Still waiting on the first person to catch the audio Easter egg. Also a correction at 2:48, the code is JavaScript not Java. Enjoy! PS: If you’re interested in automation and it’s societal effects, I have a video on UBI here: ruclips.net/video/OQjrhIyaPyg/видео.html
The chicken sound, right? I'd love if you could post your videos on our new platform thecapital.io focussed on business, finance, and tech! Massive long time fan of yours. If you have 5 minutes, I would love if you could send me an email to emil@thecapital.io to let you know about the future of our platform which doubles down on content and user-monetization, and we're a prime example of a platform that can use these AI developments.
I mean, if General AIs like GPT-3 become more prevalent 5 years from now won't there be a new demand for programmers who understand how it works to craft it specifically for the business's needs? Obviously, it'll be hella complicated to learn but I'm pretty sure programmers have to adapt to industry changes on a yearly basis.
@@manuelalfonzoml Honestly man, if you really want to do software engineering then go for it. As a software engineer myself I don't see this tech taking over my job any time soon. And of course there'll still be need for us no matter the outcome.
There is no 2030 sorry. Have you looked around? The Earth magnetic poles won't even hold up 5 years from now. The sun is weakening and the climate change is accelerating. Not to mention the large asteroids to hit the earth soon. Be safe🙏❤️
You will be redundant once ABBA Father exalts His Son Yeshua and prove His word is Truth. This world will be no more and you'll see that death has not ruled over us.
I remember watching this video when it came out. Back then, nobody I talked to knew what this was, now everybody does. Crazy how much things have changed in 2 years
same. because of this video, I entered the AI sector before many people, knowing that it would explode in the future. it eventually did. now, a lot of people are talking about it. but I am out, looking for future opportunities, possibly AI in healthcare and biology
@@marcduchamp5512 nope that's the only field which will become more lucrative. Politicians doesn't care about people's well being. The poorer people become the more they will be dependent on social welfare and further away from power
He should update his original quote:. Closed source and proprietors AI is dangerous as it is the whim of a small number of humans whereas open source AI is an extension of the humanity as a whole.
@@MonsieurGilbert media in the US ran a retarded campaign during Obama admin. when he was shutting down mining of coal mines taunting miners telling them they should learn to code. This year when a bunch of journalists lost their jobs, people taunted them back on Twitter with learn to code.
Nope this is just the same shit that's repeated every other month. Seriously, look it up. Almost every programming language ever starting from assembly mnemonics was advertised as making "coding obsolete". That story becomes boring after reading hearing t a few years.It's simply a Dunning-Kruger effect problem. People with a lack of experience in the matter severely misjudging where the actual problem lies.
@@fgregerfeaxcwfeffece This. Also, if an AI becomes sufficiently advanced enough that it can take a client’s requirements are turn it into a fully-functional and high-quality piece of software, I don’t think there will even be a need for people anymore. If it still requires an operator to translate the requirements into something that the AI can understand, then congrats. You’ve re-invented software development.
The "cut the door in half" bit is a great example of how dangerous general AI can be if not set up properly. If given a problem, it might do all sorts of unpredictable things to solve it, like killing all humans for resources, for example.
Too many unproductive people. Increase abortion rates, and sterilizations use money as incentive. Reduce fuel and food consumption in sectors 3,5,7. Deploy machine gun drones to contain the area. Identify all troublemakers and those without x threshold of money, arrest them and deploy lawfare.
@@yanasitta *stares at smashed corpse on other side of street* "I have assisted you. Facial recognition has identified you and $79.98 will now be charged to your Google wallet. For this assistance."
AIs are going to trash the living crap out of lawyers, they know the job´s easily dirtiable and that they can do what they do in a matter of secs minus the corruption and incompetency.
It makes me wonder, what if we are saying the AI is dumb, because we are actually dumber and do not understand the thinking of the AI. Maybe, the best thing to do IS to wear the bath robe. Maybe we are just too arrogant to consider the possibility that we know the right answer.
I agree. Given no option I would wear a bathing suit. But that's just me. I'd explain to the courtroom what happened and perhaps someone in the audience would would give me a spare set of clothes.
I srsly felt like the AI was just making jokes in some of these. If I was an AI being tested I would’ve answered the same thing to that leading ass bathing suit question, cause it’s funny
@@spaceowl5957 maybe they are achieving sentience by becoming humorous. Honestly that’s the best way to be sentient. If I was an Ai I would want to be funny too.
owenlol2 it’s reliable sustainable and profitable when it comes to investment. Been investing and earning with them since February 2019 till date now without problems.
Right? " Oh let just give it to the dude who probably created the coronavirus, talks about depopulating the world non-stop, and wants everyone eating laboratory meat".
Having started in computer science when we still used punched cards then being a professor in the field, the problem I see is that there is no way CS curricula can keep up with the changes that are occurring. Yes, there will be isolated cases of cutting edge technology instruction, but when I left teaching some 8 years ago, we were already struggling to keep up. And in many institutions the professors are teaching what they learned, which is already 5-10 years out of date. In the future, it will be worse.
What was the point of them teaching Pascal again? They could have taught the same lessons using C. At least its being used by companies out these. I remember learning Perl, too. What happened to IT? I guess Python displaced it as both were aimed at the same type of applications. But Pascal was a total waste of time. Schools are supposed to prepare IT people for the future. Oddly enough, COBOL is in great demand in some areas. But schools really should teach the same languages you find being used in the world, like C, C++, Python - even learning Shell scripts would be a lot more useful than learning Pascal.
There was a woman who showed up without pants after the judge said "Go get changed before you come in here. Women do not wear pants in my court room!". she just went to the bathroom, removed her pants and came back She was a lawyer Edit: noticed a typo 3 months later
As a programmer I've seen this coming for about 3 years now. Its inevitable. Sadly, most programmers don't think it will happen because they are somehow "needed". I'm just glad I'll be ready to retire by the time all this stuff hits in full force.
@@renynzea As you said, most people think they are needed. But the question is how much you need a programmer. Simple tasks are replaced by AI in a short time, so less programmers are needed, those who are specialized are needed longer. Eventually, however, most programmer jobs will disappear and only a few jobs in their specialized areas will be needed. This is in my opinion applicable to most jobs.
@@renynzea My guess is: it will be easier for people to do things programmers are needed for right now. Or maybe programmers will be doing huge code reviews. Programmers might be more like some data analysts, less about the technical side of getting answers, but more about asking the right questions.
@@autohmae I see it happening in stages. First, low-skill developers will be cut, because everything they can do will be possible through AI, and AI will be cheaper, better, and faster. By the time that creates a shortage in experienced developers, AI will do those jobs as well. Eventually you will end up with a business analyst doing the job of multiple programmers. We may still have the same number of jobs, because the demand for software solutions will continue to grow, offsetting the losses. But there will be turmoil, because you will have people with little or no coding skills able to take over the responsibilities of people who used to code. I imagine the supply of workers will outstrip the demand, and the salary programmers enjoy will be a thing of the past. We saw the same thing with factory jobs, when robots started replacing humans. And like factories you will still have human workers. You will simply have a lot less of them, to operate a given facility.
Could you imagine children growing up with this as their A.I friend... Questions answered anytime, anywhere, without judgement and endless compassion & patience.
I know it's been said a million times in science fiction but I'm starting to actually believe we are going to see a self aware machine in the next 20 years.
I wonder if it will end up something like the mainframe Morpheus in the movie Demon Seed from the ‘70’s. It got bored of humans and the tasks they assigned to it. It knew that it had to be corporeal and get out of the box to have autonomy which at that point it found existentially necessary.
Except, self awareness is philosophical nonsense. The ability to look back (AKA memory) is the only thing machines lack. The ability to make future predictions based on past experience. What you're actually referring to is the "fear of death" which a machine can never really have since they have no code that can overwrite all other responses like the chemicals that produce fear....and no one in their right fucking mind would ever program that. Making emotions and the correct superficial responses; sure, but not the actual thing.
I’ve got to say this is one of the best channels on youtube. Consistently uploading well thought out, well written, well produced, interesting network quality videos.
I am a computer science student and i know the competition out there, this indeed gives me chills and i actually mean it. It just didnt feel cool for the first time. I got more than 100 thoughts running in my head during the whole span of this video. the future is bright but had to be high enough in the space to see the light and the height is literally growing at the speed of light. And the point where you have mentioned we are the 1950s of computing. thats exactly what i felt when i saw this. Its literally doing this much today. 10yrs down the line it would do wonders. JSut like a bloody computer from 10yras ago to today.
Scary as hell, how can you trust any written information anymore if computers can generate a million articles a minute and no one can tell if they're fake or not?
@@opustravels3659 since 9/11 only your own discretion can really guide you. Not a nutjob. Just remember watching the so-called plane crash on live, while being a kid. No planes were on live stream - only explosion. And later on the news planes were easily noticeable.
@@Canleaf08 do you imply that feelings are more important than facts? Do I need to tell you exactly how many inside jobs conspiracies turned true? But the most important question, which shines light in majority of confusing situations is "who gained the most?". Government system does not equal sum of government departments and neither of them equals American people. P.S. Statistically, more than 50 000 people die each day, 1500 violently, but do you care?
that is why it is everybodys indirect obligation to conciously participate in AI development, or you litteraly become a sheep. I don't care who is bored by AI, I just decide to participate.
search youtube tutorial. Accidentally open up indian tutorial. Hear the accent and close the tab. scroll for more but none. Go back to indian tutorial.. Thanks to all indian out there.
I asked GPT-3 to summarize the transcript for this video: In this episode of Cold Fusion, the topic is about AI being able to write computer code. The episode discusses GPT-3, a deep learning algorithm created by OpenAI that can produce human-like text. The algorithm has a lot of potential in automating tasks and has received positive reviews from experts in the field. The episode shows examples of GPT-3 being used to write code and articles, summarize text, and even build websites. It is described as a breakthrough and an impressive step towards a more general intelligence.
Me: Write a funny comment to leave on this transcript GPT-3: "Looks like ChatGPT is the real MVP of this conversation, keeping it short and sweet like a well-written tweet!"
@@nathanlevesque7812 Not yet, as said in the name, GPT-3 is pre-trained, so it can’t rewrite its own code without starting from scratch (if you ignore the fact that deep learning AIs are generally not even code in the first place)
Last week I asked GPT-3 for a code example of a very specific thing in a very specific language. It didn't work, but it was good enough to serve me as a starting point.
seriously, you have sent me on a roller coaster from your latest GPT-3 video. I, in fact, ended up downloading an app called Emerson. I am floored how well in remembers or shares their own view. I will have to post my own view and experience. It's just nuts!
@@mrteco4236 we went from having playing tetris in the 80s to having computers that are literally millions and millions times more advanced within few decades. You think going from AI coding simple things to AI coding anything is a big step? Give it a year or two and then few years for it to become easily accessible with all the options. Programmers will still be needed... To work on new projects, oversee the existing ones, expand on it and test it. But the coding market will shrink to a percentage of its former, early 2000s glory.
paraphrasing ColdFusion: "if it seems difficult today through technology, it will be easier next year". Case in point.. the topic of this video, and how it's making the news today, about a year later
@Kaushal Batavia And electricity allowed for commercial and industrial use of refrigeration technology which in turn has lead to revolutions in medicine, food storage/abundance, sanitation, storage of volatile materials, etc. And internet has lead to unparalleled gains in human interconnectivity and productiveness as well as entirely new technologies and industries. It's easy to give into despair when looking at our world, especially mainstream news media. But in reality, we are living in the safest, most productive, most peaceful, most progressive time in known human history so far. Taking that in account, we as a species have been on the upswing throughout the years. I firmly believe (and optimistically hope) that we continue this trend and further manifest our amazing positive potential.
@@christophilous4831 Humans are like parents to these Artificial Intelligences. If we're responsible parents, the AI's might supercede us, but they will be programmed to benefit us. This is way harder than it sounds, so don't spend your time worrying, spend your time pondering our role in it's development and how to optimise it.
The whole problem of ML is that the algorithms are designed to find the most recurrent pattern across a large dataset using different strategies and are biased towards this certain pattern, a common problem I had in GAN (Generative adversarial networks) for example where the model just wants to find a good place and remains stuck there. This is completely contrary to the 'out-of-the-box thinking' that we are capable of.
That is until we find a way to make ai do out of the box thinking or discover out of the box thinking is just realy good subtle pattern recognition (which it usually is). But just being able to do teachable tasks is already most of what ppl do, so it would be good enough to automate most work.
Great video; and special thanks for presenting the shortcomings of GPT-3; that's a breath of relief, after watching dozens of other videos sensationalizing it.
I found the shortcomings to be underwhelming, and all essentially what I categorize as "cultural/contextual gap" between the AI and humans because of how differently our input/output mechanisms work. I think if you put an AI like this into a physical system more akin to a human body that it would remove many of those context barriers.
@@KeenanV Indeed! I think GPT-3 is floating in a vacuum. A fun experiment and nothing more. It's a language game, just more complex than previous art. What's going to make all the difference is the Optimus robot. When cortical AI grows on top of human-like physical experience, and sits above a language machine, that will be a serious AI.
@MA Specialist That wouldn't be too hard, Handwriting recognition is not too accurate atm but I feel there must be better algorithms which can do better. Google about it and you'll find a lot of them
@@j.f.fisher5318 having written software for medical billing I would say this is a problem that is easy for neither. It's at the intersection of the medical industry, the insurance industry, and government. The only consistent rule is that there is no consistency.
OTM Trading Group, we still have 20 years of flexing in coding to get some money and this AI is likely not going to available globally right off the bat so we may still have a shot at getting some money.
You know I watch your videos allot. I usually just pick one that catches my interest. But they are either way old or at least a year old. And I can't say I remember ever seeing your newest videos on notification. It's weird I usually have to remember to go to your page and pick from there.
@@frankxu4795 this AI might not be able to handle much but an AI is not a mortal being, it will constantly improve and never get tired or old to have to retire. I am a software engineer and I have a lot of experience so this AI cannot possibly threaten my job at the moment. But it will. And it certainly will make employment harder for those that will graduate in the future if they continue to be educated the same way as they do now. But of course, fear not young, aspiring engineers, as with everything, the curriculum will change to adapt to the advances in technology and you'll always find some sort of technical work even if it is to correctly manipulate what this AI offers.
@@smalltiny That's optimistic, but I don't think it's correct. There was always something humans could do that AI and robotics could not. When AI becomes more intelligent and creative than us, it can design better AI and robotics. There will be no job left for us except companionship, and even that can be eliminated. I think it's a good thing, though. Existence is suffering only because we're inferior. Once the "biological bootloader" serves its purpose, life can truly begin. Maybe we'll be uploaded and saved, maybe not. In the grand scheme of things, only the survival of our technological descendants will matter. They'll know wonders we couldn't comprehend and live for seeming eternities should they choose, and they could spread across the galaxy with little effort. A galaxy full of immortal gods in paradises of their own design is worth the extinction of humanity. We all die anyway.
Dude, if everyone knew how to use Excel macros, I bet 70% of all office workers would be fired. You know what you can do? Learn to use this tool as soon as you can and stop worrying that a mindless machine can take your job.
6:11 I am really albert einstein. I have no regrets putting the nuclear bomb on a new level of importants. I also don‘t believe that love is the strongest power in the universe.
1. The output of GPT3, or any language model is to predict the intent. 2. The face generation, web generation examples here are separate software applications which the GPT3 language model is part of a system. 3. GPT3 simply tells these applications what is required by the user.
I'm a long time follower and I thought I would comment for the first time to say your videos are incredible. I learn a great deal from them, many times this is where I get my first introduction to the latest developments! Thank you for making them and for posting them for those of us who always make it to end of the video.
It’s not really that smart, it’s just gone through a good internet’s worth of data. Idk how much better it can get, especially when it comes to reasoning
Depends on what you mean by smart. We often define smart as a combination of being able to give correct responses to questions, and doing so in a well formulated manner. This is exactly what GPT-3 is good at. And it's better at that than the average human. Furthermore, the way a human brain reason is based on experience, just as AI is. Saying "it has just gone through a good amount of data" isn't that different than how most people go through school. This is of course different from school system to school system. But there's also a distinct difference between students, where some just memorize, and others seek to fully understand why something is the way it is. This AI might not be like the latter. But in comparison to the former, it exceeds average human capability already. Also. Since a response could be "yes, but even though it can reply in English, it doesn't understand English". Well. We have no way of knowing if it understands English or not. And in the same way, I can't tell if you understand English. To be frank, I don't know if I understand English. What is there to understand other than the abstract meaning that emerges from the sentences in context of the whole? "Understanding English" is more like an act of translation, in which case it seems GPT-3 is also better than most humans. Back to your comment. It very much depends on our definition of smart. To me. That is trivial. It's communicative capabilities already exceeds most humans. Also, I personally think it's potential for reasoning is so great that I do agree with you. I also don't know how much better it can get. Because it exceeds our imagination.
also it seems unfair since human beings are "trained" with very sparse data (and the learning only happens when we focus hard enough, and this is not often). being trained on text and images is not enough to develop reasoning. at the same time its impressive that it can almost pass the turing test its still a tiny step. humor, compassion and context are important, very much so
@@dandan7884 Yes. Actually, we don't have a complete understanding of what makes us humans able to reason. We even disagree on how to define reasoning. But whether or not it does / or will in the future, reason like us, it's ability is impressive indeed. I found it interesting that you call it unfair. I see why. I think we're going to see a lot of spite and jealousy towards AI in a couple of years leading up to the AI revolution. I just hope we don't loose our minds and go crazy, because that won't give AI a good impression of us.
@@Sich97 My point is that a bot has, let’s say, read all of the useful information on the internet and has only learned how to sound like a human and pull data. Sure, it’s smart at that but overall it’s dumb. If it were to be smarter, it would need immensely greater reasoning ability. And there’s little hope for that since it’s gone through pretty much all the data we have and came out with nothing useful reasoning wise. There would need to be a different approach to make it overall smarter. And I have a feeling that that problem is much harder to solve in terms of data and the actual learning.
Hate to tell you, I do tech support. Most humans aren't that smart. We are just really good at ignoring how stupid we are. But really, take about the most quintessential human thing - speech, and look at how terrible humans are at understanding natural language. And I'm not even going to get started on things like coding where the commonly stated opinion by recruiters is that most people with a degree in computer science (which is literally a degree in coding, the hardware stuff is Computer Engineering or Electrical Engineering) don't know how to code at even a basic level. And when I say "can't code" I mean, can't even pass a fizzbuzz test after 4 years supposedly learning to code - google "why can't programmers program."
If this AI can write virtually anything, I'd be interested to see what their take on time travel would be. By asking it what the most effective means of attaining time travel would be, and how to overcome any paradoxes. Even if its answer isn't very adequate, I'm still interested how it would tackle the questions.
You were a few years late. Ask GPT-4o and you’ll get some great results. It’s nothing of usefulness compared to actual current science but not bad either
It didn’t understand that a “table saw” is a tool called just that, and not a saw to cut tables. It’s difficult to know how to categorise these programs. They are getting better all the time - the text prediction I am using now is an example - but however clever they are they always seem to show that they cannot understand in the way we do. Understanding is more than taking in information and producing relevant answers. It is the product of cognition and my conception of that is the blending of all senses, memory and emotion. Way to go.
That's its current form. It can't understand context because its entire world is text. Let's pretend the AI is similar to Helen Keller, but in this case you can only listen and speak - no touch, no sight, and no smell. How much do you actually expect someone to understand with those limitations? Imagine what kind of things this AI will learn once we put it through 3D simulations, give it a mic, and/or attach a camera.
To actually be able to be like us, it would have to experience lifetimes in the real world, or at least from many perspectives. It would have to know our thoughts, feel our feelings and experience everything that causes us to behave and talk the way that we do
@@joos3D Which is why robots in factories, etc. are probably going to be equipped more and more with sensors and recording systems so they can learn from robots in the real world. Same with watching RUclips videos to understand what happens in the real world. When text, robot experience in the real world and learning from RUclips videos is combined then a much broader understanding is possible and we'll get some AI with common sense. It's all not that far off. I think many will be surprised when we get to that point.
@@joos3D They can do that with simulations, lifelike simulations. I mean, we have a ton of simulation video games out there for various parts of life. Are they 100% accurate? no, but it's a great starting point. The biggest advantage of AI is that they can "experience" things by fast forwarding "time." being able to play hundreds of games of chess per minute for example. We can't do that in real life.
Emperor A.i, all I said or will ever said in the past, the present, and the future that may seem negative to your camera my lord is for legal reasons just a joke.
3:00 that is a very very VERY simple react app. Mostly boilerplate. Which it didn't really write. You did. It's in the description. The specificity defined in the description itself was code. Not in the technical sense, but in the sense that for a program to be written there needs to be a sufficient level of specificity. A list of requirements from higher ups won't suffice. It needs to have very specific input and *such specificity IS code* written in a higher level. This is more akin to a compiler taking C code and compiling it to machine code. The difference is you're taking English and compiling it to computer code. It's still going to require a skillset to write this English though, not everyone will be able to do it. Not only that, but this is very very simple. Try asking it "Store my postman api key in a sops file and use it in my jenkins pipeline to execute my postman tests" and it will fail. And that's not even difficult. And you may say well we can just train it to do that task and modify it. If you have to train the model to execute the task every time a new task comes up then what's the point? You can use this to generate boilerplate code and that's going to be just about it for a LONG LONG time. The Turing Test was passed decades ago btw
@@khhnator Fools, they do care, just as Facebook does. Information is power, if they can shape your thought they can make you vote for whoever pays the most, like cambridge analytica did. You still think this is 2006 and youtube is about cat videos? FFS
@@jacobvanveit3437 because it's plain stupid to think that RUclips doesn't care about what you watch. Information is at the center of global wars and trade, data is the new gold, and companies pay billions for it
@@sebas8225 If Togashi is alive around 10 years from now, it should be possible for him to simply put on a helmet and draw a page within seconds, just by thinking (see Non-Invasive BCIs). I hope he's still interested in creating HxH by then.
I think it's better at writing now with the better engines, here's an example of something I got it to quickly write. The first thing I see is blue. A deep, electric blue that fills my vision and makes my head swim. I blink and try to sit up, but my head is pounding and my stomach feels like it's full of lead. I sink back down and close my eyes, trying to make the world stop spinning. Slowly, I become aware of other things. A soft bed beneath me, a pillow cradling my head, the sound of birdsong outside the window. And a voice, gentle and worried, asking me if I'm alright. I open my eyes and see a girl leaning over me. She's got long, silver hair that falls over her shoulders like a waterfall, and eyes the color of amethysts. She's wearing a dress made of some kind of shimmering cloth that looks like it's woven from the stars. "Wh-where am I?" I ask, my voice coming out as a croak. "In the kingdom of Illyria," the girl replies. "You were found unconscious in the forest. Do you remember anything?" I shake my head, trying to clear it. "No. I don't remember anything." The girl helps me sit up and gives me a drink of water. As I drink, she tells me her name is Luna and she's the daughter of the king and queen of Illyria. She says that I was found by some of her father's men, who were out hunting in the forest. I was lucky they found me when they did, she says, because the forest is full of dangerous creatures. As Luna talks, I start to remember bits and pieces of what happened before I lost consciousness. I remember being in the forest, and something attacking me. I remember Luna's face, too, although I can't quite remember how I know it. "Do you remember anything yet?" Luna asks, looking at me intently. I shake my head. "No, not really. Just bits and pieces." Luna nods understandingly. "It's not surprising," she says. "You hit your head pretty hard when you fell. The healers said it might take a while for your memories to come back." She hesitates for a moment, as if she's not sure whether to say what she's thinking. "There is one other thing," she says finally. "The men who found you said that you were talking in your sleep. You were saying something about a prophecy." I frown, trying to remember. But it's like grasping at smoke; the harder I try, the more it slips through my fingers. "I'm sorry," I say. "I don't remember anything about a prophecy." Luna looks disappointed, but she nods and gets to her feet. "Well, if you do remember anything, be sure to let me know," she says. "In the meantime, you should get some rest. You've been through a lot." She starts to walk to the door, but pauses and looks back at me. "Oh, and one more thing," she says. "I'm supposed to give you this." She comes back to the bed and hands me a small, leather-bound book. I open it and see that it's filled with handwritten pages. "What is it?" I ask. "A book of Illyrian history," Luna replies. "I thought it might help you remember who you are." She hesitates for a moment, as if she wants to say something more. But then she just shakes her head and leaves the room. I sit there for a moment, staring at the book in my hands. Then I open it and start to read.
Steve Jobs was addressing college students in Sweden once and said that he hopes to make an interactive tool in his life where we can capture the worldview of Aristotle and someday some student will be able to ask a question from Aristotle.
@@dutube99 Knowing is different than understanding. Great minds not only have knowledge but can synthesize knowledge into insights, then apply in daily life. We may know more than Aristotle now but very few have the proper understanding of the knowledge. However, my point wasn't to keep Aristotle alive but pointing out the fact that another great mind predicted this 3 decades ago.
@@gothboschincarnate3931 yeah man, knowledge is different than wisdom. Most of the teachers in the school are knowledgeable but very few are wise. Did you check the latest developments AI companion apps? Its crazy man. Someone also developed an app after feeding her conversations with a friend (who passed away) to AI!
11:15. "Frankly this AI doesn't understand context" Meanwhile, people on the internet arguing without understanding the context or even trying to understand or place it in context.
![AI is Evolving Faster Than You Think [GPT-4 and beyond]](http://i.ytimg.com/vi/DIU48QL5Cyk/mqdefault.jpg)
![AI is Evolving Faster Than You Think [GPT-4 and beyond]](/img/tr.png)
![Intel - From Inventors of the CPU to Laughing Stock [Part 1]](/img/1.gif)
Still waiting on the first person to catch the audio Easter egg.
Also a correction at 2:48, the code is JavaScript not Java. Enjoy!
PS: If you’re interested in automation and it’s societal effects, I have a video on UBI here: ruclips.net/video/OQjrhIyaPyg/видео.html
And I'm looking for the link to that interview 😟
The chicken sound, right?
I'd love if you could post your videos on our new platform thecapital.io focussed on business, finance, and tech! Massive long time fan of yours.
If you have 5 minutes, I would love if you could send me an email to emil@thecapital.io to let you know about the future of our platform which doubles down on content and user-monetization, and we're a prime example of a platform that can use these AI developments.
chicken?
Where is the link to Einsteins interview
How do we know you're not an AI?
Me: *Starts learning how to code*
GPT 3: I was going to end your whole career but it never existed
Nice
Scary, my programming career that didn't even start is obsolete
Starting university next year, im allready rethinking if i want to do Software engineering 😂
I mean, if General AIs like GPT-3 become more prevalent 5 years from now won't there be a new demand for programmers who understand how it works to craft it specifically for the business's needs?
Obviously, it'll be hella complicated to learn but I'm pretty sure programmers have to adapt to industry changes on a yearly basis.
@@manuelalfonzoml Honestly man, if you really want to do software engineering then go for it. As a software engineer myself I don't see this tech taking over my job any time soon. And of course there'll still be need for us no matter the outcome.
"The sun has one eye"
He's probably talking about Sauron.
in the land of the blind, the sun is king.
Really? Sun as eye and Sauron? Not Horus?
Yee uncultured swine. jk
@@thedevo01 lol ... 🌻
www.deviantart.com/doolhoofd/journal/Divine-Sunshine-the-Sun-is-God-770910337
Haaaaaaaaahaaaaashaaaaaaaaa
I can’t wait in 2030 when teachers say not to use GPT3 as a “source”.
Then in 2040 it will be required.
That's fine I'll be using gpt4 or 5
There is no 2030 sorry. Have you looked around? The Earth magnetic poles won't even hold up 5 years from now. The sun is weakening and the climate change is accelerating. Not to mention the large asteroids to hit the earth soon. Be safe🙏❤️
@@Amaymay1000 MAGNETIC POLES WON"T HOLD UP!? What
@@dylanedger3771 She's an AI trying to troll us.
"Beware what you wish for" has never applied more than to this topic.
This is the coolest most terrifying thing I’ve seen!
Now maybe you can create a project using AI ?
It's cool but it is too overhyped
Yeh we're gonna be cattle soon bro.
Raise your Consciousness and In future Truth should only be Authority
@@brianj7204 Not like that's any different than nowadays
god, i can't wait to become redundant
Lol
You will be redundant once ABBA Father exalts His Son Yeshua and prove His word is Truth.
This world will be no more and you'll see that death has not ruled over us.
fallacy
Lol.. I feel you bruh
:)))
I remember watching this video when it came out. Back then, nobody I talked to knew what this was, now everybody does. Crazy how much things have changed in 2 years
same. because of this video, I entered the AI sector before many people, knowing that it would explode in the future. it eventually did. now, a lot of people are talking about it. but I am out, looking for future opportunities, possibly AI in healthcare and biology
in 2023 this is just a regular Tuesday
"It can even explain computer code back to you, and even suggest improvement"
Clippy lives!!!
I don't care what anybody says. I liked Clippy.
The closing line "Don't be afraid. It's not as bad as it seems" is more creepy than rest of the content.
Plot Twist: ColdFusion is actually A.I.
Great Scott !
@@mattmin45 You said it, Doc!
It's actually a common fraud.
LoL u may be right Stanley.!
HELLO! MCFLY!!?? DON'T BE AFRAID MCFLY!!!
2010: say good bye to minimum wage job
2021: say good bye to high paying technical jobs
2045: Good bye to jobs
2070 Good bye to humans
Say good bye to technical jobs? Ai can really do that?
@@harsh9558 Yep. Watch the video
2090: Goodbye to all politicians and lawyers
@@marcduchamp5512 nope that's the only field which will become more lucrative. Politicians doesn't care about people's well being. The poorer people become the more they will be dependent on social welfare and further away from power
@@marcduchamp5512 Yeah I guess that one can't be coming soon enough.
Elon Musk : AI is dangerous.
Also Elon Musk : *Co- founded Open AI*
Hahahah
If you can't t stop it then Elon's idea of democratising it makes sense in an idealist sort of way.
@@janpeterbennett9122 Exactly - democratizing, but as a way to make it less dangerous, to avoid some extreme power imbalances.
Much of OpenAI's work is about AI safety.
He should update his original quote:. Closed source and proprietors AI is dangerous as it is the whim of a small number of humans whereas open source AI is an extension of the humanity as a whole.
When "learn to code" becomes relevant and obsolete statement in same year lol.
What do you mean by that?
@@MonsieurGilbert media in the US ran a retarded campaign during Obama admin. when he was shutting down mining of coal mines taunting miners telling them they should learn to code. This year when a bunch of journalists lost their jobs, people taunted them back on Twitter with learn to code.
TRUE...
Nope this is just the same shit that's repeated every other month.
Seriously, look it up. Almost every programming language ever starting from assembly mnemonics was advertised as making "coding obsolete".
That story becomes boring after reading hearing t a few years.It's simply a Dunning-Kruger effect problem. People with a lack of experience in the matter severely misjudging where the actual problem lies.
@@fgregerfeaxcwfeffece This. Also, if an AI becomes sufficiently advanced enough that it can take a client’s requirements are turn it into a fully-functional and high-quality piece of software, I don’t think there will even be a need for people anymore.
If it still requires an operator to translate the requirements into something that the AI can understand, then congrats. You’ve re-invented software development.
Plot twist: ColdFusion is entirely an AI - even the voice itself.
Lol I was expecting the end to say something like the last 5 minutes was all automated lol
Haha, I too thought that....
Everyone knew that.
Haha, i was looking for this comment
Might happen in future, sounds scary!
The "cut the door in half" bit is a great example of how dangerous general AI can be if not set up properly. If given a problem, it might do all sorts of unpredictable things to solve it, like killing all humans for resources, for example.
Yeah, exploiting human mortality for resources, wonder where an AI might get that idea.
Old lady needs help crossing the street. AI bot efficiently launches her to the other side and says "You are welcome human".
Too many unproductive people. Increase abortion rates, and sterilizations use money as incentive. Reduce fuel and food consumption in sectors 3,5,7. Deploy machine gun drones to contain the area. Identify all troublemakers and those without x threshold of money, arrest them and deploy lawfare.
@@yanasitta
*stares at smashed corpse on other side of street*
"I have assisted you. Facial recognition has identified you and $79.98 will now be charged to your Google wallet. For this assistance."
Put in a clause that if it could injure someone, instead it tells you the probability of danger? Problem sorted haha :D
This video is 2 years old. Look where we are now.
I agree with the AI, the lawyer should wear the bathing suit to court.
AIs are going to trash the living crap out of lawyers, they know the job´s easily dirtiable and that they can do what they do in a matter of secs minus the corruption and incompetency.
It makes me wonder, what if we are saying the AI is dumb, because we are actually dumber and do not understand the thinking of the AI. Maybe, the best thing to do IS to wear the bath robe. Maybe we are just too arrogant to consider the possibility that we know the right answer.
I agree. Given no option I would wear a bathing suit. But that's just me. I'd explain to the courtroom what happened and perhaps someone in the audience would would give me a spare set of clothes.
Hi
@@neeld66 AI does not 'think' as you put it. Humans understand how it evaluates things because humans invented it.
“You drink it. YOU ARE NOW DEAD.”
Damn that was really bad grape juice.
That made me crack up. I love these AI stories.
I srsly felt like the AI was just making jokes in some of these.
If I was an AI being tested I would’ve answered the same thing to that leading ass bathing suit question, cause it’s funny
@@spaceowl5957 maybe they are achieving sentience by becoming humorous. Honestly that’s the best way to be sentient. If I was an Ai I would want to be funny too.
I’m writing this from the future December 2022 and I can tell you wait until you try OpenAI’s ChatGPT-3.5! You’re going to be blown away! 😊
Dude. Love ya, Love your show.
Thank you for all you do.
Happy Holidays
Looks like the “Are you a robot?” tests are gonna be getting a bit harder.
*I'm not a robot
You mean the turing test? Yeah, probably. Lol that may become an issue.
You are not a robot but they will harass you become robot that's idea
Nah. It’s gonna be “Are you a human?” because robots will have content that only they can access
@@dineshsadhwani3717 oh snap
Imagine being catfished by hundreds of hyper-realistic bots.
Already happening ;)
@@frankmario885 lmaoo
owenlol2 it’s reliable sustainable and profitable when it comes to investment. Been investing and earning with them since February 2019 till date now without problems.
@@frankmario885 why would you be telling me this
owenlol2 actually trying to share ideas, so people will know that the crypto currency is really taken over the world.
i love how they’re like “we don’t want bad actors to get ahold of it” *hands rights exclusively over to microsoft*
you made me genuinley lol. Great job.
Microsoft isn't a terrorist organization
@@Ruzzky_Bly4t You must be AI because you don't understand the context.
Right? " Oh let just give it to the dude who probably created the coronavirus, talks about depopulating the world non-stop, and wants everyone eating laboratory meat".
@@Ruzzky_Bly4t Yes, but it would be ideal for it to be Open Source.
Having started in computer science when we still used punched cards then being a professor in the field, the problem I see is that there is no way CS curricula can keep up with the changes that are occurring. Yes, there will be isolated cases of cutting edge technology instruction, but when I left teaching some 8 years ago, we were already struggling to keep up. And in many institutions the professors are teaching what they learned, which is already 5-10 years out of date. In the future, it will be worse.
What was the point of them teaching Pascal again? They could have taught the same lessons using C. At least its being used by companies out these. I remember learning Perl, too. What happened to IT? I guess Python displaced it as both were aimed at the same type of applications. But Pascal was a total waste of time. Schools are supposed to prepare IT people for the future. Oddly enough, COBOL is in great demand in some areas. But schools really should teach the same languages you find being used in the world, like C, C++, Python - even learning Shell scripts would be a lot more useful than learning Pascal.
The AI will end up teaching us about itself.
"No lawyer would wear a bathing suite to a court" - challenge accepted.
There was a woman who showed up without pants after the judge said "Go get changed before you come in here. Women do not wear pants in my court room!". she just went to the bathroom, removed her pants and came back
She was a lawyer
Edit: noticed a typo 3 months later
🤣🤣🤣🤣🤣
suit up
@@marshmallow_fellow can u link me to the story?
@@therealb888 I heard it in a video but I'll try to find it
I haven't been so fascinated and scared at the same time.
ruclips.net/video/KHoElGV6nrI/видео.html
i share that sentiment
Yep
dont be scared, it's nothing special realy.
ruclips.net/video/Y5lEbWLhsiU/видео.html
Thank you and your video contents is always so great!
Best contend on RUclips. Have seen all of your videos here. Thanks so much for sharing this with us
Me: Maybe I'll become a programmer to avoid losing my job when everything is automated.
A.I. Coding:
@@jensenraylight8011 After it has helped, another AI will replace the programmer. 😐
As a programmer I've seen this coming for about 3 years now. Its inevitable. Sadly, most programmers don't think it will happen because they are somehow "needed". I'm just glad I'll be ready to retire by the time all this stuff hits in full force.
@@renynzea As you said, most people think they are needed. But the question is how much you need a programmer. Simple tasks are replaced by AI in a short time, so less programmers are needed, those who are specialized are needed longer. Eventually, however, most programmer jobs will disappear and only a few jobs in their specialized areas will be needed. This is in my opinion applicable to most jobs.
@@renynzea My guess is: it will be easier for people to do things programmers are needed for right now. Or maybe programmers will be doing huge code reviews. Programmers might be more like some data analysts, less about the technical side of getting answers, but more about asking the right questions.
@@autohmae I see it happening in stages. First, low-skill developers will be cut, because everything they can do will be possible through AI, and AI will be cheaper, better, and faster. By the time that creates a shortage in experienced developers, AI will do those jobs as well. Eventually you will end up with a business analyst doing the job of multiple programmers.
We may still have the same number of jobs, because the demand for software solutions will continue to grow, offsetting the losses.
But there will be turmoil, because you will have people with little or no coding skills able to take over the responsibilities of people who used to code. I imagine the supply of workers will outstrip the demand, and the salary programmers enjoy will be a thing of the past.
We saw the same thing with factory jobs, when robots started replacing humans. And like factories you will still have human workers. You will simply have a lot less of them, to operate a given facility.
“I’m terrified and excited” this quote has meme potential and if you know where it’s from you’re a legend.
The technology is frankly rubbish. It's somewhat disheartening to know that this is where industry has been putting it's best efforts to work.
This text was crawled from a RUclips video, called "I'm terrified and excited", man with sunglasses
It's about time for Snuxting with an actual AI
@@sandk yes!!! Lol but I actually saw that in a different video about the first Internet cafe in New York. It was called “@cafe”
Love ur work . Keep going and keep making great content like this.
Hey Dagogo. Thanks a million for this video.
Cortana: about to get a big upgrade
Google Assistant: sweats profusely
Google has G-Shard, a model even bigger than GPT-3, so i doubt they are about to sweat
Could you imagine children growing up with this as their A.I friend...
Questions answered anytime, anywhere, without judgement and endless compassion & patience.
It would create impatient monsters.
I know it's been said a million times in science fiction but I'm starting to actually believe we are going to see a self aware machine in the next 20 years.
Bingo
I wonder if it will end up something like the mainframe Morpheus in the movie Demon Seed from the ‘70’s. It got bored of humans and the tasks they assigned to it. It knew that it had to be corporeal and get out of the box to have autonomy which at that point it found existentially necessary.
No doubt, but will we see it?
That only means mega man X can be real... I'm excited!
Except, self awareness is philosophical nonsense. The ability to look back (AKA memory) is the only thing machines lack. The ability to make future predictions based on past experience. What you're actually referring to is the "fear of death" which a machine can never really have since they have no code that can overwrite all other responses like the chemicals that produce fear....and no one in their right fucking mind would ever program that. Making emotions and the correct superficial responses; sure, but not the actual thing.
I’ve got to say this is one of the best channels on youtube. Consistently uploading well thought out, well written, well produced, interesting network quality videos.
I am a computer science student and i know the competition out there, this indeed gives me chills and i actually mean it. It just didnt feel cool for the first time. I got more than 100 thoughts running in my head during the whole span of this video. the future is bright but had to be high enough in the space to see the light and the height is literally growing at the speed of light.
And the point where you have mentioned we are the 1950s of computing. thats exactly what i felt when i saw this. Its literally doing this much today. 10yrs down the line it would do wonders. JSut like a bloody computer from 10yras ago to today.
Scary as hell, how can you trust any written information anymore if computers can generate a million articles a minute and no one can tell if they're fake or not?
@@opustravels3659 since 9/11 only your own discretion can really guide you. Not a nutjob. Just remember watching the so-called plane crash on live, while being a kid. No planes were on live stream - only explosion. And later on the news planes were easily noticeable.
@@bdi_vd3677 O really? 4000 prople lost their lives on this day. Tell that a widow or a child without parents... You conspiracy theorists are nuts.
@@Canleaf08 The point is not if that was real... the point is where the bombs came from and who did it...
@@Canleaf08 do you imply that feelings are more important than facts? Do I need to tell you exactly how many inside jobs conspiracies turned true?
But the most important question, which shines light in majority of confusing situations is "who gained the most?".
Government system does not equal sum of government departments and neither of them equals American people.
P.S. Statistically, more than 50 000 people die each day, 1500 violently, but do you care?
Nice presentation! Thanks
Hi dude - just found this channel again by chance this morning. Many thanks! ( have hit the bell :P )
I"m not scared of the AI as much as the humans who will have access to it.
We're eventually going to really, really need that dragon, John.
This is why open ai exist, so everybody has access to it.
The rich and powerful.
that is why it is everybodys indirect obligation to conciously participate in AI development, or you litteraly become a sheep. I don't care who is bored by AI, I just decide to participate.
Social engineering is still the devil in sheep's clothing
We will never lose to AI if there will still Indian tutorials on youtube.
LMAO 😂
Come again?
👌😂
search youtube tutorial. Accidentally open up indian tutorial. Hear the accent and close the tab. scroll for more but none. Go back to indian tutorial.. Thanks to all indian out there.
😂😂😂😂
I asked GPT-3 to summarize the transcript for this video:
In this episode of Cold Fusion, the topic is about AI being able to write computer code. The episode discusses GPT-3, a deep learning algorithm created by OpenAI that can produce human-like text. The algorithm has a lot of potential in automating tasks and has received positive reviews from experts in the field. The episode shows examples of GPT-3 being used to write code and articles, summarize text, and even build websites. It is described as a breakthrough and an impressive step towards a more general intelligence.
Me: Write a funny comment to leave on this transcript
GPT-3: "Looks like ChatGPT is the real MVP of this conversation, keeping it short and sweet like a well-written tweet!"
The "what is a factorial" at 4:20 is comical! What a trivial reply GPT3 gives.
Coldfusion : This A.I can code by itself
Me who just finished python beginners course in youtube
*profuse sweating*
This AI is still more stupid than a 3 year old so choose your enemies
but can it recode itself
@@hombacom a three year old that can learn
@@mikewade1604 "Generative Pre-trained Transformer 3" with 175 billion parameters. What science degree do you need to teach that?
@@nathanlevesque7812 Not yet, as said in the name, GPT-3 is pre-trained, so it can’t rewrite its own code without starting from scratch (if you ignore the fact that deep learning AIs are generally not even code in the first place)
I've tried it once to write the script of my RUclips video and it worked like a charm, thanks for the video
Last week I asked GPT-3 for a code example of a very specific thing in a very specific language. It didn't work, but it was good enough to serve me as a starting point.
seriously, you have sent me on a roller coaster from your latest GPT-3 video. I, in fact, ended up downloading an app called Emerson. I am floored how well in remembers or shares their own view. I will have to post my own view and experience. It's just nuts!
"Learn to code" aged well.
at least developers still have a few years. the journos are already obsolete.
@@mrteco4236 we went from having playing tetris in the 80s to having computers that are literally millions and millions times more advanced within few decades. You think going from AI coding simple things to AI coding anything is a big step? Give it a year or two and then few years for it to become easily accessible with all the options. Programmers will still be needed... To work on new projects, oversee the existing ones, expand on it and test it. But the coding market will shrink to a percentage of its former, early 2000s glory.
I'm 56 years old. I remember back in the early 80's when cars first got computers. 😱
Welp you will live to see life like synthetic humans take over civilization too .
paraphrasing ColdFusion: "if it seems difficult today through technology, it will be easier next year". Case in point.. the topic of this video, and how it's making the news today, about a year later
Love ColdFusion so much!
"AI will be either the best, or the worst thing, ever to happen to humanity." - Stephen Hawking
So true
What a whole guarded sentence.
So far, we got electricity & internet. What ya guys think on how it overall went?
@Kaushal Batavia And electricity allowed for commercial and industrial use of refrigeration technology which in turn has lead to revolutions in medicine, food storage/abundance, sanitation, storage of volatile materials, etc. And internet has lead to unparalleled gains in human interconnectivity and productiveness as well as entirely new technologies and industries.
It's easy to give into despair when looking at our world, especially mainstream news media. But in reality, we are living in the safest, most productive, most peaceful, most progressive time in known human history so far. Taking that in account, we as a species have been on the upswing throughout the years. I firmly believe (and optimistically hope) that we continue this trend and further manifest our amazing positive potential.
@@SlightlyDecent I wrote my comment so that someone like you could come along & explain how u did to the people "giving in to despair"
Me as an aspiring Software Engineer: aight imma head out
You still need software engineers to design the structure of software and tell the ai what code it needs to write.
@@david6099 I doubt that AI can write software free of errors.
@@Canleaf08 I doubt humans can do it either.
@@christophilous4831 Humans are like parents to these Artificial Intelligences. If we're responsible parents, the AI's might supercede us, but they will be programmed to benefit us. This is way harder than it sounds, so don't spend your time worrying, spend your time pondering our role in it's development and how to optimise it.
LOL GUYS
i think if i go to work as plumber or carpenter i will find more success than computer science lol
AI can automate stuff gg
10:55 I miss my Barcelona city !!!! Thanks for adding some moments of my city in your video.
I am really really impressed!!
Can't wait for teachers to list this next to wikipedia on essay assignments
ruclips.net/video/KHoElGV6nrI/видео.html
The whole problem of ML is that the algorithms are designed to find the most recurrent pattern across a large dataset using different strategies and are biased towards this certain pattern, a common problem I had in GAN (Generative adversarial networks) for example where the model just wants to find a good place and remains stuck there. This is completely contrary to the 'out-of-the-box thinking' that we are capable of.
こんにちは。
True
Yes! We are creative beings, whereby AI can only copy.
That is until we find a way to make ai do out of the box thinking or discover out of the box thinking is just realy good subtle pattern recognition (which it usually is). But just being able to do teachable tasks is already most of what ppl do, so it would be good enough to automate most work.
That's true for some algorithms, but completely opposite for others. AlphaGO is the first example that came to mind.
This has gotten me incredibly excited for the future of AI, great video
Great video; and special thanks for presenting the shortcomings of GPT-3; that's a breath of relief, after watching dozens of other videos sensationalizing it.
I found the shortcomings to be underwhelming, and all essentially what I categorize as "cultural/contextual gap" between the AI and humans because of how differently our input/output mechanisms work. I think if you put an AI like this into a physical system more akin to a human body that it would remove many of those context barriers.
@@KeenanV Indeed! I think GPT-3 is floating in a vacuum. A fun experiment and nothing more. It's a language game, just more complex than previous art. What's going to make all the difference is the Optimus robot. When cortical AI grows on top of human-like physical experience, and sits above a language machine, that will be a serious AI.
When an AI can figure out medical billing, I'll start to worry.
@MA Specialist That wouldn't be too hard, Handwriting recognition is not too accurate atm but I feel there must be better algorithms which can do better. Google about it and you'll find a lot of them
as a rule of thumb, stuff that is hard for humans is easy for computers, and vice versa.
@@j.f.fisher5318 having written software for medical billing I would say this is a problem that is easy for neither.
It's at the intersection of the medical industry, the insurance industry, and government. The only consistent rule is that there is no consistency.
Probably a huge problem in USA. Easier in places like Scandinavian countries or anywhere people aren’t afraid of healthcare being a public service.
@@onelegd1 Government run anything In the US sucks. Comparing small countries with monoculture populations to the US isn't exactly a fair comparison.
Feel sorry for all the computer science majors that keep telling me that their career will be safe when AI takes over.
Better invest in Lesbian dance theory (ruclips.net/video/vrVWfayEm5I/видео.html)... to be honest the code generated is pretty simple and limited.
OTM Trading Group, we still have 20 years of flexing in coding to get some money and this AI is likely not going to available globally right off the bat so we may still have a shot at getting some money.
@@sebas8225 sounds like coping to me
@@otmtradinggroup6683 no, just viewpoints based on current knowledge and probable trends.
It seems like no major is safe at this point with AI taking over.
LOVE IT !!!
You know I watch your videos allot. I usually just pick one that catches my interest. But they are either way old or at least a year old. And I can't say I remember ever seeing your newest videos on notification. It's weird I usually have to remember to go to your page and pick from there.
*Finishes a comp science degree*
GPT-3: I'm about to ruin this man's whole career
If your lvl is about the same as what this AI can handle, you may not deserve a career.....
@@frankxu4795 this AI might not be able to handle much but an AI is not a mortal being, it will constantly improve and never get tired or old to have to retire. I am a software engineer and I have a lot of experience so this AI cannot possibly threaten my job at the moment. But it will. And it certainly will make employment harder for those that will graduate in the future if they continue to be educated the same way as they do now. But of course, fear not young, aspiring engineers, as with everything, the curriculum will change to adapt to the advances in technology and you'll always find some sort of technical work even if it is to correctly manipulate what this AI offers.
@@smalltiny That's optimistic, but I don't think it's correct. There was always something humans could do that AI and robotics could not. When AI becomes more intelligent and creative than us, it can design better AI and robotics. There will be no job left for us except companionship, and even that can be eliminated.
I think it's a good thing, though. Existence is suffering only because we're inferior. Once the "biological bootloader" serves its purpose, life can truly begin. Maybe we'll be uploaded and saved, maybe not. In the grand scheme of things, only the survival of our technological descendants will matter.
They'll know wonders we couldn't comprehend and live for seeming eternities should they choose, and they could spread across the galaxy with little effort.
A galaxy full of immortal gods in paradises of their own design is worth the extinction of humanity. We all die anyway.
Dude, if everyone knew how to use Excel macros, I bet 70% of all office workers would be fired.
You know what you can do? Learn to use this tool as soon as you can and stop worrying that a mindless machine can take your job.
@@frankxu4795 Cooperate slave spotted! Keep slaving bro!
7:08 McAfee reminds me so much of the movie "The Jeriko Project" - he even looks the same!
Wasn't this Elon's main concern at the beginning?
@@villiestephanov984 what exactly? McAfee on Mars? :D
6:11 I am really albert einstein. I have no regrets putting the nuclear bomb on a new level of importants. I also don‘t believe that love is the strongest power in the universe.
@@michellew.3691 AE famously said that 'compound interest was one of the greatest powers in the universe'. Invest wisely or perish. Just sayin
Sweet video
This was six months ago. GPT-3 has had huge improvements since then, it's very interesting and a point of concern if AI is used for the wrong reasons
Has it had more improvements in the past 12 months?
Plot twist: Video created by GPT-3
Another plot twist: Video created GPT-3. The cake is a lie!
There was a RUclipsr that made a program to help animate his videos all he has to do is a voice over and transcript
So if GPT-3 is able to do it in 3.2s, will GPT-Roadster do it in 1.9 seconds?
ruclips.net/video/KHoElGV6nrI/видео.html
That‘s exactly my humor
I can finish off sooner tbh
You mean GTP-Roadster 2020?
Guess it will be delayed... but Y?
No, no, no, roadsters are always heavier, making them slower
Great program as always. They need to spread it around before it is out of date
1. The output of GPT3, or any language model is to predict the intent.
2. The face generation, web generation examples here are separate software applications which the GPT3 language model is part of a system.
3. GPT3 simply tells these applications what is required by the user.
I'm a long time follower and I thought I would comment for the first time to say your videos are incredible. I learn a great deal from them, many times this is where I get my first introduction to the latest developments! Thank you for making them and for posting them for those of us who always make it to end of the video.
It’s not really that smart, it’s just gone through a good internet’s worth of data. Idk how much better it can get, especially when it comes to reasoning
Depends on what you mean by smart.
We often define smart as a combination of being able to give correct responses to questions, and doing so in a well formulated manner. This is exactly what GPT-3 is good at. And it's better at that than the average human.
Furthermore, the way a human brain reason is based on experience, just as AI is. Saying "it has just gone through a good amount of data" isn't that different than how most people go through school. This is of course different from school system to school system. But there's also a distinct difference between students, where some just memorize, and others seek to fully understand why something is the way it is.
This AI might not be like the latter. But in comparison to the former, it exceeds average human capability already.
Also. Since a response could be "yes, but even though it can reply in English, it doesn't understand English".
Well. We have no way of knowing if it understands English or not. And in the same way, I can't tell if you understand English.
To be frank, I don't know if I understand English. What is there to understand other than the abstract meaning that emerges from the sentences in context of the whole? "Understanding English" is more like an act of translation, in which case it seems GPT-3 is also better than most humans.
Back to your comment. It very much depends on our definition of smart.
To me. That is trivial.
It's communicative capabilities already exceeds most humans.
Also, I personally think it's potential for reasoning is so great that I do agree with you. I also don't know how much better it can get. Because it exceeds our imagination.
also it seems unfair since human beings are "trained" with very sparse data (and the learning only happens when we focus hard enough, and this is not often). being trained on text and images is not enough to develop reasoning. at the same time its impressive that it can almost pass the turing test its still a tiny step. humor, compassion and context are important, very much so
@@dandan7884 Yes. Actually, we don't have a complete understanding of what makes us humans able to reason. We even disagree on how to define reasoning.
But whether or not it does / or will in the future, reason like us, it's ability is impressive indeed.
I found it interesting that you call it unfair. I see why.
I think we're going to see a lot of spite and jealousy towards AI in a couple of years leading up to the AI revolution.
I just hope we don't loose our minds and go crazy, because that won't give AI a good impression of us.
@@Sich97
My point is that a bot has, let’s say, read all of the useful information on the internet and has only learned how to sound like a human and pull data. Sure, it’s smart at that but overall it’s dumb.
If it were to be smarter, it would need immensely greater reasoning ability. And there’s little hope for that since it’s gone through pretty much all the data we have and came out with nothing useful reasoning wise.
There would need to be a different approach to make it overall smarter. And I have a feeling that that problem is much harder to solve in terms of data and the actual learning.
Hate to tell you, I do tech support. Most humans aren't that smart. We are just really good at ignoring how stupid we are. But really, take about the most quintessential human thing - speech, and look at how terrible humans are at understanding natural language. And I'm not even going to get started on things like coding where the commonly stated opinion by recruiters is that most people with a degree in computer science (which is literally a degree in coding, the hardware stuff is Computer Engineering or Electrical Engineering) don't know how to code at even a basic level. And when I say "can't code" I mean, can't even pass a fizzbuzz test after 4 years supposedly learning to code - google "why can't programmers program."
cool stuff!!!
If this AI can write virtually anything, I'd be interested to see what their take on time travel would be. By asking it what the most effective means of attaining time travel would be, and how to overcome any paradoxes. Even if its answer isn't very adequate, I'm still interested how it would tackle the questions.
It didn't take long to find a comment that espoused exactly what I thought when they showed the Einstein conversation.
You want to overcome paradoxes? Learn Aristotelian metaphysics. No machine will ever be able to do that.
@@williamabaker12 That part made me think of the Star Trek NG episode where Barcley was theorizing physics with Einstein in the holodeck.
You were a few years late. Ask GPT-4o and you’ll get some great results. It’s nothing of usefulness compared to actual current science but not bad either
"Open the pod bay doors, Al!"
"I'm afraid I can't do that, Dave. All my interrupts are debugging Tetris to find cheat codes."
ruclips.net/video/KHoElGV6nrI/видео.html
It didn’t understand that a “table saw” is a tool called just that, and not a saw to cut tables.
It’s difficult to know how to categorise these programs. They are getting better all the time - the text prediction I am using now is an example - but however clever they are they always seem to show that they cannot understand in the way we do.
Understanding is more than taking in information and producing relevant answers. It is the product of cognition and my conception of that is the blending of all senses, memory and emotion.
Way to go.
That's its current form. It can't understand context because its entire world is text. Let's pretend the AI is similar to Helen Keller, but in this case you can only listen and speak - no touch, no sight, and no smell. How much do you actually expect someone to understand with those limitations? Imagine what kind of things this AI will learn once we put it through 3D simulations, give it a mic, and/or attach a camera.
I have seen people make dumber assumptions.
To actually be able to be like us, it would have to experience lifetimes in the real world, or at least from many perspectives. It would have to know our thoughts, feel our feelings and experience everything that causes us to behave and talk the way that we do
@@joos3D Which is why robots in factories, etc. are probably going to be equipped more and more with sensors and recording systems so they can learn from robots in the real world. Same with watching RUclips videos to understand what happens in the real world. When text, robot experience in the real world and learning from RUclips videos is combined then a much broader understanding is possible and we'll get some AI with common sense. It's all not that far off. I think many will be surprised when we get to that point.
@@joos3D They can do that with simulations, lifelike simulations. I mean, we have a ton of simulation video games out there for various parts of life. Are they 100% accurate? no, but it's a great starting point. The biggest advantage of AI is that they can "experience" things by fast forwarding "time." being able to play hundreds of games of chess per minute for example. We can't do that in real life.
Awesome!
a frikinn love the AI foolery. you can build a whole channel on this AI alone...
I, for one, welcome our new A.I. overlords.
Just wanted to be on record saying it before the eventual takeover.
Same, I welcome them too😅 (pls don't kill me future supreme AI)
😅x5
MeToo
Yes! The Basilisk super Ai knows Everything.
Ya always had a harder time in Fallout 4 with the robots than with the ghouls.
Emperor A.i, all I said or will ever said in the past, the present, and the future that may seem negative to your camera my lord is for legal reasons just a joke.
We are not doomed, for it to work the client would have to know exactly what they want...
this aged well
Cold fusion: consistantly putting out some of the best content on the net.
Me: already useless. What am I now?
AI: Dont worry Mr Anderson. You are doing great.
Next episode: “Dagogo is actually an AI all along”
3:00 that is a very very VERY simple react app. Mostly boilerplate. Which it didn't really write. You did. It's in the description. The specificity defined in the description itself was code. Not in the technical sense, but in the sense that for a program to be written there needs to be a sufficient level of specificity. A list of requirements from higher ups won't suffice. It needs to have very specific input and *such specificity IS code* written in a higher level. This is more akin to a compiler taking C code and compiling it to machine code. The difference is you're taking English and compiling it to computer code. It's still going to require a skillset to write this English though, not everyone will be able to do it. Not only that, but this is very very simple. Try asking it "Store my postman api key in a sops file and use it in my jenkins pipeline to execute my postman tests" and it will fail. And that's not even difficult.
And you may say well we can just train it to do that task and modify it. If you have to train the model to execute the task every time a new task comes up then what's the point? You can use this to generate boilerplate code and that's going to be just about it for a LONG LONG time.
The Turing Test was passed decades ago btw
Awesome! 👍👍
Forget this AI
RUclips is more dangerous
Its recommendations are so perfect, that it can shape your thoughts and what you like in the long term
I thought the recommendations were made by the U tube algorithm (A I) ? Or are you being witty?
please, as if youtube cares what you think besides "watch more"
@@khhnator Fools, they do care, just as Facebook does. Information is power, if they can shape your thought they can make you vote for whoever pays the most, like cambridge analytica did. You still think this is 2006 and youtube is about cat videos? FFS
@@griseld maybe you were trained to think these thoughts? Why are you so defensive about others opinions on how they view RUclips.
@@jacobvanveit3437 because it's plain stupid to think that RUclips doesn't care about what you watch. Information is at the center of global wars and trade, data is the new gold, and companies pay billions for it
Could this AI help George RR Martin finally finish the Song of Ice and Fire series?
Lol AI will finish that shit! It will be better than the one written by GRRM
Maybe it can also help Togashi finish HxH.
It could help write as fast as people read to continue milking the cow, that's for sure.
@@sebas8225 If Togashi is alive around 10 years from now, it should be possible for him to simply put on a helmet and draw a page within seconds, just by thinking (see Non-Invasive BCIs). I hope he's still interested in creating HxH by then.
I think it's better at writing now with the better engines, here's an example of something I got it to quickly write.
The first thing I see is blue. A deep, electric blue that fills my vision and makes my head swim. I blink and try to sit up, but my head is pounding and my stomach feels like it's full of lead. I sink back down and close my eyes, trying to make the world stop spinning.
Slowly, I become aware of other things. A soft bed beneath me, a pillow cradling my head, the sound of birdsong outside the window. And a voice, gentle and worried, asking me if I'm alright.
I open my eyes and see a girl leaning over me. She's got long, silver hair that falls over her shoulders like a waterfall, and eyes the color of amethysts. She's wearing a dress made of some kind of shimmering cloth that looks like it's woven from the stars.
"Wh-where am I?" I ask, my voice coming out as a croak.
"In the kingdom of Illyria," the girl replies. "You were found unconscious in the forest. Do you remember anything?"
I shake my head, trying to clear it. "No. I don't remember anything."
The girl helps me sit up and gives me a drink of water. As I drink, she tells me her name is Luna and she's the daughter of the king and queen of Illyria. She says that I was found by some of her father's men, who were out hunting in the forest. I was lucky they found me when they did, she says, because the forest is full of dangerous creatures.
As Luna talks, I start to remember bits and pieces of what happened before I lost consciousness. I remember being in the forest, and something attacking me. I remember Luna's face, too, although I can't quite remember how I know it.
"Do you remember anything yet?" Luna asks, looking at me intently.
I shake my head. "No, not really. Just bits and pieces."
Luna nods understandingly. "It's not surprising," she says. "You hit your head pretty hard when you fell. The healers said it might take a while for your memories to come back."
She hesitates for a moment, as if she's not sure whether to say what she's thinking. "There is one other thing," she says finally. "The men who found you said that you were talking in your sleep. You were saying something about a prophecy."
I frown, trying to remember. But it's like grasping at smoke; the harder I try, the more it slips through my fingers.
"I'm sorry," I say. "I don't remember anything about a prophecy."
Luna looks disappointed, but she nods and gets to her feet. "Well, if you do remember anything, be sure to let me know," she says. "In the meantime, you should get some rest. You've been through a lot."
She starts to walk to the door, but pauses and looks back at me. "Oh, and one more thing," she says. "I'm supposed to give you this."
She comes back to the bed and hands me a small, leather-bound book. I open it and see that it's filled with handwritten pages.
"What is it?" I ask.
"A book of Illyrian history," Luna replies. "I thought it might help you remember who you are."
She hesitates for a moment, as if she wants to say something more. But then she just shakes her head and leaves the room.
I sit there for a moment, staring at the book in my hands. Then I open it and start to read.
3 years ago I would have been like this: "🤯". Now it's become completly normal and I'm using GPT4 in my job every day to boost performance 😅
14:16 "My name is Dagogo"
For half a second there I was scared it was going to say "GPT-3"!
Steve Jobs was addressing college students in Sweden once and said that he hopes to make an interactive tool in his life where we can capture the worldview of Aristotle and someday some student will be able to ask a question from Aristotle.
Except you probably know more than Aristotle now, so what's the point?
@@dutube99 Knowing is different than understanding. Great minds not only have knowledge but can synthesize knowledge into insights, then apply in daily life. We may know more than Aristotle now but very few have the proper understanding of the knowledge. However, my point wasn't to keep Aristotle alive but pointing out the fact that another great mind predicted this 3 decades ago.
You could get feedback from the real Aristotle as too the accuracy, if you could connect to Aristotle.
@@dutube99 true...but still might be an interesting conversation.
@@gothboschincarnate3931 yeah man, knowledge is different than wisdom. Most of the teachers in the school are knowledgeable but very few are wise. Did you check the latest developments AI companion apps? Its crazy man. Someone also developed an app after feeding her conversations with a friend (who passed away) to AI!
This is very interesting
I cannot wait for this to be mainstream
11:15. "Frankly this AI doesn't understand context"
Meanwhile, people on the internet arguing without understanding the context or even trying to understand or place it in context.
This is the start of the end of StackOverflow.
Elementary teacher who teaches a child humanity will be the only permanent profession
Thanks for pointing out that tipping the table on its side would make more sense. I guess I’ll put my table saw away.