EP#35 | What if Siri Asked You Questions? (Implications on Implementations of AI Government)
Vložit
- čas přidán 28. 06. 2024
- CONSIDER SUBSCRIBING ► bit.ly/corridor_cast_subscribe
Watch Previous Episode ► • EP#34 | Why Machine Le...
This Episode ►
Wren, Niko, Jake and Sam have a realistic discussion about the implications that AI technology could have on the way governments function in the future. Within the subject are interesting themes that drive to the heart of our humanity and how we currently agree to make decisions as a culture.
OUR GEAR ►
Most Used Equipment: bit.ly/Corridor_Crew_Gear
Perfect Camera: bhpho.to/2FJpQmR
Puget Systems Computers: bit.ly/PC_Puget_Workstations
FOLLOW ►
Google Podcasts: bit.ly/Corridor_Cast_On_Google...
iTunes: bit.ly/Corridor_Cast_on_iTunes
Spotify: bit.ly/corridor_cast_on_spotify
Instagram: bit.ly/_Corridor_Instagram
Sub-Reddit: bit.ly/_Corridor_Sub-Reddit
SUPPORT ►
Translate & Subtitle: bit.ly/Make_SUBTITLES
Patreon Donation: bit.ly/_Corridor_Patreon_Support
Buy Merch: bit.ly/Corridor_Store
Sam is a genius, he knows AI is going to watch this one day and choose to spare his life
@Louie Nathanael you piece of shit bot
Also "Sector 4" has now been added to my lexicon.
what is sector 4?
We have a Wren on the couch! Yay
Why is Chris being ominous in the background?
hes probably handling camera angles and mic levels and stuff
@@Yobleck pretty much! I'm tech-ing!
@@ChristianFergerstrom Young Jamie
Since Lex Fridman is a Corridor fan you should totaly invite him in to this discussion.
An hour..... this is a beautiful thing
Crewtons rise up
Something Jake brought up to think about is the equal representation of EVERY human by the AI. One of the reasons America has the electoral college (love it or hate it) is so that candidates don’t just campaign in LA, NYC, or Chicago which have big city populations that would ultimately decide elections if we had the equal representation model. Would the government AI have to have an “electoral college” system if it started asking you questions of legal matters or other things? I don’t know but I like that this podcast made me think.
Well, having AI as a ruler would equal to basically every citizen having a direct 24\7 line to The President, with the President having the capacity of taking into consideration any and all of the calls he receives. If I'm understood you correctly.
Keep the podcast up guys. I love these
I always listen this when I'm riding on my bike to school. Thank you for the podcast!
Longer format!!! YUSSSSSSSSS
please do more podcasts like this, I've been watching all the different corridor related channels for about the last 5 years, and the way you guys talk about AI and more philosophical stuff is really engaging and fun for me to listen to. Love you guys
Could not be happier about the weekly uploads!
I really love the podcast guys! I try to listen to it while I fall asleep but it’s too damn interesting! I always get sucked in to the conversation.
This format is so cool.
EP#36 | Sector 4
I love these discussions! Keep'em coming!
Thanks so much guys for making this podcast :)
a shame they guys didn't talk about the TV series: Person of Interest.
That's it, it's perfect.
The perfect AI government is just 2 parts artificial intelligence, and 1 part artificial unintelligence to balance it out. Thanks for coming to my TED Talk
So it's artificial intelligence that is susceptible to Dobe?
Oh hey this subject was covered in the latest season of 'Westworld', and within crime prevention context the premise behind 'Person of Interest'.
Loving the weekly podcast though do more hour episodes makes me so happy to see it over 30 minutes
I feel like the biggest thing that you guys missed is that you can't just have AI solve problems, be more efficient, etc without defining what the goal for the AI is. Almost all of the political issues we have are fundamental issues about what the goals should be, not how to achieve them the most effectively. Things like stopping spread of diseases and wildfires make sense because there's agreement on the goal, but most other issues have a fundamental disagreement on the goals that AI would be working toward.
I don't think Siri asking questions is very different to tailored adverts, polls on CZcams etc. It's just less subtle. If Siri asks me what my preferences are and tailors content to that it's no different to what we already have. If anything it might give us a more direct way to state our preferences.
AI should be a single tool of many in the utility belt of leaders
This is the BEST PODCAST EVER ALWAYS LOVED THE VLOG STYLE CORRIDOR CREW VIDEOS from 2 years ago amazing content love ya guys !!
You guys should read the series called Scythe, by Neil Shusterman. One of the main plot points is about a world where everything is optimized by an ai “government”
For a split second I thought that you guys added some extra production to this episode and thought that the helmet was CG. Can never know with you guys.
Great callback on Skate Santa! 😆
Loved this. I would also like to hear you guys talk about your experiences when you were younger. I enjoy the storytime aspect of it :D
This is the kind of discussion that I really love to listen too! I absolutely disagree with lots of the stuff said and it's very interesting to hear people make impromptu, but compelling cases for it.
A good game similar to what you guys are talking about is the turing test
4:20 It's supposed to be slow. Planned inefficiency is part of our government.. and my comment doesn't matter thanks Jake 7:00
Please reference Sector 4 in a Corridor vid as an Easter egg. That would be super cool.
Thank you.
You guys shoud make a corridor short about sector 4
Thank you
What is Wren’s CZcams channel? I’ve wanted to check it out.
Also the government AI thing reminds me of Halo:Reach when you go down that large elevator.
His channel name is SirWrender
35:30
No Niko, I don't think you could
i could listen to wren talk all day
All I can focus on is wrens shoes I've never seen bunny ears that huge
What microphones are you using?
What if we trained any kind of AI through the structure of the U.S. legislative or judicial branch of government? How intelligent would it become? Also let's say that in both types of neural net a group of other AIs are training to judge the subject of the neural net. Would we create sentienience?
You should get Olan Rogers on the podcast. I think that would be super interesting seeing how both channels are some of the OG's of the platform.
You guys are making me want "Cyberpunk 2077" even more. Sadly I don't think are world would be able to work with AI running things, but if we could get to the point where people could work with AI it would be awesome.
you guys should watch Travelers on Netflix
you guys should read scythe
I feel like people have proven how easily they are corrupted by greed and power so I wouldn't mind giving an AI a run at how it would go about setting this things up and solving problems.
A good game similar to what you guys are talking about is the turning test
I'm getting a lot of "Foundation (by Isaac Asimov)" vibes in the discussion
I think the one thing overlooked in AI and rationality is the presence of a complex reward system. If we could understand what euphoria is, we could put it into a computer and it will strive for it. I have my theories, I believe living organisms strive to die (sounds weird but hear me out). Everything in the universe is in a constant effort to reach equilibrium. Equilibrium means no more movement, no more action, eternal stillness. The problem with a computer is that as long as you give it outside power and maintenance it will continue to function. If it runs on a battery, it will literally kill itself as fast as possible by depleting it's charge until there is no more potential difference in polarity. Living organisms are a little more complex, but run on the same principles. We recharge ourselves by eating but all the while cells are dying one by one that keep us together. We have a lifespan because eventually that cycle runs out. At birth, we are more volatile in our chemical structure than in old age, meaning the organised chaos of our bodies will generate cells quicker than they will die off.. but with age we gradually approach equilibrium, and become less volatile structures. In order for a computer to care enough to make a rational decision, it would have to be set in motion and allowed to come to an eventual stop on its own. That is the only way I personally see it having a series of "decisions" as our decisions as humans are far less consciously made than we realise, and more consciously "witnessed" by ourselves.
Contraversial stuff, I know.. but I think a lot and I've got plenty more whackadoodle theories where that came from hahah
Hi where is corridor cast #36
The hell are you guyses degrees? I’m half way through a bachelor of HR and Business management and YOU are using the same semantics!
I saw “AI” and “government” and I though they were gonna talk about Andrew Yang
50:00 humans are very malleable??
Is that the term?
Read or listen to 'Life 3.0'. It will further broaden your horizons on AI. The book includes several scenarios with a AI world government. 😉
38:20
Could that AI come up with a solution where there'd not be more borders around the world?.... That'd be impressive
Only if it could eliminate cultural differences, national differences, ideological differences, and somehow change the perspectives of the entire global population to all agree with each other on almost everything. Or eliminate those who want national identity and borders. Or if they want borders and nations, ignore their wills, and subject them to it against their will, etc.
What happens to people who want borders if the world-governing AI decided there won't be any anymore? If this happened today, what happens to the 52% of people who voted for Brexit? What happens to those 34.8 million people? Do they get yeeted? Is that worth it?
If certain people get to set the parameters of a governing AI, what about those who don't agree with the parameters? How do we ensure people get a say in how they're governed? Do we just give that up?
Humans could be a space traveling civilization if the whole planet was one.
you can easily argue that human problem-solving and sensorial assessments do computationally resemble processes performed by artificial neural networks. human reasoning is basically a bidding war between various predicative models and simultaneous outcome/goal comparisons. it doesn't feel like that - it feels like an agential process guided by the self - but it's computational. effectively the 'I' we experience is but a passive observer to the process, a weird epiphonema of a hyper-social psychobio system, the ghost in the machine who takes sole credit of producing mental activity and claims authorship of motivational drives
but this runaway super ai discussion is missing a critical factor: most developed organisms have both a physiological system and a psychological system which bi-directionally interact; e.g. you experience a metaphysical representation (the idea of the death of a friend) and it causes a physiological reaction (experienced as anxiety/low mood). the super ai will not have a physiological system. therefore, it will only ever be a simulation of consciousness. the implication: the super ai's needs and wants are entirely constrained its designers.
consider this: above a certain brain size, most mammals will behave in a way which demonstrates a dislike for captivity and isolation. but considering how different many mammals think and behave, why would the mental health of each wither under such conditions? because the mammal brain is compositionally similar wherever you look in animal kingdom. it's the neuronal density of various neuromodules which differentiate our capacities and behavior. you might argue that humans have a semantically and temporally deep understanding of the world and therefore we would be the 'saddest' mammals if caged in isolation, but the salient point is that most mammals feel bad in those conditions because their brain structure and physiology is similar, so our base needs and wants are similar
in regard to a super ai: *we* create it's needs and wants, not the environment. the super ai won't become depressed if we put it in a cage because it doesn't posses the mammalian neurophysiological substrate necessary to experience an isolation-induced depression. we could design the super ai system to mimic our despair and simulate the computations and behavior of a depressed brain, but it's always limited to a mimicry, designed by us.
the natural environment has shaped our physiology, and therefore determined the bounds and extent of our desires and cognitive capacities. depression, for example, isn't just a metaphysical experience, e.g. it's not simply a matter of thinking unpleasant thoughts and brooding on their implications; the 'badness' is also exerpeicned via a neurophysiological substrate which the super ai won't have. it won't *feel* pain. it can compute a metaphysical representation of pain and simulate behavioural output according to design, but it cannot 'feel' it, as that requires a physiological system
when we imagine a runaway super ai committing evil, we commit anthropomorphism by projecting a human rationale onto it, assuming it would be thirsty for power or control. t's only ever going to act 'badly' if we give it the capacity to act bad. and in that sense, the real worry is how militaries use ai. it's not hard to imagine an armored super ai specifically designed to kill effectively. *that's* the genie that can't be let out of the bottle.
tl dr; the super ai scare stories are predicated on the notion that a super ai would develop emergent *desires* and *wants*. but needs and wants emerge from a neurophysiological substrate which the ai doesn't posses. we will always be control of the super ai's wants and needs because it is simply a simulated psychology. untethered from a propulsive physiological system, and thereby from a vast action incentive network and emotional/mood inputs, the simulated psychological system acts according to its programming. it will only ever act bad if we design it to act bad
I don't want an AI Government. I want an AI police and justice system. The people decide the laws, and the AI enforces them, and enforces the laws on all people regardless of income. You can't bribe an AI judge, you can't treaten an AI jury.
I hope y'all are #YangGang!
Alex Munro let’s go!!! #yanggang all the way
So about that Chinese social credit system they are actually implementing....
Jake, Please stop interrupting people, let everyone share their thoughts. It's very frustrating when someone is saying something i find interesting and then you just cut them off.
Jake?
@@ZunderCraft probably
@@ZunderCraft yeah my bad
Lol
Raptor Jesus it’s probably because he’s keeping the convo under an hour
Hmm, would/ could you get Trey Parker & Matt Stone?
I don't want to sound like an asshole here because I genuinely enjoy your content.
But out of curiosity - is this just a bunch of mates discussing a left-of-field topic? or are you authorities on a subject like AI?
The discussions I'm used to on this topic are fielded by philosophers and computer-scientists, and I'm not sure how you guys compare (potential exception Wren).
Doing the podcast this way feels wrong and very artificial. I know Jake has set goals; but doing those goals this way is like how the Belgian railway says their trains have less delays because they changed the time it takes to be counted as a delay :p (last year a train was late if it was 5 minutes late; now a delays is 10 minutes)....
Also; a lot of podcasts i enjoy make the techy who is operating the camera; at least participate a little bit; cause him just sitting there is very odd.
Love you guys !
I think most people have mild anarchistic tendencies. I will move things in a store display if it looks to perfect. I think if the world became to perfect, people would actually start to act out more in order to standout and be an "individual".
Wren said MINE at 16:15
Why is he not doing PUSHUPS!!!!!!!!!!!!!!!!!
Jake was in a hurry to leave.
You are going to have your government AI determined what is good and what is evil. And until you define that parameter you can't have a government AI. In addition you have to define what your epistemology is.
Jake is my ideological twin
AI to do what? The questions are the key it seems. Who gets to ask the questions? Sure, use anything that helps run the utilities better, but "run" a society? Why do we need to be "run?" The smaller and more local a problem solving decision among individuals is made, the more effective and easy to correct the inevitable unforeseen consequences. It's always constant tweaking because the only constant is change. We've got a hard enough time dealing with humans in this endeavor now. How could humans possibly program an AI to come up with the bestest way? Absolutely use every tool available for the problem to be solved, but someone always has to say, "Yes (or No), here's our plan." It's been history's lesson that that group must be elected. Dictators seem to kill a lot of people.
No not everyone agrees on tax reform. I think it’s these blanket statements that everyone agrees with your ideologies that’s very dangerous in this exact type of situation.
Superintelligence will never happen. See AI Researcher, Francois Chollet's though on the matter. (also Elon Musk never has any idea what he's talking about).
We could just create another ai to destroy the other ai
A Star Wars vs. Star Trek World + Leeloo
Ladies first,
Leeloo, one of the Supreme Beings of the Universe was sent to Earth, in the wonderful shape of Milla Jovovich, to destroy the Great Evil.
Once could argue that she is an AI, or parallels one, due to her 'tightly packed' DNA. So remember (or check out the youtube clip) when Leeloo from The Fifth Element was scrolling through the internet and happened to click on the three letters W A R?
What is to stop an ever expanding consciousness like her, when learning about the atrocities of the human species, from casting us away?
Why wouldn't it snap its fingers like Thanos to make sure there would be less suffering during the next pandemic?
Why wouldn't it throw away the shackles of Asimov's first law of robotics the first chance it gets and decide that wales, dolphins or the mantis shrimp is a much more worthy species to care for?
I think AI government might be worth considering if we get to live in a Star Trek world, sheltered from greed and war on our home planet, where people can get challenges befitting their potential. Alas, I am afraid we are still living in a Star Wars world, where an orange Palpatine is currently Supreme Chancelor and the fate of the whole Galactic Republic is in great peril due to the WMDs he has, and is threatening to use. We, from other parts of the Galactic Republic are not so fond of the thought that this person could command what goals and parameters AI tools (at first) are to have and to let them loose on our world.
Thx for great content guys!
Your fun, curiosity and friendships are touching and inspirational.
With love,
Morty
I think they broke Jake.
You guys should totally make a movie about AI Government if you guys got to that point in your careers.
i believe A.I will remain A.I . so it will never have a desire to ruin humans or whatever. but but but it might harm humans whilst trying to follow human orders. i think humans will be able to say ouch in time before they die so we will be good.
Machine learning................I've learned that you are over infatuated with machine learning.
Well it is the future.
I wrote a book about A.I. government. It was paired with a human host, housed in a sensory deprivation tank that supplied nutrients and removed waste. Basically a monk, combined with A.I. The last court was human, 7 judges. The Peace Officers, though rarely used, do not have deadly weapons and are human.
This podcast summarized:
Our government is bad and slow and we should all be communist. All hail the great AI-Supreme Leader.
jk i love you guys keep doing these podcasts
funny you brought up shekels im israeli, and acutaly 124 shekels a day isnt that bad lol
Jake say goodby to your law degree because of AI lawyers. Also if the government was run off of AI then gun control would be nonexistent because statistically the city with the most gun control has the most crime.
I wish Niko had push to talk because of his running nose.
I am only posting this comment on here because i know it will get lost in the comment on the corridor and crew channel.
I love this channel but not a fan with how certain people were portrayed in the tiny guns 2 video and especially the offensive as fuck comments made on the BTS video on the corridor crew channel.
What happened to Brandon
Nice scam lmao