Mark Downie on Nostr: debug-gym ... "They found that the best models could indeed improve their results by ...
Published at
2025-04-01 01:42:35Event JSON
{
"id": "a914021070d0283cee203a58c5073b0736397865e6e760ae841da04034c1d9f0",
"pubkey": "01c72ee2f371d0507d54883d77605111391a4dc38b4f0e5558a9ccbe399ae227",
"created_at": 1743471755,
"kind": 1,
"tags": [
[
"proxy",
"https://dotnet.social/users/poppastring/statuses/114260164991339763",
"activitypub"
]
],
"content": "debug-gym ... \"They found that the best models could indeed improve their results by running pdb as a tool.\"\n\nhttps://simonwillison.net/2025/Mar/31/debug-gym/#atom-everything",
"sig": "b741a69d4a2f27a2cec0bdcd15f39534c481aae407bad4c2d226329df744ab459764b775e4ed683e755970e5cbc302ac6e6c301c1adb21acf9b50cfca03d0800"
}
