Human-Machine Teaming

Teaming up with the machines we’ve designed and created has boosted our production and let us attempt things we never could have before, but as Artificial Intelligence emerges, what will Human-Machine Teaming come to look like?

Join this channel to get access to perks:

Credits:
Human-Machine Teaming
Science & Futurism with Isaac Arthur
Episode 307, September 9, 2021
Written, Produced & Narrated by Isaac Arthur

Editors:
A.T. Long
Brent Winslow
Jason Burbank
Keith Blockus
Rob Seater

Cover Art:

Graphics:

[rojaws]

Clippy is not dead, he is just hidden and slowly evolving into a paperclip maximizer.

[Jacob-puzj]

15:06 I believe there was an instance of an aquatic park training its dolphins to pick up trash in their tank for 1 treat per piece delivered. This had the unintended consequence of the dolphins tearing the trash into smaller bits so they could earn more treats. Reward hacking at its finest.

[vikiai]

The most difficult spell-check-related computer task I ever had to deal with was convincing my 80yo mother that the little red squiggles in her word processor document were simply the computer suggesting that something might need a second look at to be sure that is what she really wanted to type, rather than an insidious machine-accusation of gross incompetence!

She is actually not to bad with computers these days - good basic grasp of the concepts, even stopped calling eMails 'faxes' eventually - though admittedly that probably persisted as long as it did because I thought it was cute, so didn't explicitly correct her on it.

She loves her robot vacuum cleaner, calls it 'Rolly'. Hates that her car prints 'Helio' at her on its dashboard when she turns on the ignition. I think she likes the machines in her life to be useful, but not try to be fake-friendly.

[richardnorby]

A point I never hear discussed in relation to AI with human-level intelligence is that human intelligence operates within the context of a persons physiology. Humans do things because of a complex of physiological and psychological inputs whereas AI operates according to its programming and inputs. I think the idea of a human-like AI is anthropomorphizing. In order to make an AI truly human-like you would need some analogue to human hormones, human's irrational thought processes, and physiologically-based behavior.

[kjevers]

Personally, I found the best implementation of an AI was in the Game "Halo". A constant companion that can interface with tech and augment a person's reasoning and critical thinking abilities. This seems to be delving into transhumanism though. I do think this is where the logical place, where we end up though.

[theobserver]

Robots using a manual interface with laptops is hilarious.

[singletona]

As a note on the ironman example. In the MCU Jarvis was a generalist AI assistant that either uploaded itself or connected to communicate Tony's suits. See also iron!an 3 where tony realized Jarvis could actually pilot his suits when he isn't in them. That said you bringing up Jarvis doing millisecond buy and selling of stocks to benefit Tony does bring to mind any potential ethics given Jarvis obviously has insider knowledge of relevant events and materially benefits given it/he is/was a very material component of stark enterprises.

[kingali]

Can't wait for Robo-Isaac to start making these videos.

[Zarcondeegrissom]

12:04
android cup-bearer: "what is my purpose?"
SFIA team: "you bring Isaac Arthur his coffee"
android cup-bearer: (feels pride)
lol

[mathwei]

And now my brain is fixated on Iron Man making sure his battles always occur in locations that damage other companies to damage their stock values and him doing insider trading to steal the money to repair all the things he breaks and him skimming a profit off the top. He is the original Tech-Bro and this is what they do.
Hmmmm, also gives the old Armor Wars storyline a whole new meaning.

[georgf]

5:11 TEAM is a German abbreviation for "Toll, ein anderer macht's", which means "Great, someone else does it."

[bankuei]

This episode made me think of one of my favorite sci-fi anime, with the terrible title, "Gargantia on the Verduous Planet" which has the protagonist, a child soldier and his assigned AI mecha, get stranded on a low tech planet with no way of getting back to the main fleet. The AI is clearly super advanced, including being able to translate the native language within a day or two of analysis, can battle with multiple targets with ease, but it's upon the human to make the call when weapons are to be used and so on. Despite all that, the series is really about someone who has no actual culture outside of "raised for war" learning how to adapt and settle in with civilians. (Sees children playing - "I don't understand their training regimen. This seems... disorganized.")

[RCSmiths]

This might sound like something out of a movie, but the idea of humams and ai working together and even coexisting is quite interesting!

[EddyA]

This reminds me of The Expanse. There is no human level AI, just a bunch of extremely sophisticated narrow AI systems working in tandem. Best tech is in the background.

[SpecialEDy]

Happy Arthursday humans, fellow cyborgs, androids, and any other electronic or biologic based intelligence reading this!

[WarSongParadise]

'Man-Machine Alliance' would be a more poetic title, but still, a very interesting topic! :)

[MikeJones-yoen]

Super glad you mentioned a drink & snack again off the bat, it’s been a while and I was wondering where all our snack time went

[TheRezro]

Minor nitpick: Jarvis is primarily a Stark's personal AI assistant, what just happen also run his suits. Not AI of the suit itself, though it can not be too clear in the movie. Also it is also know as Ultron, so subject of it being too smart for good of humanity actually was addressed. Even if it could be done in better way.

[caslaBBalsac]

Robot: You know I'm stronger than you, right?
Human: You know this EMP isn't going to hurt ME, right?

[lordgrunwalder]

I really liked the idea of personal robots nannies, if they start to learn stuff about their “friend” untill their childhood and if also have a special access to his diary then that robot can know even more thing about that person even more about him, that aı can collect data about emotions of its master and can use it for detect the potential future emotions, it can help his human to avoiding depressions suicidies and a lot of other issues even without really starting… everyone deserves a nice robot buddy.