{
"$type": "site.standard.document",
"bskyPostRef": {
"cid": "bafyreicdogd3tv7byxhr5dbcikth5zs6drozxqnnoxgxrct6zp5exb3hqe",
"uri": "at://did:plc:lk3jfj3zq4k4wxnk474axylu/app.bsky.feed.post/3mo3mtxxm2v32"
},
"path": "/t/ive-been-experimenting-with-codex-cli-lately-and-it-feels-like-were-moving-from-ai-autocomplete-to-ai-teammate/1383480#post_1",
"publishedAt": "2026-06-12T10:59:20.000Z",
"site": "https://community.openai.com",
"textContent": "The interesting part isn’t that Codex can write code.\n\nIt’s that it can:\n\n * Understand an existing codebase\n * Make multi-file changes\n * Run commands\n * Work directly from the terminal\n\n\n\nThat said, I’m curious how people are using it in real projects.\n\nAre you treating Codex CLI as:\n\n 1. A coding assistant?\n 2. A code reviewer?\n 3. A debugging partner?\n 4. An autonomous agent for repetitive tasks?\n\n\n\nMy biggest takeaway so far: the bottleneck is no longer writing code, it’s deciding what should be automated and what still needs human judgment.",
"title": "I've been experimenting with Codex CLI lately, and it feels like we're moving from \"AI autocomplete\" to \"AI teammate.\""
}