at://bnewbold.net/app.bsky.feed.post/3jxzob6vpy62e

Back to Collection

Record JSON

{
  "$type": "app.bsky.feed.post",
  "createdAt": "2023-06-13T06:36:31.173Z",
  "reply": {
    "parent": {
      "cid": "bafyreia3w5hmbxnqodkcqzk3vhdv6a577gysxpbgph3jyzimpvbqmrk5c4",
      "uri": "at://did:plc:gtjszliizbclosvvbluq53qh/app.bsky.feed.post/3jxzk4xr3lb25"
    },
    "root": {
      "cid": "bafyreigqb6ahnbonnjzsdbmh7ho5lpmdh2guv3fojaozip5onzpt64nafm",
      "uri": "at://did:plc:gtjszliizbclosvvbluq53qh/app.bsky.feed.post/3jxzj2l2hnj2e"
    }
  },
  "text": "I don't really get the access-to-content narrative for most AI/ML stuff, unless recency bias is more powerful than I expect. like, the old dumps are no-cost and easily available in bulk form as torrents.\n\nI guess the idea is to source info in inform queries, more than LLM base training?"
}