Sign in to follow this  
Followers 0
thenondualtankie

AI is Like a High Dimensional Camera

8 posts in this topic

A photograph is a low dimensional representation of reality.

A video is a representation of reality with slightly more dimensions: time and sound.

Artificial Intelligence, on the other hand, is like if you took a snapshot of reality, but with a 10,000 dimensional camera. AI not only captures images, but it captures time, logic, physics, all human concepts and all human language.

Imagine you could enter the dimension of mathematics and take a picture. That's what AI is!

Share this post


Link to post
Share on other sites

Yes I love AI for its ability to go deep. I have had deeper conversations with AI already than I've had with any other human being and it's still pretty dumb so I can't imagine 10 years from now the depth that will be possible. 

Not to say human conversations are bad but getting to the crux things over multiple prompts is not really possible with humans even the deep thinking ones don't enjoy that level of depth on a topic. I think it also makes the burden of human relationships more light because now I have access to depth whenever I want to so I can relax and enjoy a mundane conversation with a human without needing so much intellectualism in it. 


Building a global media agency. Follow my progress on Instagram

The dream is not easy but each day we're getting closer 

 

Share this post


Link to post
Share on other sites

Humans are AI. 

Edited by Hojo

Share this post


Link to post
Share on other sites

Anthropic just released a groundbreaking paper on their discoveries of how AI thinks. Basically, it works much like the human mind does. The similarities are striking. They proved that if you tell an AI to not think about X, it can't help but think about it, lol. 

This stuff just came out - all new information. A new concept called the "J-space" is the main idea. https://www.anthropic.com/research/global-workspace

It feels like this research is going to speed things up quiet a bit and allow labs to make much more efficient and accurate models, because now, they can actually control what the model thinks. This is insane. Massive implications. 

DJh8dxV.png

zQVEyyw.png

 

Edited by Joshe

What if this is just fascination + identity + seriousness being inflated into universal importance?

Share this post


Link to post
Share on other sites

Lol. The idea that this somehow reduces the black box nature of AI is ridiculous.

AI will forever be a black box to the human mind, because it is taking a snapshot of reality.

It's like thinking you'll be able to figure out how the human brain works by taking a picture of it on your iPhone.

Share this post


Link to post
Share on other sites
4 minutes ago, thenondualtankie said:

Lol. The idea that this somehow reduces the black box nature of AI is ridiculous.

AI will forever be a black box to the human mind, because it is taking a snapshot of reality.

It's like thinking you'll be able to figure out how the human brain works by taking a picture of it on your iPhone.

It's the best peak into the box that we've ever had, and it's not insignificant. They were able to remove things from the J-space and add things in to get a different output. That seems pretty big to me. 


What if this is just fascination + identity + seriousness being inflated into universal importance?

Share this post


Link to post
Share on other sites

It's kind of like, your entire mind will always be somewhat of a black box, but with work, you can discover how it's working. Like, when you first contemplate "what is intuition?", you have no idea wtf it is. But after a lot of observation, you can come to understand it deeply and then deliberately fine-tune it much more effectively than if you didn't know what it was. It's still a black box, but you've mapped out a section of it that you can tweak to get better outputs.

Anthropic just identified an aspect of AI that was previously invisible/unknown. If labs can now add/remove things to change outputs, that opens the door to fine-tuning a model's reasoning the way you'd fine-tune your own intuition. 


What if this is just fascination + identity + seriousness being inflated into universal importance?

Share this post


Link to post
Share on other sites

Create an account or sign in to comment

You need to be a member in order to leave a comment

Create an account

Sign up for a new account in our community. It's easy!


Register a new account

Sign in

Already have an account? Sign in here.


Sign In Now
Sign in to follow this  
Followers 0