frederik_diamond on Nostr: It’s funny how, after a few prompts, LLMs seem to get worse and worse at coding. ...
It’s funny how, after a few prompts, LLMs seem to get worse and worse at coding. They stop reading what I actually want and instead start changing the code however they like. That’s been my experience using #ChatGPT and #Claude.
Published at
2024-09-23 10:34:40Event JSON
{
"id": "f771dd1ef6887eef537cf2c8c1972f93ecb73fb6a8dc7b06cb75c0e45b28ded6",
"pubkey": "9c9f81ed795f0f5efa558932824687d84fc7e6a4cfa6db5d6d3b50fcb7ffaec2",
"created_at": 1727087680,
"kind": 1,
"tags": [
[
"t",
"ChatGPT"
],
[
"t",
"Claude"
]
],
"content": "It’s funny how, after a few prompts, LLMs seem to get worse and worse at coding. They stop reading what I actually want and instead start changing the code however they like. That’s been my experience using #ChatGPT and #Claude.",
"sig": "c7ccd15ffcaad6ba03e272bc210a403a40814067bcb4ecf27ed5624817f0243bf118cb507623b5d851a26d3cd866b9ee68d152fbfd1881ef39a2ce967e98c12f"
}