CLIP | Lawrence Krauss: Artificial Intelligent Physicists Are Coming

Will anything ever come of #artificialintelligence besides the end of human dominance in chess, Go, and that annoying Microsoft Paper Clip assistant, 'Clippy'? Explore the possibilities for “AI Feynman” with a cameo by my friend Max Tegmark. See Lawrence’s reaction to “Galil-AI-O”, my proposal for a project involving Galileo Galilei's collected works and GPT-3.

Watch our full conversation here: Lawrence Krauss: The Physics of Everything

Join this channel to get access to perks:

#lawrencekrauss #originsproject

Watch my most popular videos:📺

Be my friend:

-~-~~-~~~-~~-~-
Please watch: "Neil DeGrasse Tyson: Plays the Race Card!"
-~-~~-~~~-~~-~-

[janwar]

It takes a big man to post argument against oneself. Good on you Dr. Keating.

[cybrfriends]

these summaries/abridged videos are sometimes easier to absorb than the ones that are over an hour. i really enjoyed this one.

[SteveC]

I enjoyed this, I haven't yet had a chance to watch the full interview so this was nice.

[gregoryhead]

AI can't tell us what literal constant(s) to multiply the electron mass is g. The two literals are known, yet unknown.

[vesawuoristo]

The recent news about the IBM chat bot being maybe sentient certainly was very interesting.

[resonant_theories]

the worst thing with AI is when it will be "wrong" and when everyone will have learn to trust it blindly..

[wulphstein]

You literally don't need AI physicists. I can already tell you right here and now how to unify quantum mechanism and relativity in a way that leads to an experiment and the warp field. It's all right there is a Modern Physics book.

[picksalot]

Regarding AI generated Art, I suggest everyone take a look at this YouTube video. The results are startling: "End of An Artist's Career? A.I. Creates Art for You! - Complete Guide"

[peterz]

Why should we assume that we can bridge the gap from automatic operations to consciousness by heaping on more and faster operations. And the claim of substrate independence also seems to be a huge leap. Many natural phenomena seem to be substrate dependedent. Fusion takes place under certain conditions involving certain elements with a certain configration. Only a simulation of certain aspects of fusion can take place within a computer. Why shouldn't consciousness also be dependent on the physical materials as well as the configuration of those materials? So much we don''t know that it appears ludicrous to make pronouncements about silicon based AI system leaping to consciousness.

[cammm]

Would a human brain without any, zero sensory inputs be just like AI? Then flip that around and we will "see" if robots can experience elation and fright at the same time.

[theklaus]

A very interesting subject. It has a kind of dooms day attached to it. ( Sam Harris for example) if I understand it correct it is just a matter of time before we have it. Nano technology supporting our immune system sounds so good. Self replicated robots repairing whatever damage in our body. But a risk for developing an intelligence we can't control is a problem ! Harris often talks about if it first has access to the internet then it's a highway for knowledge. So as good as it sounds we know the biological evolution have no chance keeping up with the technical evolution. The question seems to be are we able to control it ! None the less it's very fascinating. I will find the complete podcast and gather information here ( and there )👍. So lucky all this knowledge just a click away.!

[blengi]

Given GPT3 can't even do simple arithmetic generically and robustly - can any AI model? - and hence shows essentially no meta understanding of even the most basic mathematical principles/logic after seeing perhaps billions of examples, gotta wonder what the software nerds have been doing for the past 10 years...

[neIntangible]

If you tell a created intelligence a joke... would it "get it"?... could it really understand the nuance implied?

[oneshot]

Intelligence requires "self consciousness".
This AI is not even sentient.
This whole AI thing is a one big joke.

[wulphstein]

While the physics community is frivolously toying with irrelevant ideas, the need to develop warp drive physics becomes greater. Only a warp drive capable drone can outmaneuver and destroy an ICBM. There are madmen in dictatorships around the world who have nuclear weapons. We don't have a reliable enough nuclear defense. The prerequisite to warp drive physics is quantum gravity. If the physics community doesn't focus on quantum gravity, there might come a day when a madman will push the button, and it will be too late for human civilization.

Rethink your philosophies about what kinds of physics you are willing to think about.

[theduder]

"AI physicists" Excellent joke. Excellent joke indeed. Should probably copy-write that before everyone begins using it. lol
Just goofing around.
However, without human input/interaction, AI is still currently unable to hold up its side of a conversation very effectively.
It will have the math down no problem, but I have to remain doubtful of "AI physicists" based on what our current "AI" is still incapable of. AI still can not determine which is the better object to impact with a vehicle, a child or a cardboard box.
I do not foresee AI being able to process the complexities behind scientific discovery. Not any time soon.
The weird part... I want to be both correct and wrong all at the same time. lol

[poksnee]

Computers are tools nothing more. They do what they are told.

[resonant_theories]

has every human played chess with a machine in the first place?

[AlistairGale]

There will be fascists featuring Ben Shapiro.