Zeitpunkt Nutzer Delta Tröts TNR Titel Version maxTL Mo 27.05.2024 00:00:20 11.941 0 976.622 81,8 Mastodon.green 4.2.7 500 So 26.05.2024 00:00:23 11.941 +1 975.746 81,7 Mastodon.green 4.2.7 500 Sa 25.05.2024 00:00:12 11.940 0 974.830 81,6 Mastodon.green 4.2.7 500 Fr 24.05.2024 00:00:29 11.940 -3 974.003 81,6 Mastodon.green 4.2.7 500 Do 23.05.2024 00:00:31 11.943 +1 973.065 81,5 Mastodon.green 4.2.7 500 Mi 22.05.2024 00:00:11 11.942 -1 972.010 81,4 Mastodon.green 4.2.7 500 Di 21.05.2024 00:00:14 11.943 0 971.249 81,3 Mastodon.green 4.2.7 500 Mo 20.05.2024 00:02:58 11.943 0 970.403 81,3 Mastodon.green 4.2.7 500 So 19.05.2024 00:00:11 11.943 +1 969.554 81,2 Mastodon.green 4.2.7 500 Sa 18.05.2024 00:00:19 11.942 0 968.565 81,1 Mastodon.green 4.2.7 500
Gerry McGovern (@gerrymcgovern) · 10/2022 · Tröts: 7.535 · Folger: 3.428
Mo 27.05.2024 17:51
“First, Meta successfully trained its AI to excel in the pursuit of political power, albeit in a game. Second, Meta tried, but failed, to train that AI to be honest. And third, it was up to us outside independent scientists to, long after the fact, disprove Meta’s falsehood that its power-seeking AI was supposedly honest. The combination of these three facts is, in my opinion, sufficient cause for concern.”
https://www.msn.com/en-us/news/technology/why-it-is-so-dangerous-for-ai-to-learn-how-to-lie-it-will-deceive-us-like-the-rich/ar-BB1mnGOU
Meta tried to make its AI honest? You're not serious, are you?
[Öffentlich] Antw.: 0 Wtrl.: 1 Fav.: 2 · via Web