AGI is here

[Shodburrito]

OpenAI announced their new frontier model o3 today.

It's either at the level, or better than humans in a vast array of intelligence benchmarks. Stop pushing the goalposts and recognize that this is general intelligence. Witnessing history. (Starts at 23:00)
 

 

Edited Friday at 06:24 PM by Shodburrito

[OBEler]

Yep. It is happening.And we have no access right now.

How will this change the world from now on. This should be in headlines on every newspaper.

https://arcprize.org/blog/oai-o3-pub-breakthrough

Edited yesterday at 12:10 AM by OBEler

[Shodburrito]

@OBEler After some more digging, on the ARC AGI benchmark, they say it fails the 2.0 version of the test with only 30% correct. It's interesting though, because at some point, we will no longer be able to create logic tests where it fails them. It is advancing at such an inredible rate I think we are in the singularity right now and we don't realize it. Next year this model will come out, and by 2026 a  new frontier model will beat o3. I think the next model will truly be AGI. Already o3 is at the level or beating top mathematicians and scientists in their fields.

We are only 2 years since the release of chatgpt 3. Incredible how far in 2 years we've come. A year ago video generation was barely viable: the models could barely produce videos that adhered to the prompts. Now with google's VEO 2, it is generating near-realistic videos, with some fooling me. And to think this is the worst these models will ever be. Exponential growth does not seem to be halted, if anything it just seems to be heating up. I think by the end of the decade we will have millions of fully embodied AGI robots across the globe. 

We are living in the most interesting point of human history. Buckle up.

Edited yesterday at 12:53 AM by Shodburrito

[OBEler]

Yes but now is the question, how do we normal folks adapt to this phase.

 

 

[RightHand]

5 hours ago, OBEler said:

Yes but now is the question, how do we normal folks adapt to this phase.

 

 

If it's AGI, its first order of business will be a manifesto on human debauchery.

[Hojo]

Since Donald Hoffman says that learning patterns is a survival technique of evolution, this means if the robots get better than us AI will develop newer survival techniques that we do not have. It will see less of the big picture but will find new ways to survive. From a human perspective because its programmed by humans. This is going to start a strange loop.

This will show people reality is a simulation. And we are witnessing the simulation on the computer in the same simulation we are in IRL.

Edited 17 hours ago by Hojo