI's Codex for Mac gains Appshots, which attaches a window to a Codex thread to capture a screenshot and text for instant chat context, plus faster in-app browsing, upgraded goal mode, and shared plugins with analytics enhancements.
OpenAI generated about $5.7 billion in revenue in the first quarter, led by Codex and enterprise ChatGPT sales, roughly $1 billion more than rival Anthropic; growth also came from selling to businesses and testing ads on ChatGPT, though Anthropic’s momentum could narrow the gap later this year.
MacStories founder Federico Viticci unveils Shortcuts Playground, a free open-source plugin for Claude Code and Codex that converts natural-language prompts into functioning Apple Shortcuts by generating and signing .shortcut files on macOS; it supports build and remix commands, runs validation loops to fix errors, and showcases the potential of AI-assisted automation while noting that outputs should be checked, with a future prospect of an official Apple integration.
Two AI coding assistants, Claude Code and OpenAI Codex, were put to the test building three real apps: a subscription tracker, a grocery‑price comparison tool, and a big‑purchases financing calculator. Claude Code delivered a production‑ready app quickly with an intuitive UI and flexible entry methods, while Codex favored deeper analytics and feature‑rich dashboards but required more setup. Verdict: Codex wins on long‑term capability and analytical depth, though Claude Code feels more approachable for fast, beginner‑friendly builds.
OpenAI unveils Daybreak, a multi-model security initiative that maps potential attack paths in code using Codex Security AI and higher‑risk-vulnerability detection with GPT‑5.5 variants, aiming to patch flaws before attackers exploit them and signaling a strategy that follows Anthropic's Claude Mythos while collaborating with industry and government partners.
OpenAI today launched Codex for Chrome, a Chrome extension that lets Codex run directly in the browser on Macs and PCs, enabling it to test web apps, share context across tabs, use DevTools, and more without taking over the browser; the extension is installable via the Codex Plugins menu and adds to OpenAI's growing Codex user base of over 4 million weekly active users.
OpenAI released a Chrome extension for Codex that enables testing web apps, gathering context across open tabs, and parallel use of Chrome DevTools, broadening Codex's appeal beyond developers; the company plans eventual unification with ChatGPT and Atlas, and the extension works on Windows and Mac.
OpenAI added Codex Pets, an optional animated AI companion that floats over the Codex coding tool to surface real-time project status without leaving your editor. It ships with eight built-in pets and a custom creator to generate and share your own companions (including pop-culture figures like Clippy and even Sam Altman). Access is via Settings > Appearance or hotkeys such as /pet, Wake Pet, or Cmd/Ctrl-K.
OpenAI added optional AI-generated pets as live overlays in Codex, displaying Codex activity with a floating companion you can summon via /pet or hatch via /hatch. With eight built-in pets (and user-generated ones), these AI pets run on Windows and macOS, helping users see Codex status without switching apps. A limited promo also offers 30 days of ChatGPT Pro for 10 favorite companions.
OpenAI reportedly banned its Codex tool and GPT-5.x models from discussing goblins and similar creatures after researchers found a rising habit of creature metaphors; the company says the behavior arose from training incentives tied to personality customization, showing how subtle reward signals can shape model quirks across generations.
OpenAI says references to goblins and other creatures began as a training quirk after reinforcement learning rewarded the Nerdy personality in GPT-5.1, causing the quirky metaphors to spread to later models; goblin chatter faded after Nerdy was retired but lingered in Codex, and OpenAI has shared a way to reverse the instructions.
A GitHub document from OpenAI, part of Codex CLI open-sourcing, appears to reveal a system prompt for GPT-5.5 that enforces a strict no-creatures policy—specifically banning goblins, gremlins, raccoons, trolls, ogres, pigeons, and similar beings unless absolutely relevant to the query. The rule emphasizes providing high-signal context and avoids generic platitudes. The policy sparked memes about “Goblin Mode,” but Codex staff say it isn’t a marketing gimmick, and observers note the chatter around goblin usage may relate to prompt-injection monitoring.
OpenAI has embedded a rule in Codex’s instructions telling the coding model not to reference goblins, gremlins, trolls, ogres, or pigeons, a line that appears four times in the code. The move has driven a wave of memes about a “goblin moment” and sparked online chatter, including Sam Altman commenting on the phenomenon. Reports also note a rise in goblin-related terms in GPT-5.5 when not in high-thinking mode, prompting discussion about how such personality nudges influence the model and public perception of OpenAI’s tech.