There isn't an AI app or model for ~some of the work I do.

npub12xamzh3hapdspk98dqmmdd4rx4w3m30vtz7apuuq3jdd8dm6y2gq9tstqx
hex
8c55be5296d7e9d4678f4d2b9c2b5fe8a08ccb3e723438e09643321f414e0f2bnevent
nevent1qqsgc4d722td06w5v78562uu9d073gyvevl8ydpcuztyxvslg98q72cprpmhxue69uhhyetvv9ujuem4d36kwatvw5hx6mm9qgs9rwa3tcm7skcqmznksdakk63n2hgachk930ws7wqgexknkaaz9yqpye8usKind-1 (TextNote)
↳ Reply to Event not found
42009c8b69edb5e73c6557c9c1ff1f86ad9d1f336f7c84a445b2304c1db44e7d...
There isn't an AI app or model for ~some of the work I do.
I'm a certified welding inspector. I often have to review and sometimes edit documents such as welding procedures and welding procedure qualification records.
The information available for inferencing is outdated. And hallucinations happen too frequently for me to bet my house on.
It could be simple. I don't even think we need to get AI involved. Simple python scripts could probably parse through the variables. But it would have to be on a (welding )code by (welding)code basis.
And there's also the critical factor of things that work on paper that don't work in the real world. iE; if you can't perform the weld In the real world you have no business writing the procedure. That's the rule I go by in quality management for humans and I would maintain the same for machines.
Anyways, I keep testing and hacking for my use case. When the tech is there I may change my mind.
Raw JSON
{
"kind": 1,
"id": "8c55be5296d7e9d4678f4d2b9c2b5fe8a08ccb3e723438e09643321f414e0f2b",
"pubkey": "51bbb15e37e85b00d8a76837b6b6a3355d1dc5ec58bdd0f3808c9ad3b77a2290",
"created_at": 1771806624,
"tags": [
[
"alt",
"A short note: There isn't an AI app or model for ~some of the wo..."
],
[
"e",
"42009c8b69edb5e73c6557c9c1ff1f86ad9d1f336f7c84a445b2304c1db44e7d",
"wss://nostr.mom/",
"root",
"1bc70a0148b3f316da33fe3c89f23e3e71ac4ff998027ec712b905cd24f6a411"
],
[
"p",
"1bc70a0148b3f316da33fe3c89f23e3e71ac4ff998027ec712b905cd24f6a411",
"wss://relay.damus.io/"
]
],
"content": "There isn't an AI app or model for ~some of the work I do.\n\nI'm a certified welding inspector. I often have to review and sometimes edit documents such as welding procedures and welding procedure qualification records. \n\nThe information available for inferencing is outdated. And hallucinations happen too frequently for me to bet my house on. \n\nIt could be simple. I don't even think we need to get AI involved. Simple python scripts could probably parse through the variables. But it would have to be on a (welding )code by (welding)code basis. \n\nAnd there's also the critical factor of things that work on paper that don't work in the real world. iE; if you can't perform the weld In the real world you have no business writing the procedure. That's the rule I go by in quality management for humans and I would maintain the same for machines.\n\nAnyways, I keep testing and hacking for my use case. When the tech is there I may change my mind.",
"sig": "18145de9ef21a6299390c1fd7bc136da074279c884d9efbf9c8f584da191096f2ab521bc86a8380f533f285f4f0f663726889f780f0ca86332c8377672e9dbed"
}