Post Ai0pam3Jffdj3ebdVg by Grimblob@mastodon.social
 (DIR) More posts by Grimblob@mastodon.social
 (DIR) Post #AhatosJwVjtyG8P7mC by nixCraft@mastodon.social
       2024-05-05T21:51:20Z
       
       1 likes, 2 repeats
       
       the guy is losing it slowly and wants to make a remake of 1984
       
 (DIR) Post #AhovaGShONEKuS6F4y by david@theblower.au
       2024-05-05T23:39:46Z
       
       0 likes, 0 repeats
       
       @nixCraft maybe... and hear me out here, maybe he should go outside and get some actual friends. You know, like talk to people, except not about AI or funding.
       
 (DIR) Post #AhvUexFEvleC6JXH9c by dozymoe@mastodon.social
       2024-05-06T02:38:40Z
       
       0 likes, 0 repeats
       
       @nixCraft targetted ads
       
 (DIR) Post #Ai0pam3Jffdj3ebdVg by Grimblob@mastodon.social
       2024-05-06T01:06:40Z
       
       0 likes, 0 repeats
       
       @nixCraft I think Sam should get his wish. A lot of AI is just "people in poor countries in cubicle farms" so he should release all the details of his life and the wisdom of crowds can be competent for him.
       
 (DIR) Post #Ai31wULnkpIgIGIDo0 by mozz@mbin.grits.dev
       2024-05-06T01:39:29+00:00
       
       0 likes, 0 repeats
       
       @nixCraftThe crazy thing about uncontrolled AGI is that at a certain point, when it exceeds human reasoning, whatever trajectory is dialed in at that point is just what's going to happen. There is no rewind back to before it got activated and the humans were in charge. It's a point you only get to reach once, and then once it's been reached, you better hope all the safety and alignment is set up right, because from that point forward, it's going to continue on the trajectory that was set, and it'll just disable any human attempts to control it, if it doesn't want to be adjusted in whatever way.We're still a ways away from that, obviously. But the research on AI safety has to keep pace with the pace of increasing capability. Again, we only get one shot at it.The top AI researchers used to seem like they understood that (better than the general perception that existed of what was dangerous about it). OpenAI actually attracted a lot of the top people with the promise of making a top priority out of safety. I don't really know, but it seems like now I don't hear very much about it anymore.
       
 (DIR) Post #Ai3vXK82DXSumfQlto by w4tsn@darmstadt.social
       2024-05-06T06:37:15Z
       
       0 likes, 0 repeats
       
       @nixCraft BREAKING: tech bro invents friends and colleagues
       
 (DIR) Post #Ai6Deg7P2rEkVCiBv6 by dpino@mastodon.social
       2024-05-06T10:55:18Z
       
       0 likes, 0 repeats
       
       @nixCraft by definition, a colleague that knows absolutely everything cannot be a super colleague, it can only be an asshole