Antti Peltola ⁂ on Nostr: LLMs have become technology that has failed to deliver on its promises the most since ...
LLMs have become technology that has failed to deliver on its promises the most since Pentium 4.
Failure of Pentium 4 however didn't put a loaded gun pointing at the head of the working class as failed CPU microarchitecture is incapable of destroying the economy as it didn't have USD 2 trillion of CapEx depending on its success.
Same 10 % hallucination rate despite how much we throw data and compute on these transformer models. Holy fuck we're in trouble.
#AIBubble
Published at
2026-08-14 06:03:08 UTCEvent JSON
{
"id": "8fa076eb35fdbb6a5d07d87b7469ffac186e50fde08da1fb0213ba0183f480b3",
"pubkey": "776838356f6ed3bf1ab5a5c17a00d4211e4d18d6b548f42cf362f668f2f88d68",
"created_at": 1786687388,
"kind": 1,
"tags": [
[
"proxy",
"https://mastodon.world/@anttipeltola/117092344692403567",
"web"
],
[
"t",
"aibubble"
],
[
"proxy",
"https://mastodon.world/users/anttipeltola/statuses/117092344692403567",
"activitypub"
],
[
"L",
"pink.momostr"
],
[
"l",
"pink.momostr.activitypub:https://mastodon.world/users/anttipeltola/statuses/117092344692403567",
"pink.momostr"
],
[
"-"
]
],
"content": "LLMs have become technology that has failed to deliver on its promises the most since Pentium 4.\n\nFailure of Pentium 4 however didn't put a loaded gun pointing at the head of the working class as failed CPU microarchitecture is incapable of destroying the economy as it didn't have USD 2 trillion of CapEx depending on its success.\n\nSame 10 % hallucination rate despite how much we throw data and compute on these transformer models. Holy fuck we're in trouble.\n\n#AIBubble",
"sig": "d78e5e009f316b99a6e87a3b4094d526881e600ab4671738c434ca2316f48d90a7917ff75c3a660cd94c3e2e4fc2cbc25049884524b999b6e61e45c0dbeae838"
}
