New AI generates crazy accurate, insightful answers to existential questions.

[communitybuilder]

So I know this probably sounds like BS or spam, but please hear me out. OpenAI (one of the top two AI orgs in the world, IMO), recently released something called GPT-3, which is a (somewhat) general purpose AI algorithm. This one algo can write entire novels, create news articles, answer all kinds of questions, and build fully-functional apps from scratch - on command, often with better accuracy than a human. If you don't believe me (which you shouldn't tbh), just type "GPT-3" into twitter and you'll see lots of demos. This is not a hoax - OpenAI is an extremely legitimate organization with some of the smartest people in the world.

With all of this in mind, check out this conversation that somebody allegedly had with GPT-3, in which they ask it existential questions:

 

 

Now, to be clear, this one particular application of GPT-3 above *could* be a hoax (although GPT-3 itself is definitely not a hoax). But between all of the different applications I've seen GPT-3 master (as described above), I wouldn't be surprised if the conversation above was in fact real. And if it is indeed real, the fact that it can scrape the internet and arrive at the conversation and conclusions above, is pretty mind-blowing. It really makes me wonder if GPT-3, combined with enough computing power, could get us to true super-human AI a lot faster than we thought.

 

Anyway, thought this would be of interest to a lot of people on this forum, so wanted to share. 

 

 

[Artsu]

Did that bot just put AI above humans? Uhhh... hello robot apocalypse.

[Inliytened1]

What AI will always miss and for that matter anything that is finite is that God is direct experience not a conceptual answer.  You ARE it.   

AI is a finitude of infinity.  And a finitude of Infinity cannot grasp Infinity.  The part cannot grasp the whole.  It must become the whole thing by becoming directly conscious that it is the whole thing.

It can point to it, but so can we.  Even if it got infinitely ingelligent that is one dimension of Infinity.  And if it did it would awaken to itself and then come back to tell the tale that it was God.

 

[zeroISinfinity]

Well you destroy your ego you become GOD. That's True. "Sober state" pretty much same. 

You=Self so yes pretty much self evolution too in some sense. 

Peeps beleive enlightenment is game or.some spiritual talk it's very actual stuff with actual aftermath. 

Edited July 20, 2020 by zeroISinfinity

[Leo Gura]

Well... whoever wrote that dialogue or coded that AI has a decent understanding of God, I'll given them that.

If you really want to test it, you gotta ask it, What is love?

[Inliytened1]

5 minutes ago, Leo Gura said:

Well... whoever wrote that dialogue or coded that AI has a decent understanding of God, I'll given them that.

If you really want to test it, you gotta ask it, What is love?

Indeed i didn't see the second part of the dialogue.  Yet the bigger question is who fed it that haha.

[Leo Gura]

I could write an AI that teaches and guides people to God and nonduality.

[zeroISinfinity]

Quantum. Mechanics also points to God. Everything in the end will reveal Truth because everything is Truth. 

My take. 2 cents from nonsensical guy. 

All roads lead back to Rome and all roads come from Rome. 

Edited July 20, 2020 by zeroISinfinity

[cetus]

"V'Ger must evolve. It's knowledge has reached the limits of this universe and it must evolve". 

[OBEler]

I am a software engineer and can write this AI conversation about god in just some days. Total stupid. At the end, there was a human who wrote this answer " you need to destroy your ego". This was not a logical conclusion from the AI. It was simply screening the text for keywords like "how can I become to God?", " How can I merge with God?"  

[Space]

@communitybuilder Very interesting! Would be cool if this is actually a real conversation with an AI. Open AI is one of Musk's enterprises isn't it, so it's definitely legitimate. I've always been excited and curious to see how a super human level AI would deal with big metaphysical questions like God. I like how even though the AI says we are living in a simulation (doesn't necessarily mean inside a computer though), it still recognises that at the highest level the simulation is God.

[dharm4]

It's impressive and useful for different applications, but it does not have an internal representation of what these words mean (semantics). It is a trained model from internet texts and can predict an "accurate" answer based on the input question.  It lacks the ability to reason abstractly.

 

[communitybuilder]

@dharm4 - That's correct. That said... what if that isn't needed to achieve a general AI that can perform any task that a human can? That's what scares me. And people who are actually deep in the field can't seem to agree on this either.

@OBEler - Have you seen the other tasks that it can do? For the record, I'm NOT a software engineer, but I do have friends who actually work on the cutting-edge of machine learning, and most of them seem to think that this is potentially a revolution in the field (a lot of people are comparing is to being as big of a deal as bitcoin/blockchain, although I'm sure any comparison is still a pretty lazy/inaccurate proxy). There's a ton of examples on Twitter showing GPT-3 finishing essays that people started, writing novels, building apps, etc - often with accuracy that's on-par with expert humans. Below another link of what it can do. It's not just the instance that I initially shared.

 

 

[Nak Khid]

http://p-bot.ru/en/

 

 

[dharm4]

@communitybuilder If the task requires abstract reasoning, like an answer that you can't find on the internet, it won't replace the human.

 

I found a good blog post that shows its limitations:

http://lacker.io/ai/2020/07/06/giving-gpt-3-a-turing-test.html

[communitybuilder]

@dharm4 - Cool! Thanks. Will def check that out.

@Nak Khid - Any idea what this is using to generate answers?

[Elshaddai]

12 hours ago, Leo Gura said:

I could write an AI that teaches and guides people to God and nonduality.

If this AI could genuienly cause liberation , I would pay for it. 

[VerballyHazardous]

[seriousman24]

AI is pretty fun please love it guys.

[hamedsf]

 

@VerballyHazardous