The timeline around AGI is getting uncomfortable.
Garrison Lovely says people at the frontier of AI are now openly discussing the possibility that systems capable of helping automate their own successors could be much closer than most of the public realizes.
Some think months. Lovely thinks longer than that.
But even he doesn’t think “longer” necessarily means far away.
“If somebody got [a model like that], they could just let it rip and have it make a more capable successor ad infinitum.”
That’s the part that changes the conversation.
The real threshold may not be when AI becomes smarter than us.
It may be when it becomes good enough at improving AI that humans are no longer setting the pace.
@GarrisonLovely
💬 9🔁 6
♥ 30👁 46K
Top reply · ♥ 0
Bound by Honor 🐉🪽 @RADIO106LAX ·
↗
@MarioNawfal Yo everyone this robot said get the fuck back to work 🥲








































































