Future Timeline

Yuli Ban wrote: ↑Wed Apr 20, 2022 4:33 am

I already posted this. Regardless, I like his presentation style. He's able to properly articulate base principles, and he's not a quack. Unfortunately, those are two things that I cannot say for most singularitarians.

Ray Kurzweils book has been delayed again.

I am 90% sure Ray is no longer writing a book and book websites automatically push back the date as he hasn't notified them it is never coming out.



I really respect that Ray has probably unironically sped up the singularity and human level AGI some years by promoting the theory of said event.

Sad, because I would gladly read the whole book, even if I don't fully believe it.

Ozzie guy wrote: ↑Mon Apr 25, 2022 8:39 am Ray Kurzweils book has been delayed again.

I am 90% sure Ray is no longer writing a book and book websites automatically push back the date as he hasn't notified them it is never coming out.



I really respect that Ray has probably unironically sped up the singularity and human level AGI some years by promoting the theory of said event.

Shame. I remember reading The Singularity is Near about...7/8 years ago I think. Even then I thought his estimates for how soon in the future some of those technologies would be were a little (a lot) on the unrealistic optimistic side, but it really got my mind racing with the possibilities of what I might live to see this century.

Still hope his sequel comes to fruition some day. See if he thinks if we're still on track.

People are saying Multi-Game Decision Transformers is bigger than Gato perhaps people don't care as much as we just had the Gato breakthrough.

An AI in the same category of Gato was publicly announced called "Unified-IO"
https://unified-io.allenai.org/

I am not saying they are of the same quality one is probably better than the other however I am in no position of authority to argue one is better. The important thing is they are in the same category they are both AXI/less narrow AI/pre proto AGI what ever you want to call them. To me this signifies we are in a new era. The norm is now for companies to release these AI's in the grey zone between narrow and proto AGI, this is coinciding with state of the art Ai now being able to pass the Turing Test some of the time with average Joe's who don't care about AI and are not trying to detect it.

Thankfully and amazingly this era should be short lived as we will soon leap into the Proto AGI era once Gato is scaled.

Jesus christ, recently I saw a video from the youtube channel "Edan Meyer" called "AGI is NOT coming soon" where he states learning maths from language models is one of the biggest problems, I also read this link from Yuli Ban viewtopic.php?f=16&t=48&p=22234#p22234 which I believe said the same thing in regards to maths being a problem.


Math's is no longer a problem





(Note I did not link the video from Edan Meyer as the video poping up in the middle of the post would distract people from the news and make them think I was just linking a video)