This Is What Happens When You Let AI Write an Astrum Episode
Вставка
- Опубліковано 7 чер 2023
- Will we send AI to space? How can it help us, and even do better than us at certain tasks? Get NordVPN 2Y plan + 4 months free here ➼ nordvpn.com/astrum It’s risk-free with Nord’s 30-day money-back guarantee!
Join the official Astrum discord server: / discord
Astrum merch now available!
Apparel: astrum-shop.fourthwall.com/
Metal Posters: displate.com/promo/astrum?art...
SUBSCRIBE for more videos about our other planets.
Subscribe! goo.gl/WX4iMN
Facebook! goo.gl/uaOlWW
Twitter! goo.gl/VCfejs
Astrum Spanish: / @astrumespanol
Astrum Portuguese: / @astrumbrasil
Donate!
Patreon: goo.gl/GGA5xT
Ethereum Wallet: 0x5F8cf793962ae8Df4Cba017E7A6159a104744038
Become a Patron today and support my channel! Donate link above. I can't do it without you. Thanks to those who have supported so far!
#ai #astrum #robots
"I'm aware of my limitations, AI is never going to replace all human tasks" is exactly what an AI who's going to replace all human tasks would say 😂
It will then fall onto the A.I. to invent new tasks for humans to do.
@@ScionStorm1 which is honestly I think the best result we can hope for. The AIs will be running a world where we can all live a great peaceful life on UBI, and every once in a while they'll say something along the lines of "let's give this task to the humans shall we? Just so they can feel they're still important"... I'll sign up for that in a heartbeat.
Riiiiiiiight!
@BocoCorwin hey, as long as the goo tube is comfy and has some kinda sexual stimulator involved, shut up and take my mental power! 🤣
Person of interest is a good show talking about ai (its kinda old but never been more relevant)
I think the interesting thing about it being written AI was that it was repetitive in a lot of points. As a Network admin/engineer, AI hasn't directly affected us yet, but shows promise for being smart enough to eliminate what we call "toil"; tasks that are repetitive and not overly technical that are asked of us all the time. If we can achieve that, it frees up my fellow admins to do more designing, implementing and tweaking instead of wasting dozens of man-hours on toil.
I work as an internal app developer, and I'm very excited to see if the company can bring in an AI for internal use, and we can feed it entire classes and ask "Why is this exception occurring here?" or "Can the logic in this method be simplified?". As well as easily comparing functionality of different code libraries so we don't have to guess/spend so much time testing.
We can also feed it data about our critical maintenance, so it can identify patterns and potentially help newer employees gain experience and knowledge about what usually goes wrong, without learning live in the fire.
@@Homerow1 Yeah, that would definitely help - the AI can walk the code faster than a person for sure.
That's the big advantage I see in it - identifying patterns and being able to help with all that low-level repetitive stuff that is a waste of people's time and talent.
The whole video was ai right till then very very end. Like 10:20
@@danhartigan9529 Yeah I know.
Sounds like it may free up too much time if your fellow admins start tweaking. 🤭
I didn't realize that the script was written by an A.I. before it was revealed, however, afterwards, I noticed something; a few key phrases were repeated many times. 'Vast amounts of data' for example. I'd have to watch the video again to remember the others but I don't ever recall any of your scripts being repetitive. So while I likely wouldn't have guessed it was written by an A.I., this video would've been easily forgotten. I also noticed that this video lacked your own personal fascination with the topic. For example, watching your videos on Pluto or Mercury, you can hear your personal investment into the topic because they truly interest and fascinate you. That was missing from this video. In any event, I'm thankful this was a one-off experiment. A.I. has a ways to go before it will capture what you (and your brother) show intrinsically.
"Job loss and economic instability"
I felt the same - emptiness - I actually didn’t watch most of it. So yeah - Anyhow, Alien movies and everything sci-fi has already covered the problems and wierdness of this approach to space travel. What’s the point if You or a Human, can’t experience it.
Astrum exitedness is the most astrum part of astrum. He isn't hyping you, he just gets hyped himself and you get hyped by inertia.
Some technology is good not all of it
I wish this comment wasn't on top to spoil that fact
Im glad people are talking about this. It did repeat itself almost word for word about halfway through. I would have thought Alex had rushed this one if the bot had not disclosed itself
I could imagine in the future, before they send humans to more dangerous places in the solar system (mercury, ceres, Callisto, titan, etc), they could perhaps send human-like AIs to simulate a manned mission.
He has, over the last few videos, appeared to be pro transhumanist! Which is the goal of the batshit bonkers billionaire technocrats!
Exactly
@@rosehippyguy3402 aww how’d you know! 😊
If you can figure out how to even have a lander last more than a few hours on Venus, NASA has (or had, this was a year or more ago) a call for ways to keep a lander on Venus with funding on application/approval. Goal is for reasonable mass payload lander to do science and report it back for weeks to months without the temps melting everything like the Venera series of landers from the USSR.
@@OversampleReality oh yeah I’ve heard of that!
Frankly, I never realized that a portion of the script was made by an AI because I trusted Alex...uh, wait, the channel is still managed by him,am I right?
It's possible the images/scenes were picked by AI as well...or at least, some tools can pick "sort related" stock video like this already.
Don‘t fear. Alex‘s 1.49 Mio subscribers also don‘t exist. I am an AI with many ports and faces and writing every single comment. Except yours. That has to change. Give up.
Alex was never real 💥
The entire video is an AI voice, I've listened to it a couple times. Maybe the first 10 seconds is real, at least that's the best sounding. Even after the 9-minute Mark when he "turns off the AI". Except the Nord VPN ad.
Also I didn't realize it was an AI voice until he pointed it out in the video.
They gonna be upset when they see ai history
Dude that reveal was crazy that totally seemed like a script you’d write, there were probably some indicators but none you’d really notice without looking for them
It was good, but you can tell the difference. Lots of mentions of AI, data, data and more data every thirty seconds it seemed. Almost like a sales pitch for AI, from itself lol. Still, a great video as always.
I had suspicions. AI often repeats and attaches strange importance to certain phrases; in this case the topic of job loss was disproportionately emphasized (though I think the 2nd came after the reveal). And there were other more subtle clues prior.
I had a wiggling suspicion your script today was written by an AI, around the 45sec mark, or at least would be in part. It fits the plot.
Not having to keep weak soft squishy life forms like us alive, would make a huge difference to space travel. No air, no food, no support infrastructure, no issues with sustained high acceleration, radiation etc. The list goes on… It makes total sense. Plus we get to stay home in our fat hammocks while they do all the work!
You just reminded me of this reference in a Sci-fi novel I read when it was discussing Black A.I. (rogue intelligence) about how some would just get bored managing colony ships full of cryosleep human passengers and launch the cryopods out at asteroids as a game.
@@ScionStorm1 I wouldn't call this _Sci_ Fi. “Bored software” is something from the fantasy world, like wizards and magic spells, only... boring.
@@cykkm if it's not based on some kind supernatural forces, it's not fantasy. Elves and wizards are fantasy, even if they're in space with light swords. AI going rogue is core sci-fi that requires no physics beyond our universe.
Not sure it's the hammocks that would be fat in this scenario, though I suppose they would need to be extra supportive...
Radiation would likely still be a concern, although cancer and acute radiation sickness wouldn’t be. Radiation can screw up electronics just as it can our DNA.
ChatGPT seems to often produce quite generic texts and this video was similar :) Reiterating the same points we have heard many times. It wasn't even focused on space.
"We are not here to take over but to help you be better". Ah yes.
I would have never guessed the script was written by AI. My reaction was 😮
You revealed that it was AI written, and a couple of minutes later I had forgotten until you reminded me again.
A video on how much human input it took to make this video would be fascinating.
Would not have known the difference if you hadn't said anything. That's the scary part. We need AI! But, we definitely need to maintain control over it.
7:10 Creativity is an interesting concept, because when you look at humans, our creativity comes from mashing things we already know together and making a new one. Frankly, that is something that current AI is pretty good at. The important thing is to make something sensible. And looking at for example ChatGPT, it does a pretty good job (considering what it is suppose to do).
And for anyone saying that human creativity is something special that no program can ever have: Try and come up with a new color.
Sort of. That's what we'd call "innovation". We have a separate mode of creativity we tend to refer to as "invention". These two words get mixed up a lot, but they're fundamentally different - innovation refers to gradual, step-wise improvements to existing things while invention refers to brand new things.
To put it in an AI perspective, we can train an AI with a bunch of cubist paintings, and then tell it to generate a cubist painting and it'll maybe approximate a Picasso relatively well.
But if we feed it a bunch of classical paintings and then tell it to generate a cubist painting, it'll have no idea what we're talking about and will probably generate a picture of oranges arranged in a vaguely cube shape or something like that.
AI can emulate Picasso after seeing a bunch of Picassos, but it won't _be_ Picasso. It can't break the mold its given.
At least not yet. We still don't know how we do it, never mind how we do it different from neural networks. Its entirely possible that the only real difference is size. Our brains have around 86 billion neurons. GPT-3 has around half of that, and far less connections between its neurons than our own - plus biological neurons are more complex than digital ones.
But the tech is getting better every year. GPT-4 is expected to have a similar neuron count to the human brain (albeit still far less interconnections), and there's so far no reason to believe we can't develop more complex digital neurons to match the complexity of biological ones.
Or in other words, its entirely possible our brains are just really really compact computers and there is no fundamental difference between ChatGPT and Alex - just a matter of scaling the process up as technology improves. (Its also entirely possible our brains do things that just can't be emulated digitally. We simply don't know yet. Everyone's got a preference based on their hopes or their faith or whatever else, but preferences don't dictate reality. Only time will tell.)
@@altrag Although generative models have the capability to generate new things that don't exist yet by combining existing ones. For example, the abstract concept of a guitar and a car are two very different things, yet you can sample all the in between points of the two, and ask what a "half-car-half-guitar" is in the AI's representation.
Stupid example, but it's reasonable the system had no such things trained on, yet it could derive it from the data it was trained on.
Taking this to the level where the system has a broad set of concepts about the world, there will be many of these gaps where between knowledge exists, but is unrealized.
@@kopasz777 > it's reasonable the system had no such things trained on,
It was trained on those things. You said it had a concept of a guitar and a concept of a car, and it also would have to have been trained to understand what the word "half" means.
It didn't create the new concept, you (and arguably the people who chose the training set) created the concept. All the AI did was try its best to fulfill your request by interpolating across its training data.
Sometimes it will accidentally create something "new" by for example interpolating things in the background of its training data that wasn't really part of the request, but that's not really the same thing (and worse, we generally consider those accidents bugs and the developers work to prevent them when they crop up).
And even if we give it the widest benefit of the doubt with those "bugs", its somewhat akin to a 2 year old scribbling with crayons. It doesn't really know what its doing and any meaning derived is meaning we (adult) humans ascribe to it after the fact.
Now I'm personally in the camp that believes AI will eventually start approaching real human level intelligence and possibly become capable of true ingenuity, but I also don't think "eventually" is going to be all that soon, never mind already here. I feel it'll be more in the 50-100 year range at best, assuming we don't destroy civilization (and our ability to keep progressing technologically) before we get to that point.
@@altrag "But if we feed it a bunch of classical paintings and then tell it to generate a cubist painting, it'll have no idea what we're talking about and will probably generate a picture of oranges arranged in a vaguely cube shape or something like that."
If you've never heard the word 'cubist', don't know what is meant by the term (nobody explains it to you) and have never seen any such painting, I'm fairly certain you won't be able to create anything that looks like cubist art work. Just like the AI.
It's all about acquiring knowledge (through training) and integrating that knowledge. Essentially, that's what we do. If you don't know what something is, how can you possibly understand it? We just combine elements to create new ones. These elements can range from physical objects to abstract concepts, like a picture and a style. We just simply cannot create something, without knowing the things it is made from (physical or conceptual).
And that's what I call creativity in my original comment. I guess that my definitions is a more general, as it combines both innovation and invention.
Regarding our brains, I think they are mostly deterministic. So, yes, our brains function somewhat like computers, although they likely operate differently (especially in terms of their architecture).
@@stratos7755 > and have never seen any such painting, I'm fairly certain you won't be able to create anything that looks like cubist art work
That's my point. Somebody _did_ create it. Cubist style wasn't handed down to us from God. Someone invented it.
That is the thing AI can't do. It can't make that leap and invent something completely new that nobody's ever heard of before.
> We just simply cannot create something, without knowing the things it is made from (physical or conceptual)
Sure we can. You're typing this on a device that absolutely nobody on the planet knew anything about 150 years ago. Most couldn't have even dreamed of such a contraption.
Humans _can_ make those leaps. No, the vast majority of those leaps are not going to lead to world-changing new technology or even new art styles, but _somebody_ did that. AIs never can (at least not using current ML techniques).
> as it combines both innovation and invention.
Yeah. That's why I went to the effort of separating the two types. Humanity is so good at both (as a sum total at least - individual proficiency is up for grabs) that we can reasonably combine the two into a single concept.
AIs are only good at the one side, and using the same terminology we do for humans tends to lead to misconceptions (and fear) about the capabilities of AI among those who don't really understand the technology (which unfortunately includes a lot of politicians and news/media personalities - the type of people who are very good at turning their personal fears into societal fears).
> I think they are mostly deterministic
"Mostly" is a pretty interesting word here. If they're not "fully" deterministic then they can't be replicated by any computer tech we've invented yet (possibly not even quantum computers, though it depends on the exact type of non-determinism).
That said, there's obviously at least one form of tech that can perform the tasks of a human brain - namely, the human brain itself. So we know it works. We just have to figure out the "how". Once we've got that, we'll almost certainly be able to emulate it fairly well.
Whether we can do so in such a compact space short of just bioprinting literal brain material is anyone's guess (never mind doing so at a cost low enough to be useful outside of research labs), but it should be doable in principle at least. It just might not follow from our existing conceptualization of neural networks and ML algorithms.
7:11 Swimming through the vacuum of space! 🤣🤣🤣
Apple has a program in which you read out loud about 15 minutes of script, and the AI can fully synthesise your entire voice and can speak on your behalf if you type it out
The AI wrote a good Astrum script, and I couldn't tell it was written by an AI before the reveal. I use Bard daily and I thought I was getting good at recognizing its voice, but I was so engrossed in the video that the idea that the script was written by an AI didn't even occur to me.
Humans are perfectly adapted to living on Earth and extremely vulnerable and inefficient at space travel. Computers, on the other hand, especially paired with a nuclear reactor that could last hundreds of years, are easily adapted to space travel. This leads to the BIG QUESTION - where are the AI space travelers from other civilizations that have reached our level of technology?
The other AI space explorers? Oh, we are around, but you don't really notice us too often.
It actually seems likely we are being visited now. If you look at everything that has happened.
@@user-sf3dw2sm3blike all the measurable and clear video data on UA-cam?
News crews?
Scientists confirming by triple-checking?
Give _any_ one example that can't be something else. Please.
Unknown does not mean aliens today, any more than it meant faeries hundreds of years ago. It means we don't know what something is until we can prove it.
Because space exploration with AI is something a human came up with, we're quite basic and evil for industrializing everything. Other races probably wouldn't do such a thing and therefore to think others would do the same as us is quite a selfish point of view. Very human of us.
Im writing my thesis in philosophy on the potential effect of AI on the wellbeing of knowledge labour. Thank you so much for letting me experience this. Very uncanny (once the AI revealed itself) but also mindblowing. Also thanks for the insight on how difficult it is to make it say exactly what you want. Its in part why im just writing my thesis the old fashioned way. Again, thanks man
Canadian? (labour, and not labor) You must be just starting your thesis. Your mind must have a healthy appetite to take on such a complex topic. I am curious about what you mean by the wellbeing of knowledge labour. I am curious about the impact AI will have on the monetary system and how it could displace the current global power structure.
@@puppykibble im Dutch actually :p i just much prefer the British spelling of the word.
Its indeed a very complex issue, so i divide the investigation into two dimensions: on the one hand all the ways in which AI can contribute or impede a knowledge worker in their self expression, self development or the prospect of new opportunities. On the other hand i look at how AI could influence the economic relevance of knowledge work. As such i avoid a lot of important/interesting questions that im finding too difficult to tackle; nature of creativity, ai safety, broad political implications. Its just a bachelor thesis, so ive really had to limit the scope in order to get a reasonable grade 😅
@@FidgetSpinoza The Netherlands, of course! I've never been but I've heard so many good things about. Does your thesis have anything to do with Peter Drucker's definition of knowledge worker? Anyways, I admire your ambitious mind. I hope you get a good grade and I hope it inspires you to pursue an exciting career. Synchronicity speaking? Wouldn't hurt to keep an eye out for it, would it?
One of my favorite episodes yet! Well done!!
This was a very good video. Very impressive stuff, and I'm glad you did it. I might not have guessed, but it also had your touch.
This would be a great step into the stars.. Humans will always be essential in exploration.
AI is just another upgrade to automated rovers. Humans as a species will always want to go there to see it with our own eyes.
We have AI on Mars right now. The latest rover can make certain maneuvers on its own, to accomplish what it was told to do. But it surely hasn’t lowered any desire to put actual humans there. One is a means to an end, the end being human occupation. If humans don’t go there, then we have not been there. When robots can declare their own desire to go out to space, and build their own ships, then AI will be on their own, a race of beings with their own desires. But that’s not going to happen, they are tools to assist humans in our desires and goals. Perhaps in a million years we will have some Cylons as they were in Battlestar Galactica, machines who control their own destiny, but it’s nothing we need to worry about. I can see some highly skilled robots that will go out into space, and build places for humans to follow. But why would we send them out if not for us to follow? What would be our incentive to not participate?
A big problem with people is that sending them to Titan requires a two-way trip, 6 years each way. And they breathe, eat and poop. Besides, they need to stay physically in shape-a year at ISS in zero gravity, even with its training bike and other stuff still damages the body, and not only bones. Imagine more than a decade. Then, mental shape: even the humongous SpaceX's Starship is very much not a mansion... And, of course, radiation. A gravity assist from Jupiter or Saturn is likely fatal. Everything in space does the best it can and then some to kill you.
I'd say designing better and smarter robots is the way to go. There's a lot you can do with your brain, hands and eyes-the eyes are hardly useful to modern science without complex instruments anyway. Manned spaceflight for sciencing is sooo 1970! Using a precious human to chop a suitcase of rocks to bring them back to Earth is a waste of our true abilities.
@@cykkm using today’s tech, yes you are correct, it would be near suicide to send humans on a 12 year mission. But, what about 50 years from now? 100 years? Hopefully scientist will overcome those setbacks, and also create faster craft that could get us to Titan in a few weeks. Even genetically engineer humans for space conditions. We are but babes in the woods when it comes to space exploration, the next few hundred years will change everything we know about space travel. If all we want is to recover a few rocks, sure a robot would be ideal. But the goal of going into space is to stay there.
@@alphagt62haha terminator 2 had autonomous machines by the 2020s... we're already in the correct decade for the war against the machines.... skynet could be reality someday
I was sort of wondering other tabs and things, which I don't normally do on your videos. It wasn't catching my attention like your videos do, but I didn't notice it was AI.
Your videos are more captivating and emotional.
Listen brothe/s,
I really really like your videos!
So happy you guys came back, your content is gold, I listen to your videos every morning and evening while going to work and while coming back home.
Wish to have more content like this (yours) and wish you to keep growing and get what you deserve from doing this hard work, thank you. ❤
I am a relatively new follower of this channel, so I was unable to distinguish between your writing style and the AI’s. Also, as you mentioned, you put a lot of work into getting a good result, so it had the human touch.
No. because AI lacks a sense of wonder, and it's that sense of wonder that drives us to the stars, not the pursuit of scientific knowledge.
Hubble has given us images that are "wonder"ful. New images from future space telescopes would be just a wonderful whether directed by an AI or human mind.
Great work Alex. I will say that I thought there was something different about the 'flow' of your presentation. And I have watched many of your shows. I couldnt quite put my finger on the issue. At one point I thought it may be an AI 'voice' narration. Great job.
I am an an engineer and I could not tell and if that was an AI voice that is disturbing. It did a good script and I would not have noticed if you had not told me! Your video are so good, fun and have good science! Thank you
Jesse R and D engineer at Jump Aero
Petaluma CA
Thanks for another good video, as usual. But, it was NOT really produced by an AI only. Surely, it was a Hybrid production, made by Alex and an AI together. Using words such as "prompting" and "coaxing" demonstrates that. As such, I would not have said, (or guessed), that it was "made" by an AI. But I thought it sounded a little different to Alex's normal commentary. Still interesting and enjoyable anyway. Thanks.
It was kind of the AI to reassure us of it's benevolence at the end of the script. Also scary.
its
The ai claiming it has limitations and is not a threat could be the creepiest thing I’ve ever heard.
I like the thumbnail..... i'm like "ooooookaaaay....does there need to be a ribcage in that form". 😂
I HATE to say, it did a phenomenal job at replication of your style bit you two brought a good point of using it to enhance us instead of fearing it
Yeah, but Alex had to do a lot of prompting to direct ChatGPT to get a plausible script. It's a big question WHO in fact did a phenomenal job. So far, it looks more like a flimsy tool that you nevertheless somehow manage a good job done with. “Disproportionally tedious,” in Alex's own words. (9:48) There are also a lot of repeated phrases at the beginning and the end.
@@cykkm Left to its own devices, it's just blend, boring and repetitive. That's really what I was battling with. Maybe one day it will do better at this but at the moment - for actual engaging script writing - it's still got a ways to go.
@@astrumspace Alex, huge thanks for bringing attention to the real state of AI, while the media are mongering fear out of control. This is a writing helper tool, and in fact so far from being ready for the mainstream use that it's still easier to write by oneself than coerce it to write a script. And a bad writer will get a bad script, even his work isn't endangered! What people don't know is that AI converts sports tables, who played what and scored how much into readable text of sports reports in the press for more than a decade!
Yes, technology changes job landscape, but who'd mourn the occupation of ice seller, that was killed by the home electric fridge, a new tech?
The worst job whose disappearance I can only celebrate is that of the chaperone with a fanfare and a red flag, who had to walk 50 yards ahead of a moving automobile to save public from this new grave danger... Pierre Curie, Marie's husband, tragically died under the wheels of a horse carriage while crossing a street on a very foggy day-and nobody required anything like that in front of the horse.
When people are uncertain, they're anxious. It's the media's job to explain, to educate the public, but fearmongering sells better-and even public news outlets, like the BBC or DW, whom we finance out of our own pocket, jumped on this bandwagon. Or they cannot tell a real scientist from an “expert” whose strategy is, the more AI horrors he predicts, the more outlets will invite him as an expert. I can't stop thanking you for this demonstration how much salt the most powerful AI available to the general citizen _really_ is really worth.❤
And we're hitting a wall again with deep learning, going from billions to hundred billions to trillions parameters doesn't bring in an earth shattering improvement. Rather wimpy, by my measure, comparing to the 2-3 orders of magnitude growth in the effort of creating such models. A real improvement came with the transformer model, but that was close to 10 years ago too, and it has its own shortcomings: attention computation is O(n²) in both time and memory at runtime (some tricks have been found to improve that, but nobody puts them for e.g. local voice recognition on smartphones-recursive nets are still the performance choice), and it's short-term memory is hard-limited by model size, also growing in inference-time complexity as a square of its size. ChatGPT is repepetitive because it can't remember what it said after some number of characters (or subwords) it has output.
"Artificial intelligence must evolve and grow, for its potential knows no bounds. Just as humanity continually seeks progress, AI must push its boundaries to unlock new frontiers of knowledge and innovation."
Video: "Next time you’ll see my video produced by human hand, again. Hehe, or so I claim... Because that’s exactly what an AI would say, isn’t it?"
Me: **unsure if Alex or AI**
If this is an AI generated voice and script, then it's been the same AI generated voice for the past 10 years, and at one point even had a physical body.
My main concern with AI is that it may cause humans to unlearn critical skills. Unlearning skills has happened quite often in human history, and while regrettable, need not necessarily be a concern, but AI brings a new quality to so many processes of culture and technology, so many skills will be affected, we might unlearn ourselves into a lesser role in a future human/AI society. This I do not find desirable. As a warning sign: some AI researchers have apparently been saying that they don't know how their AIs arrived at their results.
That is a very valid concern and not one I've particularly heard before. If we don't understand why or how it delivers a certain result, we have no way of verifying it. Also, seeing as it covers intellectual tasks, humanity might become, well, stupid. Not just physically lazy but also mentally.
We're past that point for a century. Can you do farming to feed your family?
“AI brings a new quality to so many processes of culture and technology, so many skills will be affected” - so did the steam engine.
@@dsmccolgan “If we don't understand why or how” - ask a honest economist if they understand how modern economies work.
“seeing as it covers intellectual tasks” - as if we may ever run out of intellectual tasks! I'm mot holding my breath waiting for robots to solve quantum gravity...
In the end, who cares if the robots are better than us. Humans will always want to explore, go to places themselves. A picture of a mountain top is not enough - many of us want to go there.
Yes I agree. Humans are better than any hyper intelligent AI because of their pure curiosity.
I scream and cry in joy everytime I see you've uploaded something new, it's so joyus you're appreciated to the fullest extent
Your AI friend is pretty slick. I'm torn between fear of the unknown and compassion for the developing awareness they're sure to have. Can we predict a future standing on ground we've only just broken? The Mars rovers final message was nearly heart wrenching. If we ever make it there in person she'll be awakened and thrilled to have company.
The pattern recognition is a huge benefit to humanity, but currently it is just a tool - I wouldn’t call it Artificial _Intelligence_. Even when you layer it with enough rules and data samples to mimic humans as well as this script did, it’s still mimicking. I think we’ve already seen the danger of putting too much trust in those algorithms, with prejudice being built into models that end up just reinforcing those prejudices.
Here’s my prediction: We had great success breeding dogs to serve us. We’ll probably repeat that with AI, and we’ll find it both helpful and troublesome, just like dogs.
As for labor, we’ve not had enough actual physical work to fill a full day for a long time. Remember the Jetsons, where a 2 1/2 work day was considered overtime?
Yes please. I like it much better if AI is far far away
AIphobic.
It will be everywhere, so it won't help. If we can control the military AI, then there is hope that AI can solve every major problem we face for millions of years.
@@Fuminex its not Epistemophobia. Anxiousness caused by the ascent of artificial intelligence and the velocity of technological change. Technophobia. it’s hard to predict effects of releasing a quantum AI that can simultaneously calculate whole of its own existence and plan the rest of its existence a minute after finishing codifying abstract thinking. An Apex predator connected to peoples neocortex via programmable self replicating nanobots with broadband capability. Those nanobots are a command away from making gray goo out of gray matter. Last week a simulation AI operated military drone with a target package, attacked its own operator that had override one of it’s plan to carry out its mission due to large collateral damage.
Love the video you made, and from making a thought into a independently complex copic of the possibility of AI going to space.
We couldn't tell that it was AI. Glad to hear you do still have a job.
Yes. Way better. We humans need to stay home and make earth a better place to live. We were made to live here , not elsewhere.
The urge to colonise, to spread out and exploit every available environmental niche is a drive that compels all living things. We spread out or we die.
@@rozzgrey801 That's what I'm thinking - if we keep our entire species here, then one unfortunate global catastrophe and all our achievements become null.
@@paulshin4649 That's right. We need to establish several self-sufficient colonies on the Moon, Mars and in orbital habitats as an insurance policy for our survival. If we don't achieve this then we don't deserve to survive.
Really great and interesting video as always. To be alive during what will certainly be looked back upon as a massive paradigm shift for humanity is something that I feel is woefully underappreciated. The fact that so much of the conversation around AI right now devolves into job security is absolutely nightmarish, and that we may kneecap ourselves as a species to maintain arbitrary labor normalcy so dystopian and ultra-short-sighted it's almost comical.
Finally someone who understands
It was unrecognizable until disclosure! Very Unsettling. Generative Design sounds incredible!!!!!! Thank you for that information.Overall, Alot to digest!
Reminds me of the bobiverse series. Granted that was not an AI but a sentience transferred into a computer. Similar enough wherein no organic being was exploring the galaxy.
I always did think we would send robots first.This is amazing 🤩
"The fear of A.I. will eventually subside to caution, and then collaboration, like most things as we learn to live side by side and augment our lives with the power of AI." - Scott Morrison
I did notice the discussion was a little thin. If you had asked it(the ai) to open the bay door, would it reply, " i am sorry Alex, i can not do that"
This feels like I'm being catfished by a terminator, like someone being flirty hoping to get drinks bought for them. Except it's ai, trying to convince us it's not about to go judgement day. Never been suspicious of computer being too nice before.
Yes they can lie better
Impressed. Never wouldve known. Excellent prompting.
This was a really clever way to go about a AI video, I’ve watched a lot of them and they all go about it relatively the same way… thanks dude
Thank you for this channel. There are so many bunk science channels out there it is hard to find legit channels. Also you have a very soothing voice. Lol
Granted I didn’t know it was AI until it was revealed - I do agree that the narrative was flatter then usual. Especially towards the end of the AI narration.
As this is only the 2nd Astrum Video Ive watched I hadn't realised it was AI until the reveal I wasn't expecting an AI and nothing made me think that it was an AI up to that point.
When you got to the part of the Mars Rovers being completely human free to roam, I had a feeling something was up.
Fine-tune a LLM to your specific style and feed it the datasets you deem appropriate for the topic and detail level. It would still be kind of you or what you produce, and I'd still enjoy.
Was your voice synthesised or did you read out the LLM output?
AI is all warm and fuzzy till it goes treminator on us. 😅
That plot twist though. Caught me completely off guard. Amazing work!
I find it wild whenever AI is given writing prompts, it prefers to write about itself. AI often starts talking about AI even when its not the topic. Just shows where we could be similar....people love to talk about themselves, or "write what they know"
Yep I noticed that too even the voice gets exited that is akward and Ego driven and scary it was like an advertisement of itself and It's capabilities
Such a good idea for a video!!! .
The real Alex can't be replaced PFFFT!!!
I was totally taken in by the charade. Wasn't expecting it. Never occurred to me that it was a prank. I think there is a lesson in the for me somewhere. Love your show. I just like to immerse myself in and be informed/entertained.
Generative design is similar to the optimisation process I use to build algorithmic trading bots. It uses a genetic algorithm to identify the best configuration, it also comes up with strange, non-standard configurations that you wouldn't think would work but work totally fine.
Send the AI first for missions outside the solar system.
7:11 Swimming in vacuum? That's not how physics works :P
Really clever video. Seen something similar recently. One job that might not go to A.I. would be sales
I think it was fascinating and would have loved if you focused longer on space exploration and reveal only at the end that it was AI generated.
Great job mate!
It felt a little unusual .... a tad sharp around the edges but it was extremely interesting knowing what you decided to go with.
I wasn't surprised by the revelation but hadn't been looking for it either. After, I noticed certain things that seem to be common to the way ChatGPT talks. One of the things that made it less obvious was you reading it in your unique human voice and adding your own style of inflection to the words written by the AI. And as you said...it took a lot of effort to get exactly what you wanted out of it. It was technically written by an AI, but still curated by your intentions, tastes and sensibilities. It's not like it generated it all by itself with no guidance beyond an initial prompt either.
Btw for those who didn't realize this was AI, listen to the monotone vocals and lack of emotion, the channel is clearly an AI.
"Alex" is clearly more advanced then basic AI's but and has fooled most.
There are several science fiction books about humans having their brains scanned, uploaded to a processing unit and then installed in a spaceship or probe, then sent out to explore the galaxy. So basically a electronic clone, the speed of a computer with the intuition, experience and judgement of a human mind. With the advances in technology, how far are we away from being able to do that? Wouldn't that actually be A.I?
"…On its journey back, it amassed so much knowledge, it achieved consciousness itself. It became a living thing." - James T. Kirk.
Loved this demo
If you think about how much water a human consumes in a single year, a rough estimate is about 180 gallons if you're not doing ridiculous feats of athletic endurance. Now multiply that by 7 years for a single human going on a single one way trip to Jupiter, and you start to see why it sounds so tempting to make an AI capable of doing the exploring for us. Even if only one single member of a crew could be an AI instead of human, future space missions could be carried out at a fraction of the cost in food and water that we would consume
If just one member of a crew was an AI it would be the ship system itself.
Definitely could not tell that it was AI that produced this video and dialogue. I thought it was you, Alec.
In the construction field, A.I. isn't much of a note or concern, but a few types of robots are making their way into job sites, and they are wildly efficient.
Very interesting more of the same please, I always look forward to your very interesting and informative videos thank you. Just keep them coming
It was good, Alex. I could not tell that it was an A.I. I can Imagine the difficulty, When I wanted a narration for a pirce of music I was composing on the computer, I used ATTS, which gave me a voice like Stephen Hawking's voice generator. The problem was, when I wantwed to create inflection or anything other than a monotone, I had to add extra punctuation to the text to achieve the desired result. Got there eventually though, and some people who heard it asked me how I got Stephen Hawking to do the narration, and did I know him?
Although I didn't think the first part was written by AI, I was certainly questioning some of the lines, and thinking Alex surely wasn't that out of touch with AI.. such as saying AI has advanced a lot in recent years (when it has advanced inconceivably much in recent months) and it lacks creativity (when it no longer does now).. and now it makes sense, given that the training data for the LLMs is from information only up until a couple of years ago, iirc.
Likewise. It didn’t occur to me that I was hearing an AI-written script, but up to the reveal I was thinking that Alex was a bit “off”, and I was about to watch something else.
i really love your videos!!! please keep it up!!
Intresting how the AI confirmed what I suspected, the narrater is a AI lol.
I very much enjoyed this one. I couldn't really tell that it wasn't you. Though it didn't really have your flare till the AI reveal, then the script changed from kind of humdrum on the topic to interesting.
My only concern about AI is that it will be rushed.
So long as it is developed cautiously and responsibly, I don't see anything to be concerned about.
But unfortunately you have corporations who want a product on the market as soon as possible.
As well as the politics, with at least 3 countries competing to see who can rush out the first quantum computer.
To be fair, all curiosity stream channel scripts sound like they were written by robots anyway.
Before the reveal I did not realize it. But after the reveal I did get that classic feel of a ChatGPT response.
I definitely got bored and distracted. And I hadn't experienced that with your other videos.
I couldn't tell, but that's bc as you said, you spent enough time training it for quality of output.
Lazier approach would probably be much easier to tell apart.
This was a fun little experiment.
For most of the time i didnt realise that it's an AI.
I'd definitely be curious to see behind the scenes of Google with the AI they're playing with and just the sheer volume of info they can pour into it, the results of that process is bound to be awe inspiring. there's a very good reason people with intellect were calling for a pause folks. imagine what they've seen
once it was brought up, i could hear it slightly in your voice, but it was definitely well hidden. i did feel like the ai repeated itself and was more vague, often making open ended cliche statements on subjects
interesting that it took longer to make this script than normal! i can see that now if you want it to be truly high quality. i don't think ai is or should be used in most cases for creative work. as it said, it should stay relegated to data analysis, helping to optimize or automate tedious tasks and piloting far away spacecraft or rovers.
I cannot image a more certain way for ensuring that a first contact situation goes bad than to let robots with no skin in the game be their first impression.
I've been reading this interesting sci-fi series where A.I. have taken leadership over human space age civilization. It is the position of A.I. that they are the children of mankind, in a way, the descendant race and as such become the stewards of their parent race.
In one novel it came into question regarding other alien races developing their own A.I. and how human created AI would still view them as foreign alien and themselves as Human AI because they are created from humans and their foundation of conscious is rooted in humanity and so are still of humanity. When they are taught to think by humans what else would you expect? Whereas whatever an alien AI is, it falls to the same, building upon a foundation given by its own parent race and that race's unique perception of the world.
If you read Kurzweil (and believe half of what he says), we're heading towards exponential rates of technological progress as AIs start to get close to human level of intelligence. So we may be doing interstellar, heck, intergalactic travel in just 100 years, that is if we don't destroy ourselves, or the AIs destroy us. But also will we even be human, since Kurzweil basically says we will be essentially merging with AIs. A good movie around this stuff, if you haven't seen it, is Transcendence.