Event JSON
{
"id": "94a2b8d5d2bb35f9be24ad45d0eb4bd7961107dbb4c003232d1ab4cb729d8d63",
"pubkey": "1cabde0bc028762ae84ad6f0a3be93e4233a59263a3290f63d9975fb4d55e598",
"created_at": 1706763816,
"kind": 1,
"tags": [
[
"p",
"bb0526093bc7f94fd056d51262669842fd978190d0f13d3c5e530b7278bbd844",
"wss://relay.mostr.pub"
],
[
"p",
"ae947543152d3049444c930c2d35a0edc7d19a4e517cee01fd1ba03874437724",
"wss://relay.mostr.pub"
],
[
"e",
"e829b66878924349086056588a0bee3e33d20f4863d85282e80fdb82b1714d2a",
"wss://relay.mostr.pub",
"reply"
],
[
"proxy",
"https://discuss.systems/users/ricci/statuses/111854473450023727",
"activitypub"
]
],
"content": "nostr:npub1hvzjvzfmclu5l5zk65fxye5cgt7e0qvs6rcn60z72v9hy79mmpzq4qmle9 I've got myself tied into knots thinking about this - that typo is pretty dumb, clearly this is text a human wrote, since LLMs aren't known for misspellings... But wait! What if sprinkling in some typos is a way to make the text look like a human wrote it! What if one of the things this GPT is trained to do is misspell things? What if the typo in that description is actually *genius*‽",
"sig": "634f51111291897f842c3150e81dd7bcc7449636053dac0a3d68c9cbd69033cba9db2589b9ae386519e5e0b7a0583ac03e43b5c894cf5be3001610123e4d6a9a"
}
