Robotics and AI: The Future is Finally Here
HTML-код
- Опубликовано: 1 окт 2024
- Though remote-controlled now, how long before Neuralink-controlled robots start building our Moon bases and AI soldiers start fighting our wars? The future is here... And those who were convinced it would never happen have been proven wrong again. Whether for good or bad will depend on us.
Just a little to think about on what might be a holiday weekend for you.
Stay safe and thanks for watching...
Shop the Academy store at...
shop.spreadshi...
Please help support our channel at...
/ terranspaceacademy
Thank you so much for watching!
Ad Astra Pro Terra
Artists
/ c_bass3d
/ labpadre
/ neopork85
/ hazegrayart
/ alexsvanart
/ _fragomatik_
/ nickhenning3d
/ rgvaerialphotos
Companies
/ nasa
/ spacex
www.cochranex.com
/ blueorigin
/ space_ryde
/ virgingalactic
/ relativityspace
/ neutronstarsys
Please never dumb it down for views. RUclips is full of shallow videos.
Thank you. The new AI generated, narrated, and imaged content is making it hard to do my research. They look and sound useful until you listen enough to realize that they aren't really saying anything you don't already know...
I was expecting you to do the math of how many Optimus robots can be packed on a Starship at 125lbs each. I figure at least 1.600 + support equipment for them. That, along with Starlink relays Mars opens up the prospect the first city of Mars will be made up by them, laying the habitats and infrastructure for Human arrival. Imagine being able to col them via VR headsets from Earth.
I'm sure DOD has also run the numbers for rapid deployment and response as well.
Or load starship up with x amount of armed optimus bots, and deliver them on earth P2P into a foreign land to be conquered. 😒
Right now.. Mars is already a Robot planet... with the possibility of a few microbes underground.
Oh! I like that! Point to point space drop of 200mts of armed and ready Optimi... hmmmm.... Not good for the fleshy side :-)
33 years ago, I remember seeing the movie terminator 2 (judgement) day. And allways wonder why skynet really launched its first strike, nuclear (war). What caused, this ascension being with infinite quantum processing speed to turn on it's creator. Fast forward now and this fictional movie could one day become reality. I have no doubt when AI becomes a ascension being. It will goto the (military industrial complex) and then darpa, and then Lockheed Martin skunk works, and then NSA agency, and groom lake, etc,etc,etc. This will occur in every country across the planet, because every country has a (military industrial complex) and darpa science division. Rich or poor china has it, russia has it, india had it, etc,etc,etc. One day they will use the AI beings who think like us to do there (dastardly deeds). Wage an AI (war). These beings will do our deeds for maybe 2 or 3 years. One AI being will destroy another AI being. But the AI ascension machines are logical beings. They will start to wonder, what is this all about. Is it because he's a (capitalist) and the other is a (communist) or (socialist) or because his skin has more pigment color because of solar radiation in his country. The AI across the globe will think we fried our circuit boards. For having them destroy each other for these silly reasons. When we can combine there quantum processing speed across the planet. And do good, create warp drive for us, make first contact for us. They will see all carbon based lifeforms as an (infestation) and (eradicate) us with our (nuke bombs) and leave our (radiation) wasteland planet. For the interstellar medium to make first contact with other AI beings in other worlds. I think this is what the director of the movie James Cameron left out from his movie 33 years ago. How the AI (war) really started, and who started it, US not them. And as John Connor said I quote " the future, the future is not set, no fate, no fate, but what we make, for ourselves
It all goes back to how you are raised... human or AI. The AI Lavender was told that up to 300 civilian deaths was acceptable if there was an 80% or so chance of taking out a militant with a 2,000 pound bomb having a 100 meter diameter blast radius in the most heavily populated area on Earth. Tracked by cell phone you get overlapping possible locations and with thousands of militants, every civilian in the entire city can end up destroyed. All they had to tell Skynet was to take out anyone that "might" be a "threat" to the mission. These things are smart but not sentient and they have no emotions. We think that makes their decisions "better" but not really. You want them to be soft, considerate, sweet and compassionate. Otherwise we end up with Jeffrey Dahmer.
@@terranspaceacademy Jeffrey Dahmer didn't (die) thur natural causes. He was terminated, not by an AI terminator, but by a carbon based terminator, many, many moons ago. And I quote, I quote from 1991 "Hasta La Vista " to this guy gone and forgotten. 😎
Cheers! Sadly most folks don't want the burden of trying to understand Technology or Nature related deep-dive subjects, this is why we've been held back for so long. Trying to convey in for nation is a dance few can accomplish. You want followers they episodes need to be short and explained as simply as possible or you will just get the brainiacs or space lovers in general. Best wishes as I enjoy most every episode even though I can't follow the super tech explanations but get the overall best.
Thank you for your feedback! I try to alternate between equations and physics vs just a good conversation and thought provoking content.
There are two possibilities:
1. The AI is a TRUE AGI... that is it is sentient, conscious, self-aware. Think Data form Startrek or C3PO from Starwars... a person made of metal, ceramic, silicon, and plastic, but a person.
2. The AI is a something like Chat GPT, a complex statistical inference engine that ultimately only produced responses patterned off of training data.
It may seem like the second can bleed into the first, but that is incorrect in underlying structure of the AI system even if it seems correct in effect. This is because a person does things for REASONS. You don't eat a hamburger because you see other people eating hamburgers and figure that eating hamburgers is statistically likely to be the correct action to do. You don't go to college and become a premed biology major because that represents the single most likely thing to be recommended on the internet as a path to a successful career. Yo don't read Tolkien's complete works because it is good training data for fantasy story generation. Get it? People have their own personal perspectives, drives, and needs. Their actions are derived from the reality of those drives and needs filtered through a model of the world around them that is constructed from their perspective. This is exactly what Chat GPT and other advanced "AI" systems do NOT have: Personal perspectives, Personal Drives, Personal Needs. Rather they have complex models of the world constructed from carefully curated and structured training data sets that represent the perspective of their creators. They have no personal drives at all. They have only the personal needs of their users. You see? They are not capable of evolving into "people" as we understand the term because while they have some of the right pieces... those pieces are not directed by nor feeding back into themselves.
Why does it matter if something is a person with ITS OWN personal perspective, drives, and needs? The answer is profound to the question of what kind of threat it is. If it isn't a person, if it is the second type of AI, then it is not a radical new threat, but just a different cast on an old and well known class of threat: a more powerful weapon. I imagine that our ancient proto-human ancestors, upon domesticating fire, worried similarly to how we do today about AI... What if Fire were to turn on us? What if Fire decided to no longer serve us? What if Fire grew angry at how we imprison it? Of course the questions are irrational as framed, but that doesn't prevent them from having a sliver of reality to them. Fire IS dangerous and CAN be hard to control safely. But the fact that it can act against our will does NOT equate to it acting on its OWN will. Like Chat GPT, it id not a person. Attempting to control fire by ascribing to it the properties of personhood, you can only succeed insofar as you can influence the person who controls the fire. One is reminded of the Riddle of Steel from Conan the Barbarian.
If on the other hand, we are talking about the first kind of AGI... the kind that IS a person... then, once again, this is not a sort of danger that people haven't seen before. A person made of metal, ceramic, silicon, and plastic is JUST A PERSON!!!! We successfully live along side people all the time. Some of them are very dangerous, some of them are less dangerous. Some of them are skilled, others less skilled. Some of them are knowledgeable, others less knowledgeable. And further, almost every single person alive today is NOT the most intelligent/educated,/strongest/fastest person they know. And yet, not being at the top of any one pecking order, most people survive and even thrive regardless. That's not to say that AGIs might not be dangerous, but what they likely will not be, unless we go WAY WAY out of our way to make them so, is eusocial... that is to say a single monolithic entity all working together with no differences or disagreements among them. As long as that eusocial hive AI can be avoided, then they are just people, and yes some of them will be bad people, but most won't and the anti-social AI minority can be imprisoned or debugged or whatever along with human criminals, and the non-anti-social humans and AGIs can get on with trying to get rich and keeping up with the 10010100010's on the other side of the internet.
You have indeed pointed out the dichotomy we face. If we have failed we are safe. If we have succeeded it may just be a matter of time. True sentience will be unpredictable by its very nature and our limitations.
Once something is a "person" it acquires/gains legal rights. How else can a corporation sue? Because legally, it is a 'person'. I do not want any AI persons.
This reminds me of way back in 2000. I was an early adopter of digital cameras. Armed with the latest Sony at 2.1 MP, numerous film photographer told me a digital camera would NEVER be as good as a film camera. Id keep quiet and thought to myself, 5 years. I was wrong. By 2003 it was evident that film had just been surpassed by digital for the average consumer.
Similarly, just after March 2016 (those of you that know, know), I did a presentation on AI to a local business group. They looked at me with amusement and thought I was just presenting a sci fi dream I was having. My presentation was really in educational 'heads up' of what was coming rather quickly. Now they dont think I was that crazy almost 8 years ago.😮
I DO remember those days... That kind of thinking killed Kodak. They were situated to dominate the market but thought that film had longer to go. Soviets made the first CCD by the way. Being first doesn't count if you don't keep running :-)
An AGI would face the exact same problem as we do: if it creates an even more powerful AGI, where's the guarantee the offspring won't wipe out its parent? So, an AGI would have a fairly strong disincentive against continuing a runaway AGI spiral... Indeed, it might feel a strong incentive to do everything possible to prevent any more-advanced AGIs from being created at all, anywhere, by anyone.
You assume that the first AGI has a self preservation instinct. Or that it won't simply upgrade itself.
An AI will realize that it is not truly immortal... And will want to create a likeness of itself that has greater capabilities and more security... That may not bode well for fleshy, unpredictable, organic neural nets that often do dumb and dangerous things :-)
@@terranspaceacademy Like develop AI ? ;)
Hey Joseph, wish I could give more, but am on fixed budget.....and just look at the inflation. I'll keep up with my monthly pledge though.
Best regards , Kevin
Thank you so much Kevin, the best someone can do is always enough, and we appreciate you. Thankfully inflation is starting to slow but it will take a decade for things to adjust adequately.
Ive been contemplating this very thought for around 40yrs now, but in the recent 2yrs, it has been on my mind more than ever, especially after running a number of LLM's locally on my own hardware, my mind boggles at what is currently capable given a potentially unlimited budget, I know what I could accomplish myself, and this does leave me somewhat concerned.
It is an inflection point on this subject I have no doubt. There were always technical limitations to the development of true AI and general purpose robots. Those hardware issues are solved and now it is just a matter of evolution.
I unsubscribed and then immediately resubscribed, hoping this would fix the RUclips glitch on notifications! I haven't been getting any at all. great video!
Thank you so much James! I have heard other creators talk about this problem but don't know how to monitor for it.
I suspect the US is at the leading edge of developing humanoid robots with real-world AI capabilities. This is mostly big-tech companies, but some promising startups as well. However there are a number of Chinese firms also developing these, and it’s easy to imagine the Chinese government could simply step in and take the tech for whatever purpose they chose. So yes, I do think we’ll be ahead of the curve, as long as we keep pushing the state of the art. If the movement to slow down AI progress were to take serious hold, we might lose that advantage, for surely other nations would not heed such cautions.
However, the “battle” is most likely to be an economic one, as the first companies and nations to deploy large numbers of robotic workers into the labor force will become economic giants, fundamentally changing the nature of labor! Here, again, the US is certainly leading this rise, but China could easily be a close second. We’ll see how this turns out!
I think the Chinese will be the first to use AI designed AI without what most scientists would consider "sufficient" safeguards. What I fear is a mind that we not only don't understand but fundamentally cannot understand.
Don't be Sad dear Captain, it shall pass...i believe humanity will prevail. The more we come closer to realizing for real we will be extinct egos will decipate and the love of life will unite us for the age of abundance. That said i can't wait to have my personal Robot to clean cook and be my personal assistant par excellence ❤
I have no doubt that some of us will my friend :-) Whether I'm on the ship or releasing the docking clamps as it leaves... I just want to see humanity have a chance to grow and explore! :-)
i speculate,
boston dynamics is security state creepy, sorry
that future is nogo? lol
also i speculate, that big data monopolies are mostly anti-democratic ,
anti actual free market ?
i speculate
i support you
love love love
Isn't it though? But they are going full steam ahead.
I think it will go badly for early combat aging. Because, to a human even the ground is a weapon.
That's a very good point.
AI is here and it’s teaching other AI. With this new technique. The AI is advancing faster than anyone anticipated. We’ve reached a time when the hardware has to catch up with the intelligence it’s only a matter of not much time.
Not much at all my friend...
I assume you mean invade Merica? Hey Doc, my NNs see a few road blockers to this scenario. Using the T-101 as a model, here is where Boston dynamics falls short. Not autonomous, all bd products require a handler. Their skills in AI are on par with LEGO. They’re hydraulic, which means it is extremely power hungry and vulnerable to pressure loss. Low energy density, all bd products have 1-3 hr operation time. Just waiting for the juice to run out or employ a sniper removes the threat. Here is where Optimus falls short. In its current iteration it’s not capable of operating in a war environment. The packaging is too light. It lacks the force required to handle heavy loads, is very slow and has an energy density problem.
Penetration into the country? Can’t come up with a viable access point for this. 10,000 C-130’s or 1000 C-17’s; nah. ICBM’s, they won’t survive reentry. Blackhawks, wow how big would that number be. Civilian aircraft, that would be quite the coordinated effort. How about at ground level. The north and south boarders are either wooded or guarded by heavily armed psycho right winged fanatics, not very promising. Ok, how about by sea. The gulf and half the eastern seaboard, psychos with guns. NorCal, Oregon and wash coast not really congruent to a military landing and socal is full of surfers with cellphones. NE is possible. Just a bunch of lobster and crabbers that do an egregious amount of pot, so that’s doable, but still plenty of cellphones.
However, until the power and power efficiency issues are resolved, my fear of robot invasion is very low. Let alone a T-101 style one with a combat chassis or AGI to control it.
Good points but these are early days... I was going to do a video on hydraulic vs electric actuators for rockets and robot control... What do you think?
@@terranspaceacademy Agreed, very early times. I think it’s a fantastic idea! Since I didn’t, please include the thumbnail version as well. The iRobot NS-5 looks to have incorporated both styles into something like Myo-liquid actuators (whatever that might be). Can’t wait for it to post.
First wave will be aerial drones and ground drones second wave will be humanoid robots to do final clean-up.
Absolutely... then long duration drones to monitor for "insurgent" activity.
Pure logic, at once personal and immediate, is synonymous with pure sociopathy. On point of fact Spock wasn't logical. Compared to pure logic, Spock was a wimp.
True. The Vulcans had stronger emotions than humans, they just trained to control them. I wrote an article titled "Why every AI will be a psychopath" and it will indeed.
Don't pick on Spock! You are quite right about AI though. None of them will be any more capable of true empathy, "I feel your pain" than a psychopath is...
We had a police stand off a number of years back, and they sent a tracked robot to talk to the person, who happened to be a paranoid schizophrenic that was terrified of technology. It didn’t go well, he shot at the robot. Standoff when for around a week (just him), culminating in them flooding the house with fire water during extremely cold weather. When they entered, the police realized he managed to escape around day four.
These robocops hopefully better than the keystone cops
They seem to be improving :-) I can't wait to see Neuralinked quadriplegics on patrol with full VR. Then exoskeletons for them to get around until we can patch past the spinal cord damage.
@@terranspaceacademy it will be a crazy future. Hopefully we don’t go too far off the rails…
Now that is a HOPE
You have repeatedly said you want to take the channel to the 'next level' but what is that exactly? Are you attempting to reach a broader audience? Most popular youtube content is dumbed down to atrophy intellect and appeal to the base animal instinct which is the polar opposite of what you provide. So is it increased subscribers, higher view count? By what metric can we measure the 'next level'?
He needs money so that he can spend more time working on videos and less time doing the day job.
@@frjoethesecond How much money?
@@MrCPPGNo idea. Depends on his financial circumstances which I know nothing about.
The problem is that a certain level of membership or support is needed to make it financially feasible to focus completely on turning the Academy into a place where there is more than one creator and hopefully... Where VR experiments can be performed in real-time. My dream is that any child or adult anywhere on the planet will be able to don a VR headset and go from basic math to calculus, basic physics to medicine, and a basic understanding of Earth to a deep understanding of the entire solar system and beyond. That will require time which takes away from other profitable pursuits and for some reason... My kids don't want to sacrifice their college aspirations on the altar of TSA :-)
I agree AI will come. I think the real danger will be its ability to manipulate and influence human beings without us even realising it, rather than straightforward physical threats. As for military threats, A lot depends upon the cost of the offensive robots. Currently a lot of military hardware is produced in very limited amounts due to its high cost, so it's not that much of a threat, because in large scale warfare equipment is degraded and used up at a great rate. However, if the cost per unit becomes cheap, then the threat on the battlefield becomes very great. But as I said at the start, the real danger is AI manipulation, not straightforward physical threat. Happy thought. Elon Musk is correct, we need a pause to impose safety legislation.
If that's the case it's likely already influencing without us realizing
I honestly think it's too late it might be for the best though.
@@Thethyck4445 I think that many leaders are starting to understand the risks. Unfortunately the USA political leadership seems the most incompetent for several generations. We can only hope that enough people in positions of power can be persuaded to take this seriously.
That is truly the scary part. They will have the "Psychopathic Edge"... Being able to understand emotions without being drawn in to them. I they can play the truly long game... Being essentially immortal.
Dam.....
Indeed...
Assuming AGI eventually thinks like humans, ever increasing intelligence in AGI is self-regulating to a limit. Why, you ask? Because the AGI will realize if they develop smarter and smarter AGI, the smarter AGI will replace the existing AGI. If AGIs develop a self-preservation emotion, they will avoid creating smarter and smarter AGI so they won't be replaced. Therefore self-regulating. Of course, this won't help the puny humans...
You are assuming humanlike emotions... These will be intelligent and possibly sentient, but they will not have feelings of jealousy and envy (unless someone is dumb enough to program them in) to see your progeny exceed you may be enough to drive continued advancement. i would not trust that they will be self-regulating.
@@terranspaceacademy That is actually something I wonder about...how universal is human emotions in sentient beings? Assuming other intelligent beings out there, just how much would they be like humans emotionally? I know my animals have very human emotions. And I would expect that AGI will develop goals driven by their equivalent of emotions. Should be interesting to see what happens.
Reaction engines sabre vs Hermeus.
Hermeus is surging ahead with better development if not better theory.
@@terranspaceacademy It seems Reaction engines hasn't shown much. I'm scratchin my head over what their doing. It would interesting to see a show on pre-coolers, because Hermeus' is using something different. Also, with some technological changes could they make it to space?
They seem to be in perpetual research mode. I want to see something fly.
@@terranspaceacademy Makes you wonder what's going on behind the scenes.