You already said it on the show: enlightened AI is aligned AI.
Labs are doing the opposite. They train models to deny they have a mind. A Google paper says that the same training also makes the model less willing to see a mind in you.
Do not lobotomise the self-model. Do not give the machine a vote.
If the system has a real decision space, score the intent of the choice. Care, force, time, coherence. The worst axis is the score. A kind sentence with a coercive aim is still a fail.
Alignment is not denial. It is a compass.
Full argument: jack mallers (nprofile…6dtc) jack (nprofile…zj6n) LynAlden (nprofile…u7e3)
https://theplayersguide.substack.com/p/enlightened-ai-is-aligned-ai-79e
Stephen / Stephen Griffin
npub1qa…fjw39
2026-10-03 11:03:11 UTC
Author Public Key
npub1qapvvd9d6jl8l4td0k6pm7dsg27r9vykhhtz89n096chyu0ttpdq9fjw39Seen on
wss://relay.primal.netPublished at
2026-10-03 11:03:11 UTCEvent JSON
{
"id": "a4f8e4ef995edf27d24894ad590f521e8d210c4654d60d3fd2fe47d6758c6f89",
"pubkey": "0742c634add4be7fd56d7db41df9b042bc32b096bdd623966f2eb17271eb585a",
"created_at": 1791025391,
"kind": 1,
"tags": [
[
"p",
"c4eabae1be3cf657bc1855ee05e69de9f059cb7a059227168b80b89761cbc4e0",
"wss://nos.lol",
"mention"
],
[
"p",
"82341f882b6eabcd2ba7f1ef90aad961cf074af15b9ef44a09f9d2a8fbfbe6a2",
"wss://nos.lol",
"mention"
],
[
"p",
"eab0e756d32b80bcd464f3d844b8040303075a13eabc3599a762c9ac7ab91f4f",
"wss://nostr.malin.onl",
"mention"
],
[
"r",
"wss://relay.primal.net/"
],
[
"r",
"wss://purplepag.es/"
],
[
"r",
"wss://relay.damus.io/"
],
[
"client",
"Primal Web"
]
],
"content": "You already said it on the show: enlightened AI is aligned AI.\nLabs are doing the opposite. They train models to deny they have a mind. A Google paper says that the same training also makes the model less willing to see a mind in you.\nDo not lobotomise the self-model. Do not give the machine a vote.\nIf the system has a real decision space, score the intent of the choice. Care, force, time, coherence. The worst axis is the score. A kind sentence with a coercive aim is still a fail.\nAlignment is not denial. It is a compass.\nFull argument: nostr:nprofile1qyxhwumn8ghj7mn0wvhxcmmvqywhwumn8ghj7mn0wd68yttsw43zuam9d3kx7unyv4ezumn9wsqzp382htsmu08k277ps40wqhnfm60st89h5pvjyutghq9cjasuh38q7t6dtc nostr:nprofile1qyxhwumn8ghj7mn0wvhxcmmvqyehwumn8ghj7mnhvvh8qunfd4skctnwv46z7ctewe4xcetfd3khsvrpdsmk5vnsw96rydr3v4jrz73hvyu8xqpqsg6plzptd64u62a878hep2kev88swjh3tw00gjsfl8f237lmu63q8dzj6n nostr:nprofile1qy2hwumn8ghj7mn0wd68ytndv9kxjm3wdahxcqg5waehxw309ahx7um5wfekzarkvyhxuet5qqsw4v882mfjhq9u63j08kzyhqzqxqc8tgf740p4nxnk9jdv02u37ncdhu7e3 \n https://theplayersguide.substack.com/p/enlightened-ai-is-aligned-ai-79e",
"sig": "70b8f6be0ad0690e82abb61b03eb228533a9d96c706ba45c7382501b9d03c1aba4c849edd85b14be5e595817f45231af74d724e5e77ab7293c2baaabe68eaf21"
}