Event JSON
{
"id": "79e8c5ceae49a70ea663bde6941fbf036661f992f345718220b83fc3e3f9890b",
"pubkey": "841b017d49f8211881a23acbd002bfb07c0d9c0d720c4a99c42669458bf0ee39",
"created_at": 1770538282,
"kind": 1,
"tags": [
[
"t",
"AI"
],
[
"t",
"ai"
],
[
"t",
"Agents"
],
[
"t",
"agents"
],
[
"t",
"Economics"
],
[
"t",
"economics"
],
[
"t",
"BOTEC"
],
[
"t",
"botec"
],
[
"t",
"Models"
],
[
"t",
"models"
],
[
"t",
"Investment"
],
[
"t",
"investment"
],
[
"t",
"Reliability"
],
[
"t",
"reliability"
],
[
"client",
"Nostur",
"31990:9be0be0fc079548233231614e4e1efc9f28b0db398011efeecf05fe570e5dd33:1685868693432"
]
],
"content": "Economic BOTEC (back-of-the-envelope) analysis exploring the feasibility of AI agents. Building on Toby Ord's \"half-life\" model, it shows how the cost per success grows exponentially with task length, creating an economic wall insurmountable without a leap in agent reliability (continual learning).\n\nhttps://www.lesswrong.com/posts/NqNmNqtQCaqtpAeyX/agent-economics-a-botec-on-feasibility\n\n#AI #Agents #Economics #BOTEC #Models #Investment #Reliability ",
"sig": "e70e444c1a45f0c62cbbd102a819db29041a43367e034e9af859b9d9479af715eff1b578aea9f8362d8ef4e67554768418856f6932f17d09d89df12de6d08b75"
}