mike on Nostr: ChatGPT has moved memory retrieval away from static memory files to wider context ...
ChatGPT has moved memory retrieval away from static memory files to wider context retrieval. Because of this, I have just initiated a Data export, which can take up to 7 days to compile. You can do the same:
1. Open ChatGPT → Settings.
2. Select Data controls.
3. Find Export data and click Export.
4. Select Confirm export.
5. Open the email from OpenAI and download your ZIP archive.
What you'll receive:
The archive includes your conversation history and other account data. Exports can contain a conversations.json file, which is particularly useful because it allows programmatic processing rather than manually copying conversations.
I shall use my conversations as training data for building my own model, "Brian", my brain.
Published at
2026-10-20 12:30:32 UTCEvent JSON
{
"id": "71ad7fd1af2d73a8372fb5a6a7f2faa11b6ac72855d1a33ab1770160749788ae",
"pubkey": "e83b66a8ed2d37c07d1abea6e1b000a15549c69508fa4c5875556d52b0526c2b",
"created_at": 1792499432,
"kind": 1,
"tags": [],
"content": "ChatGPT has moved memory retrieval away from static memory files to wider context retrieval. Because of this, I have just initiated a Data export, which can take up to 7 days to compile. You can do the same:\n\n1. Open ChatGPT → Settings.\n2. Select Data controls.\n3. Find Export data and click Export.\n4. Select Confirm export.\n5. Open the email from OpenAI and download your ZIP archive.\n\n\nWhat you'll receive:\nThe archive includes your conversation history and other account data. Exports can contain a conversations.json file, which is particularly useful because it allows programmatic processing rather than manually copying conversations.\n\nI shall use my conversations as training data for building my own model, \"Brian\", my brain.",
"sig": "d3703bb43513a90a8bdd18eb94adafd53ccb4251dad74efcf6e14dafdf7f54f44206e00fa304809a530b044e2b92ef7f73995c11e56590f790eb4fcb0ec2fff2"
}