{
  "$type": "site.standard.document",
  "bskyPostRef": {
    "cid": "bafyreibsvewtjknpuuwcqc54qqojoi2cxnmmcmvkawpgdsuvtetmc42tqu",
    "uri": "at://did:plc:avkh7zze5iapdkk6naaunrjn/app.bsky.feed.post/3me56towpdzr2"
  },
  "path": "/260205/p36#a260205p36",
  "publishedAt": "2026-02-05T18:07:01.000Z",
  "site": "http://www.techmeme.com",
  "tags": [
    "ZDNET",
    "OpenAI launches GPT-5.3-Codex, which it says runs 25% faster, enabling longer-running tasks, and“is our first model that was instrumental in creating itself”"
  ],
  "textContent": "\n\n David Gewirtz / ZDNET:\n**OpenAI launches GPT-5.3-Codex, which it says runs 25% faster, enabling longer-running tasks, and“is our first model that was instrumental in creating itself”**  —  ZDNET's key takeaways  — GPT-5.3-Codex helped debug and deploy parts of itself.  — Codex can be steered mid-task without losing context.",
  "title": "OpenAI launches GPT-5.3-Codex, which it says runs 25% faster, enabling longer-running tasks, and \"is our first model that was instrumental in creating itself\" (David Gewirtz/ZDNET)"
}