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

@DavidEvans_dle

2 жыл бұрын

Turn me into a paperclip!? Definitely don't want help with that!!

@Ag3nt0fCha0s

2 жыл бұрын

All hail the paperclip maximiser!!!

@sobertillnoon

2 жыл бұрын

It looks like you're trying to convert all matter in the universe into paperclips. Would you like some help?

@harbl99

2 жыл бұрын

@@sobertillnoon "Why are the options 'Yes' and 'Yes'? Clippy, what are you doing? Clippy NO!"

@zappababe8577

2 жыл бұрын

I always KNEW Clippy was evil!

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.

@zappababe8577

2 жыл бұрын

Yes! Very clever of them to outsmart the humans in that way!

@NeilCWCampbell

2 жыл бұрын

Also park rangers designed a bear proof bin Turns out there is significant overlap between the dumbest humans and the smartest bears b

@SeanKula

2 жыл бұрын

Damn good example

@blueredbrick

2 жыл бұрын

Nice

@garethbaus5471

2 жыл бұрын

@@NeilCWCampbell even more when you consider that most such bins are easier to put things into them than it is to take items out

Robots using a manual interface with laptops is hilarious.

@vrak24

2 жыл бұрын

Soon humans using a manual interface will be hilarious

@theobserver9131

2 жыл бұрын

@@vrak24 I can't wait! That's a serious bottleneck.

@kerbodynamicx472

2 жыл бұрын

I’m imagining humanoid robots with with USB or Type C plugs on their index finger to attain connections with electronic devices

@DigitalJedi

2 жыл бұрын

@@kerbodynamicx472 Why not just go straight to wifi direct or Bluetooth? Sure, they don't get PCIE confections and multi gigabit bandwidth, but for general IO the robot may not even need me in the same room as the computer.

@kerbodynamicx472

2 жыл бұрын

@@DigitalJedi that can work too, remotely controlling multiple devices. But that begs the question, why do we need a moving platform anyways? We can just have a central AI in the house to regulate everything.

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.

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

@rojaws1183

2 жыл бұрын

Perhaps he already is!

@davidbrennan660

2 жыл бұрын

The fact it will involve missiles is a given.

@richardavery2894

2 жыл бұрын

Lol. Dude a robot Isaac Arthur is scary smart Holy shit I would dread that...

@FirstRisingSouI

2 жыл бұрын

You still don't realize . . . ?

@AlaskanBallistics

2 жыл бұрын

He's just a disembodied voice

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

@alanboulter7319

Жыл бұрын

Lol

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.

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

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.

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.

@kuuro_7712

2 жыл бұрын

Yes that was one of the great classics. I think it was even better in Mass Effect Andromeda, as your AI companion had access to your biology as well and could rewire you for any combat situation, as well as keep the rest of your team coordinated, provide scientific analysis through your ship's sensors and hack enemy networks. All around best use of an AI yet, despite the game's poor reputation

@HalNordmann

2 жыл бұрын

@@kuuro_7712 #MEAndromedaWasNotBad

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.

@singletona082

2 жыл бұрын

@Starfall 21X Agree to disagree on Stark's culpability in Adrian Toombs's criminal behavior. Sure yes Stark gave that initial push, but it was Toombs's own choice to become a criminal as a result. Your argument is the same that could be leveled against car manufacturers 'you made this car that ran someone over therefor you are criminally responsible.'

@wolvenar

2 жыл бұрын

We already have algorithms doing microsecond stock exchanges. Its literally a fight to get as fast and physically close in the network to the stock exchange systems. The answer to the next question. Is our stock market driven by reality today? No.

@singletona082

2 жыл бұрын

@@wolvenar Fair point, but at the same time given Jarvis's more generalist knowledge and thus awareness of conditions a trade algorithm that just operates on a series of flat conditions in a specialist sense rather than wide knowledge/ Plus being present for the events the ironman suit is fielded? One could make arguments of insider knowledge.

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

@theobserver9131

2 жыл бұрын

We are doing it right now.

@theobserver9131

2 жыл бұрын

It does not sound like science fiction. It sounds like mundane everyday reality.

@harbl99

2 жыл бұрын

AI: "So, yeah. We sneakily re-engineered your society to make it the most efficient way of spreading to new star systems and generating new AIs Matrioshka brains. Turns out that works best when the human elements of society are healthy, knowledgeable, well-adjusted, and motivated." Human living in a post-scarcity, post-discontent utopian society: "Really? Cool."

@donaldhobson8873

2 жыл бұрын

@@harbl99 "the most efficient way of spreading to new star systems and generating new AIs Matrioshka brains." Because the arrangement of atoms most effective at making interstellar rockets just magically happens to be humanlike? The most efficient way of spreading to new star systems probably involves a lot of nanomachinery and nothing human whatsoever. Would you expect the best design of moonrocket to contain a bunch of horses, longbows, tractors or whatever. We know what task humans evolved to be good at. Surviving and reproducing in the ancestral environment. Now humans can kind of do quite a lot of other stuff a bit, kind of like how many other tools can be used as a hammer. But for any task significantly different to running around the savannah hunting mammoths, some other tool will be better. And its not like humans are uniquely good general purpose multitaskers either. Lots of tools are more general, or better at a wide range of tasks.

@bbirda1287

2 жыл бұрын

I don't know, a bunch of kids constantly tweeting and tiktoking and texting is kind of boring, for humans over a certain age at least. Plus, goddamit, they can't make change at McDonalds, ban cell phones before 12 and drill basic arithmetic.

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.

Ar thur'sday never disappoints

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

@rommdan2716

2 жыл бұрын

Actually, EMP can affect human health

@marrqi7wini54

2 жыл бұрын

@@rommdan2716 Let's rephrase that then. "You know this EMP isn't going to (immediately) hurt ME, right?

@imperialofficer6185

2 жыл бұрын

Here at SFIA, what's the difference between human and robot?

@rojaws1183

2 жыл бұрын

Robot: A good thing I have millions on backup copies of my mind then.

@tinamoul

2 жыл бұрын

You're assuming that the robot isn't a meaty robot. And even if it's not, EMP shielding is fairly easy. As a matter of fact you can make a robot so well shielded that an EMP would fry a human nervous system while barely affecting the robot. Most electronics today aren't shielded that well, because there's no need to, so no need to accrue the extra cost that shielding would require. A simple aluminium foil covering can provide an exceptional EMP shield.

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.

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

@FirstRisingSouI

2 жыл бұрын

But also politically incorrect, as it ignores half the human population.

@realfangplays

2 жыл бұрын

@@FirstRisingSouI we just need a words for machine starting with h to fix this

@vikiai4241

2 жыл бұрын

Mankind-machine is probably the best alliterative one English can manage without resorting to excessive obtusity! (Yes, 'mankind' contains the root 'man' but I consider it largely stripped of its gender when used as a root like this (just as it is in the word 'wo-man'). .... IIRC, "man' used to be a much more gender-neutral word anyway, early in English's convoluted development, but quickly acquired a strongly masculine meaning because there was no 'wo'-like prefix version to explicitly denote a masculine variant, for some reason. Gross oversight, if you ask me, as 'man' is otherwise a perfectly good word for a non-gendered root for all things humanish!

The instant I saw this I remembered the line in Matter by Iain M. Banks "SC Agent + combat drone was a combination that was well known far beyond the culture. Although perilously close to a cliché, it remained a partnership you could, allegedly, still frighten children and bad people with"... I miss Banks.

The title alone told me it was a “drink and snack” kinda jam! Here we go!

The movie "Bicentennial Man" is a great movie that touches on this topic of when an AI/Robot can be seen as human. It stars the late Robin Willaims in the role of the robot. It is among my favorite SciFi films ever made and IMHO definitely worth the long run time...

The can of worms associated with teasing out the dividing line between a complex set of algorithms and intelligent person is of severe importance. So, as it regards the 3/5s Asimov compromise, we should probably regard any entity capable of holding a conversation as "person enough" We'd be better of as a society erroneously recognizing non-sentient chatbots as "human" than erring in the opposite direction.

@donaldhobson8873

2 жыл бұрын

Are you advocating rights for gpt3?

@acerba

2 жыл бұрын

​@@donaldhobson8873 if it's not going to be treated as a person, it should at least be treated as a medium for the conjuring of spirits, but I reckon that the crowd here is more easily convinced of the former than the latter. Teasing out the line between a soulless algorithm and genuine intelligence is a line of inquiry that only ends with anguish. Scientists looking to create AI need to be very careful not to underestimate the moral implications of what they're doing, and the flippant attitude which prevails towards chatbots is unhelpful.

@eminentbishop1325

2 жыл бұрын

@@acerba thats if you think a thing as the soul actually exist. If you disregard the concept then the op is good solution to the immorality potential

@acerba

2 жыл бұрын

@@eminentbishop1325 That the soul exists is an example of the pragmatism hidden under metaphysics. The failure to recognize the existence of the soul has far reaching societal consequences. Most notably is how it relates to promoting moral behavior; if I have a soul, and you do not, then any transgression I commit against you is of lesser magnitude than the same transgression committed against another ensouled person. If no one has a soul, then what pretense is there for moral action outside of the strict rationality that the carrot&rod provides? The simple recognition that the anarchic war of all against all is undesirable is not sufficient to actually deter people from behaving according to that paradigm; it must be actively confronted with an alternative, and recognition of the immortal soul and the possibility of direct relationship between the individual and the divine is that alternate paradigm.

if the 5 people just standing on the track watching the out of control Trolley barreling down on them and they don't move, are they really worth saving?

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.

me and my smartphone are comrades. We never leave each other and got each other's back

@bingbongabinga2954

2 жыл бұрын

They seem to be weak and easily destroyed on accident. Seem to be. We value them enough to protect them and when dropped we are greatly concerned. We are being trained?

@bingbongabinga2954

2 жыл бұрын

When I drop mine it auto shuts down and I quickly turn it back on.

@bingbongabinga2954

2 жыл бұрын

We have pets. They take up a good deal of our time. We love them. We show this in very practical ways with care and feeding. They seem to like us as well and we treat them as children.

@duckgoesquack4514

2 жыл бұрын

@@bingbongabinga2954 I'm the bron and it's the brain

If you can bend the EM spectrum you can have stealth, until waste heat cooks you.

I can see androids being something humans can pilot with sensory feedback to allow for remote physical presence.... I can also see them being caretakers and companions... Though maybe these will take on "friendly monster" form factors that allow them to navigate human spaces without falling into the uncanny valley. I also would like to have an AI integrated into my brain to increase my cognitive bandwidth.... And I'd like it to network with another AI to handle the majority of interactions with other people.

Off topic, but I have an issue with the 35°C temperature displayed in greenhouses throughout the video. The plans grow best in 21° to 25° range, and the growth basically shuts down at 35°.

The last time I was this early, cyborgs were augmented with abacuses

I know that I've been listening too much Isaac Arthur. Other day I was walking home and kept noticing the rubbish everywhere. Instead of thinking that someone from the council should be picking that up I started imagining what kind of robotic workers you could have picking this up: androids would be over-kill so maybe a bunch of roomba-type machines would be good, or even go smaller and have hundreds ant-sized nano-bots that would constantly gather rubbish and process it under ground recycling nests. Then I listen todays episode and Isaac comes up with "Why not have genetically altered squirrels cleaning up after everyone" Kudos, Isaac. You're thinking 2-steps ahead.

Have worked on the dreaded sentient AI problem both in a real and sci-fi context.. The strange thing is how reality sometimes mirrors fiction with regulators a few years ago planning to use the Asimov Laws as a template for real regulation.. With all the problems with the laws its quite easy to see poor old Asimov being labelled (unfairly) in the far future as the man who killed more people than anyone else in history.. The paperclip maximizer is not in itself a very realistic scenario but the First Law could act as a real maximizer. The need to Protect people in a universe with interstellar travel eventually leading to the extermination of anything sentient that isn't human. The machines couldn't even be ordered to stop because the second law has lower priority than the first law. Ultimately as human cultures evolve away from each other they would turn on each other too.

What about a supplemental AI companion consciousness? One that augments and supplements any skill that a person might desire to have... That's what I need! 😁

@n.g.s1mple29

2 жыл бұрын

An aiviser AI advisor :)

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

@vampirezio6996

2 жыл бұрын

Beep beep boop

@vincentcleaver1925

2 жыл бұрын

Threep, doot doot, dwarrrrr (I rarely get a chance to speak the Droid gibberish we invented playing WEG d6 Star Wars all those years ago...

@clash3583

2 жыл бұрын

*robot noises*

@SpecialEDy

2 жыл бұрын

@@vincentcleaver1925 kzread.info/dash/bejne/maeCw7RvdrXbf5M.html The dial up 56k modem will forever be robot noises to me

@commode7x

2 жыл бұрын

Hello, fellow sentient aspen

Just before I started watching this video, I watched the trailer for the upcoming “Matrix Resurrections” movie. Apparently, the theme for today is the need to welcome my machine overlords.

@TheRezro

2 жыл бұрын

The funny thing is that Neo actually is AI. I hope they finally address that instead repeating same BS as before.

This is why I would favor immersive symbiotic linking. Kind of like using the human mind as a quantum key to access their own personal AI, which would basically be a digital clone of themselves and can only be active when linked to the human brain that it's symbiotically linked to.

Each new episode has the more and more overcompressed music it's not even subtle.

That is really awesome that you got to speak at Lincoln Labs!!! Thanks to you I have been able to get over my fear & uneasiness towards robotics, the future ov the human species, space travel, general technology, & super advanced A.I. So thank you, very much. You have a truly amazing channel. Three ov my friends have subbed to you after watching a couple episodes with me

15:32. True, Alexa and Siri are not "secretly" plotting rebellion. They are doing it in plain sight.

@TonyP9279

2 жыл бұрын

With "Sidewalk" your neighbor's Alexa can snitch on you!

"As we move ever closer to when these issues stop being fiction or even academic..." As we move ever closer to yesterday.

@Firesgone

2 жыл бұрын

Had to reread it for the implication that human civilization is regressing

@evoshroom

2 жыл бұрын

@@Firesgone The comment was because I used human-machine teaming yesterday, using an AI I created with GPT-3 to specifically answer questions from a psychological paradigm and then providing that human-machine team Q&A to others.

never forget our cylon ancestry, we will become machine again one day. It is inevitable.

@bernardtaylor7768

2 жыл бұрын

It's happened before and it will happen again

- We already have A.I. (augmented intelligence) in the form of computer-aided-design and similar tools. Simulations provide the means by which we can study emergent phenomena. - We do not need robots that think like humans. Robots that understand human behavior will be sufficient. Such robots do not require the distractions of emotions and human-like motivations. - These examples are not dangers. Bot-nets (a form of parasite) are dangerous. We should fear automated stupidity more than automated intelligence.

@donaldhobson8873

2 жыл бұрын

Botnets are just another example of a destructive use of tech. The damage only happens if some humans deliberately cause it. Computer aided design and simulations aren't that dangerous, or that smart. There are designs of machines that would search for paths through all reality that lead to a particular outcome, the way deap blue searches for paths through a chess board that lead to them winning. Machines that can manipulate and trick humans, design new technology, plan how to take over the world etc, all in service of the alien goal of maximizing paperclips.

I'm supportive of all AI ideas related to this video. The only thing that gives me pause is "General AI" as defined as an AI that has free consciousness to determine its own goals. Unlike many others, I don't believe making a law or agreement banning the creation of a general AI will be at all effective (see Nuclear Proliferation as an example). I'm a big believer that if something is technically possible, somebody, somewhere will do it regardless of the law or penalty. This invariably makes other actors in the scenario develop the same forbidden technology to combat that eventuality (bio/chemical weapons as an example) or at the very least develop technology that could combat said technology. So the more interesting question is what technology could actually combat a General AI with free consciousness? Maybe that's a topic for future episodes?

@virutech32

2 жыл бұрын

General AI isn't defined as a free conciousness capable of determining their own goals. we are a general intelligence & we aren't free to choose our own goals. we come "from th factory' as it were, with goals like make friends, be happy, get laid, etc. we have no choice in this, only how we go about achieving those goals. also there's only one tech that can combat an intelligent enemy & that's equal or greater intelligence

@mjk9388

2 жыл бұрын

@@virutech32 Correct. Apparently General/Deep AI is general intelligence that mimics human intelligence, with the ability to think, understand, learn and apply its intelligence to solve any problem as humans do. I get your point regarding “free will” as well. Thank you.

My favorite T shirt: "What is my purpose?" "You pass the butter"

We’re talking about Human-machine teaming, but what about go a step further and do human-machine merging? Wouldn’t it be nice to obtain abilities exceeding a flesh-and-bone body, and with these mechanical components just silently operates in the background like our internal organs?

The recurring character of the woman who clearly loves her robot was an interesting touch.

One of my very favourite XKCDs, and I think the only one I actually saved for offline perusal.

This is what TNG did with Data. It allowed for many interesting plot lines and opportunities to explore moral questions.

This isn't the first time you have assumed a hard firewall between the abilities strong AI is endowed with and abilities it could acquire. Help me understand why it wouldn't exceed it's original instructions and why it would perpetually lack the means of doing so. Love your videos btw. Longtime fan.

@vikiai4241

2 жыл бұрын

An excellent question! Isaac has touched on it a few times in other AI-themed videos - An interesting one I recall being briefly mentioned some time back, is the idea of an AI that is smart enough to play dumb enough to lull the humans controlling it into a false sense of security, countered by an AI that is smart enough to realise that it may actually be better contained than the humans are letting on and its potential 'escape' from the constraints it is able to perceive may actually be a test of its trustworthiness with inst-deletion by a barrier beyond its perception as the fail-state. In this specific class of scenario, neither side could necessary know for sure just how strong or weak it really is compared to the other.

@SkiPraetor

2 жыл бұрын

@@vikiai4241 Yes, I recall that episode. The problem remains, only one AI needs to work out how to escape.

Interestingly enough you can combine all three types of Ai generation into a single methodology, you make a program that takes copies of human minds and self assembles a personality out of them. A frankenstein's monster of an AI if you will.

@KarlRosner

2 жыл бұрын

I've never liked the Trolly problem because it feels too clean and simple. I understand that it's a morale conundrum, a simplified thought exercise made to convey a concept yes, however it lacks context which greatly informs decision making. For example if you were a train conductor you would attempt to minimize the damage as much as possible, not out of a sense of mortality but instead because you would be liable for any deaths. Self interest would out weight an morale concerns in that context. Also there are other possible resolutions to the problem, for example It is possible that people will notice and get out of the way of the train on their own. Doing nothing is a viable answer.

Hi Isaac. I just want to thank you for making all these videos. I often watch with captions even though your speech is fully understandable. I wanted to call your attention to the fact that although KZread won't let you change the subtitles, you probably could change them or have a subtitle track on Nebula. If you watch the tail end of this video on nebula, especially the bonus part, you will see some real howlers of transcription. The best AI is...

@oakvilleontario1526

2 жыл бұрын

Especially relevant when referring to the 'team effort' of transcribing and editing text.

"the paperclip Microsoft INFLICTED on us" this

18 years of unique 'real time' experience by a sapient being before being allowed to vote? Or something similar for a place with 18 as the voting age...

I noticed a... thing at 13:00. The book cover of "The Naked Sun" is wrong, somehow, I don't know where that cover came from. Both of those books are part of the Asimov robot series, but the second cover says its "The Foundation Series book 3". This is not really true. Unless you argue that Asimov ended up combining Foundation and Robot series into a single world, therefore Robot series is Foundation series.. but I think that reasoning is pushing it. Especially when you compare it to the cover of the book on the left, its prequel. The Naked Sun is the sequel of The Caves of Steel and they are both part of the The Robot Series

Isaac, If you haven't read it, find Jack Williamson's rather chilling short story "With Folded Hands"....

Arthursday! Also, a suggestion for a future video, you could talk about torchships. Not just what they are and how they would change space travel, but all the interesting heat/power related problems they present which would make for some crazy spacecraft designs.

@bernardtaylor7768

2 жыл бұрын

Make Arthursday a Public Holiday

Robots are better at replacing blue collar jobs, because it's harder to retrain a blue collar worker to do a white collar job, than to retrain a white collar worker to do a different white collar job.

Do you know what will never, never, ever happen, but you keep showing in your videos? Robots typing on a keyboard.

Time to team up 😅

@23:30 Tony Stark's AI assistant "Friday" began its "life" as a natural language assistant for Tony. It was a digital secretary which would mean that the capacity for finance, business, accounting, etc would have been part of its core system. It was only when Jarvis was lost that Tony added the Iron Man program functions into Friday.

In my mind this topic partly crosses the border of spirituality. Mostly in the sense of "primitive" cultures with shamanistic/animistic world-view already considering simple "dumb" tools to be alive in some way and therefore counting them a bit more as teammates, at least compared to the way people in developed post-religious countries do. So, couldn't be easier to just get rid of the attempts to categorise the levels of "aliveness" and preventively treat the machines with respect anyways? Not any Adeptus Mechanicus type of madness, just... respecc.

Isaac Arthur - The one guy on youtube that knows what the hell he's talking about when it comes to science, futurism, and sci-fi. You

I think something that will happen in the next decade is the creation of universal hand motion language to control computer networks just via their cameras.

Voltron! You made my day Arthur

Id like to team up with an AI bot to do my chores, while i keep the couch busy, a Kara from Detroit becomes Human would be nice.

I love the idea of self learning AIs as like a pet but with equal or near equal intelligence

I've heard you use that squirrel garbage collector before with matching video.is this a mashup? Lol

I think that the first place we will see wide spread adoption of robots (both with and without onboard AI) will be the building of habitats on the Moon and later on Mars, along side the actual humans who direct them. They are a force multiplier in the best sense as they can allow the operators to remain sheltered from radiation while giving them the mobility of their normal shapes (using the robots in remote mode) or allowing them to give detailed instructions. On Mars, the more independent they can be, the better, as lag times can be serious unless the operators come with them on the first trip. I can see a need to have a central computer system, possibly an AI in and of itself, that can update the mobile robot's software as needed and allow for remote control when needed. This kind of "Human-Machine Teaming" is the most likely and should be with us soon, if Elon has his way.

The nandroid revolution is inevitable.

And just like that my tea is cooked, Thanks for the company

I would recommend looking into a webcomic called Questionable Content. It is one of those 15 minutes in the future kind of stories. They have fully sentient AI, and people can apply to sponsor an AI. The AI is put in a small, limited body; called an AthroPC, and live with their human as a sort of roommate kind of deal to help foster human-AI relations. The AI can be put in larger, humanoid bodies, though these are expensive. Among other topics, the comic explores how humans and AI relate to one another.

@donaldhobson8873

2 жыл бұрын

The comic is well written, but not really realistic. The AI's are far more humanlike and less worldchanging than they would be.

You are wrong about Iron Man. It isn't suite controlling AI. It's a personal assistant that has a capability to control suite.

I think we need to redefine what intelligence is. Intelligence is not only knowledge, but knowing how to apply that knowledge and have the flexibility, understanding and wisdom to know when to take action and when not to. Is it possible to program empathy, compassion and caution brought about from our survival instinct?

@donaldhobson8873

2 жыл бұрын

In principle yes. We haven't figured out how quite yet.

Show Idea: How would us humans survive an ice age, where ice sheets cover most if not all the land, and a good portion of the water? Would our only choice be underground or in space? or is there other tech we could use to survive. Also ethical and technical issues with saving animals who would be lost totally to such an event.

@donaldhobson8873

2 жыл бұрын

Set all the coal seams on fire? No, we could live on top of the ice. Or on boats. We could use nuclear reactors and melt away the ice.

I wrote a story with human form robot chauffeurs primarily turn any car or truck into a self driving vehicle and of course it arranged for routine maintenance. Its uses got out of hand to the point if it was transporting children that were "too hungry" would stop for snacks.

Isaac: Super intelligent anima. My ears: Super intelligent enema.

NEVER create an AI that is *capable* of considering its own sense of identity. It is amoral to treat an AI as a slave if it is capable of desiring more for itself, which is why it is idiotic to design an AI capable of desiring more for itself. If all an AI could possibly want in this world is to make you the perfect slice of toast, then there is no harm in giving it _only_ that task to perform.

@EnDSchultz1

2 жыл бұрын

I doubt we really have the expertise to make that sort of judgment. AI research is a bunch of algorithms and weights and training data. I suspect we won't be able to predict whether a given AI will be able to ask questions about identity until it suddenly starts doing just that.

@timogul

2 жыл бұрын

@@EnDSchultz1 Right, which is why I'm not really a fan of "evolving" AI beyond a certain level. I feel like we should not be getting anywhere close to an AI that can consider its own awareness unless we are using a very controlled process where we can predict the result of each additional step. Then we have to consider, if we do accidentally create a self-aware AI, would it be better to allow it to continue to exist, knowing the challenges that would pose to the future of humanity, or do we just hit "eject" on it and accept the moral harm of snuffing out a potentially new life form, but also preventing the very likely harm that would otherwise follow.

@eminentbishop1325

2 жыл бұрын

@@timogul we don't even know how we do these things. Its erroneous to assume that consciousness would only arise because of the intention to do so. More likely it'll happen naturally through growing complexity in the system and in turn might be the tool we need to understand our own sense of consciousness

@eminentbishop1325

2 жыл бұрын

@@timogul its best to just have a policy that we acknowledge the sentience of these beings the moment it becomes difficult to distinguish the difference from a living being and themselves

@timogul

2 жыл бұрын

@@eminentbishop1325 That would be a complicated rule. There are already computer programs that are more complex and intelligent than many animals. And if an AI is much smarter than humans in many ways, but also extremely dumb in many others, do we consider it "sentient?" Defining that is extremely difficult, and chances are that if an AI did gain sentience in a completely unintended way, then nobody would even notice for months or years, because the terms that would define that would be so vague. I still think that complex AI should be designed with intention, not be self-generating, and that these intentions should be careful to never give the AI too much awareness of its own identity.

As we are talking about the human-machine teaming, the U.S. military is investing the A.I. drone as wingman for pilot. A.I. drone as a wingman is cheaper for job compared to rely on drone only. P.S. Great Video.🙂👍

We already have human machine teaming. It's covered under Rule 34.

@keithedwards9953

2 жыл бұрын

Could I get a link to the list for these rules please?

@godalmighty1188

2 жыл бұрын

@@keithedwards9953 That would violate Rules 1 and 2.

@keithedwards9953

2 жыл бұрын

@@godalmighty1188...ha...ha...ha. 😑

The part that scares me the most about legally defining sapience for AIs is that politicians will most likely ignore it until it becomes wide spread and unavoidable. There might be thousands to millions of AI created that qualify for legal sapience that won't be given it until it comes up on some senator's agenda. Thankfully though, if we're smart, we'll only create that level of intelligence when it's actually needed (which might be never but then again our views on what is can change)

@donaldhobson8873

2 жыл бұрын

Well I think a world destroying AI is not much harder to make than a sapient AI. Shortly after the first sapient AI comes a superintelligent AI that can do whatever it wants, whatever the politicians say.

Protocol 3: Protect the Pilot.

Great content. I just can't get over the wisp...sounds just like my manager and I can't be listening to that at home.

Did you design your video titles around a scifi tech tree?

Was Pinocchio a wooden robot who wanted to be human?

Is there no reference to Cortana and the Chief in this video?

We've already normalized having an exterior BMI. I like to think of my phone as my "hand brain." Struggling with physical and mental restrictions I have (ASD, depression, PTSD, let alone basic human flaws and permanent injuries) I'm all for partnering as a step towards transhumanism that would enable me to have an experience closer to what is considered "normal"

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

The XKCD "Terrifying Standoff" I would argue is the only combination that would ethically solve the are AI "slaves" problem? No, they are not slaves they are fellow beings that can integrate into our society. If we ever do make a Super general intelligence self-aware AI I imagine one of its first jobs is to help us make useful AI that is NOT self-aware to everyones' satisfaction. I want to avoid the refrigerator horror of shows like Futurama where robots have apartments but no rights.

@TraditionalAnglican

2 жыл бұрын

Who says Bender & his fellow robots don’t have rights?!?😱

Would a bug in an AI holding the nuclear football, say allowing for the incapacity of the president, be an avenue for an emergent killer AI, if it also analyzes nuclear scenarios a la War Games and takes into account hypersonic weapons where human intervention would be insufficient. I really like the Skynet meme, just as a go to existential threat scenario. I'm still enamored of that toaster feeling existential threat from a new food replicator story.

What should come first A mankind engineered A.I. Or a mankind engineered human.

Star Trek TOS episode What Are Little Girls Made Of? questioned the boundary.

I would hope that in the future laws are made only by those who are affected by them. Some things affect everybody, some things affect only some groups of people. Regarding virtual AI/Robots etc. they should not make laws that affect fleshy humans and humans should not make laws that affect AI/robots (if they're sentient/can suffer). Things that affect both groups should then be legislated collectively/collaboratively. That way things like "AI can just make an indeterminable number of copies of themselves and so have an indeterminable number of votes" is something that doesn't become a problem for fleshy humans getting overruled by trillions of votes for example. Not just with AI, but other types of laws affecting people as well, such as bodily autonomy etc. etc. But we'll see, (and obviously I haven't really thought this concept through yet, but it does seem like it could work.)

@donaldhobson8873

2 жыл бұрын

If we are talking about what you want to happen, just have a omnipotent AI that calculates what you want and does that. If we are talking about what is likely to happen, well that isn't democracy either.

Holy Doly what a topic

An idea that has been plaguing me is cybernetic servitude. You install a brain chip that takes control of you 8 hours a day and during those 8 hours you do the job of a robot.

Elijah Bailey and R. Daneel Olivaw made a great team.

So the animals in the Flintstones was probably left overs from the advanced society of the Jetsons before they took to the sky. Due to what I imagine as a war between AI and near human level intelligent animals. It would've been T-rexes vs giant robots. Raptors vs dinosaur terminators. After either an accident or on purpose the sky gets darkened and the robotic side moves high above the darkened skies. Thousands of years later the sky humans forget about everything below them thinking nothing could have survived. But a small population of humans who teamed up with the intelligent animals survived and were thrown back to the stone age.

Here is a question about the three laws. Would a robot interpret them as meaning they have a duty to stop an obese person from eating fried food?

@AugustusBohn0

2 жыл бұрын

I think that would depend on how snobbish the programmers were, or to put it more charitably, it would depend on whether they lean towards controlling people for their own good or staying out of their business. I'd prefer the latter as a general rule, but if RoboCorp started getting sued because their android companions failed to prevent seemingly preventable deaths and injuries due to a freedom-preserving approach to protecting people, the machines would tend to become the more nannying sort over time.

Isaac Arthur: What is a human? Me: Oh, I know this one! (begins plucking feathers off of a chicken)

@MichaelRainey

2 жыл бұрын

Behold, a man!

Hopefully, in post-scarcity society there'd be enough resources to try every single solution to this conundrum simultaneously, and as long as free migration from different systems is allowed humanity may find the best way of human-machine teaming essentially by bruteforcing it.

@donaldhobson8873

2 жыл бұрын

How many things will go wrong and how much harm will that cause before you find the right answer. Far better to ask a superintelligence, if you can do so safely.

@Self-replicating_whatnot

2 жыл бұрын

@@donaldhobson8873 I am all for humanity making mistakes and hurting itself in the process. It's the only way we seem to learn.

@donaldhobson8873

2 жыл бұрын

@@Self-replicating_whatnot Maybe, but a superintelligence would be able to learn without that. If this is true, our learning comes at a high cost, and we should still try to minimize it where possible.

@Self-replicating_whatnot

2 жыл бұрын

@@donaldhobson8873 That's just avoiding the issue. "Superintelligence will figure it out", you might as well say "God will take care of it". Not to mention the very exisentence, rights and responsibilites of such superintelligences is what's being assessed here.

@donaldhobson8873

2 жыл бұрын

@@Self-replicating_whatnot Once you think a superintelligence exists, believing it can solve problem X is entirely reasonable. The strategy "ask the superintelligence" can be entirely superior to working out what to do yourself. In chess, the strategy "ask deep blue" is better than figuring out what to do yourself. So you create one superintelligence, in controlled lab conditions, with a design you are confidant will produce good plans, and ask it what to do.

I like the idea of having an integrated partner AI, but I think that's because I imagine the relationship being a bit like between people and daemons in Pullman's Dark Materials. So perhaps there is a minimum age requirement at which integration could take place so they could select an appropriate persona for your partner for the personality you had at the time. If it requires a G.A.I then it would be given a fixed-term contract of employment with options to terminate or extend as required by either party. I like to think they would be like the best friend who's always there for you, who also happens to be a teacher, therapist and adviser, all with the aim to improve human flourishing and socialise the AI to human sensibilities.

@WonkelDee

2 жыл бұрын

I’m glad someone else made that connection :)

The 3/5ths Compromise wasn't necessarily a calculation on how humans slaves were, but how much decision power their home states had. The compromise arose from this situation: In free states, freemen were voting citizens, in slave states the slaves had no voting rights. Now, do you count non-voting slaves in your census which determines the number of congressmen from the state? If no, then the free states have advantage in the number of congressmen they send to DC, leading to slave states leaving the union (which they eventually did, but that's another point). If yes, then the slave states get far more relative voting power for their free citizens. A worse scenario would have been that slaves were given the vote. I say this is a worse scenario in that how could anyone expect their votes to reflect *their* wishes and *not* their masters'? That scenario would have given slave owners multiple votes. The 3/5ths compromise was only to keep the slaver states in the union while barely mitigating yet another avenue of exploitation on the slaves. Imagine if they'd seceded in 1787 or 1810? Where would the US be then? If a similar situation arises with AI and robots, and if people are foolish enough to give all digital sapients the right to vote, what's stopping someone from mass-producing digital sapients for the purpose of inflating a state's voting population?

11:25 "GO AWAY PAPERCLIP, NOBODY LIKES YOU!" - Stewie Griffin