Elon Musk’s A.I. Destroys Champion Gamer!

Elon Musk’s A.I. Destroys Champion Gamer!

(Intro) You are watching Coldfusion TV Hi, Welcome to another Coldfusion video It’s no secret. That artificial intelligence is progressing rapidly. It seems like every couple of weeks There’s an unexpected development that takes people by surprise It’s starting to become clear that we’re near the start of a new era that will include artificial intelligence Those people that aren’t paying attention are likely to be taken aback by the things that are going to be possible in a Few years Some voices have spoken out about the dangers and need for safety measures that must go along with developing powerful way I one of the loudest voices of concern was that of Elon musk in a strange twist elands non-profit artificial intelligence start-up Which is called open AI and has just achieved a pretty remarkable feat in this video. We’ll take a look at what’s going on So what’s the background story? Every year the game developer valve hosts the competition for expert players of the game dota 2 a popular online multiplayer game The competition hosts professional players from all around the world all fighting for a twenty four million dollar Grand prize this year there was a special guest competitor who wasn’t human it was an AI trained by the engineers of Elon musk startup open AI When put up one-on-one against one of the world’s best players of dota 2 a crowd favorite Dendi the artificial intelligence one Even a year ago. Nobody was sure if this kind of thing was possible Dendi was surprised that an AI could outplay a human He said that the AI quote felt like a human, but a little like something else end quote this feat was achieved with just two weeks of real-Time learning by the AI The engineers state that during this training period It accumulated Lifetimes of experience they also state that the rules of dota are so complicated that if you just wrote down Pre-Programmed rules and program some code to follow it the end result wouldn’t even be as good as an average player The artificial intelligence was trained from scratch with no knowledge of the game it played against itself over and over again until it had mastered the game How on Earth does a computer learn to play Delta? Yeah? So this spot is quite unlike any what you’ve seen before so We’ve coached it to learn just from playing areas itself, so we didn’t hard code in any strategy We didn’t have it learned from human experts just from the very beginning it just keeps playing it’s a copy of itself it starts from complete randomness, and then it makes very small improvements and Eventually it is the pro. Level these are pro players these are human brain So you just telling me this robot has failed so many times It’s actually better than professional dota players it’s played for really lifetimes of experience and It’s great so many games of dota. It’s explored many different strategies learn to be learn to exploit other people who beat and it’s just played far more into the strategy space than any human as Musk is hailing this achievement as the first time that artificial intelligence has been able to beat professionals in competitive esports With esports being considered for the 2024 Olympics. This is certainly interesting So this actually isn’t the first notable achievement by open Ai they’ve actually done some pretty cool stuff in the past The company actually invented a method where humans can interact with a robotic AI and teach it just like you would a human here’s how it works First a human wears a Vr Headset does a task and the robot watches and then imitates the task in real-time without ever having done it before In this case it’s learning how to stack some lego blocks that task sounds very simple for a human But it’s actually extremely hard for a machine to do this The AI manages to do this by having its visual neural network trained on a large set of images of what blocks and simulated environments could look like This first visual neural network then sees the data to another neural Network called the Imitation Network After this training and after just one single original demonstration of what to do the robot can now stack blocks? even with different colored blocks placed in different positions every time This means that the robot has to and does perform different actions to the scenarios than it’s already seen The end goal is to create an AI that can adapt to new and unpredictable environments Okay, now back to open a is recent achievement Elon has tweeted that his feet are feeding some of the best dota players in the world with an AI is a task much more Complicated than chess or the board game go Now I’m not so sure that I agree with Elon statement and esport is more complicated than go chess. Yes, definitely Let’s go perhaps. Not some state that goes the most Brilliant game ever made Go is a 3000 year old Chinese board game in which an AI called alphago from the company deepmind recently beat the World Champion in a series of playoffs This event was held the biggest moment in artificial intelligence Not expected for another decade. He is deepmind Ceo Dennis hasset us talking about some of the complexity of go Nothing about goa. Is that he only has two rules I could teach you the game in five minutes But it leads to incredible complexity. It’s probably the most elegant game that mankind has ever devised What happens in Asia and Career in Japan or and China? Is if you show promise in the game of go of at the age of sort of five or six or seven? You get taken out of normal school and get put into go school where you study go 12 hours a day seven days a week With your peers who are also trying to become professional go players this has taken really seriously, and it’s been like this hundreds of Years Now one way to illustrate the complexity of the game Is that there are not more board configurations in the gamers go than there are atoms in the universe? So there’s no way that you can solve this game through broad group Force calculation It’s much too complex even if you took all the compute power in the world and you ran it for a million years That wouldn’t be enough computer power to calculate all the variations ingo If children are literally pulled out of their school to spend half of their lives training to master this game I can’t really just dismiss it and say that dota 2 is more complicated although Again that this could depend on your definition of combo. I hate it, but it’s just something to think about on the topic of Deep Mind I imagine that some of you would be interested in a few juicy updates about what deep mind’s alpha go is up to Also, what has humanity learned from that moment when the Go world champion was defeated by an artificial intelligence? Well while playing it’s winning match against the go champion alphago played some very strange moves They ended up giving it the advantage to win the match To be clear in over 3,000 years of humans studying and playing go we’ve never thought of playing the way that alphago did Alphago’s moves have now been used in go schools to train students and expand their way of thinking about how to play the game and Further to this deep Mayan plans on using the alphago algorithm for more general purpose Functions as it shows great promise in its ability to learn and I think of AI is This incredibly powerful tool that will augment human ingenuity and unlock our true potential In fact one way you can think about AI and indeed alphago is like AI think of it as analogous to the hubble telescope A kind of ultimate tool to explore the universe of course the go players alphago Was allowing them to explore their universe of the game of go? And I think there are many other domains in the real world that suffer from this kind of combinatorial explosion that go has Now obviously as I said at the beginning we test our systems on games because they’re the most convenient way to develop our AI algorithms But obviously ultimately we’re not interested in just being good at games We want to translate those algorithms into the real world and be useful and make huge impacts on real world situations And one reason we could believe we can do that is because we’re building general-Purpose learning systems They’re not being handcrafted for the get further. That’s which. They’d like chess They’ve actually we’ve actually built We believe general-Purpose algorithms that could be taken from the games that we test them on and apply to real world and we’re applying these to all sorts of other areas healthcare robotics and even Optimizing Data centers So a variation of Alpha Go we took Over last summer, and we applied it to Google’s data centers, and we managed to save 15% of the power that was used in those data centers by controlling the cooling systems more efficiently. So there you have it. As always it’s some very interesting times in the field of artificial intelligence It seems that artificial intelligence playing games would be a great way to enable it to learn, but further to this, the knowledge learned could be applied in other field making the act of playing games much more important than order first seem. Both Deep Mind and Open AI are going with this strategy. So what’s your view on the story? Are you at all surprised by this or is this all just normal to you by now if you’re a subscriber of this channel? I’m sure you’re pretty used to it. Anyway, that’s the end of the video. Thanks so much for watching. If you’ve just stumbled across this channel feel free to subscribe. This has been – go go you’ll be watching cold fusion and I’ll see you again soon for the next video See you guys have a good one. Cold Fusion, it’s me thinking

100 thoughts on “Elon Musk’s A.I. Destroys Champion Gamer!

  1. Just because we treat a game as more important, does not mean it's more complicated.
    You mention go's large amount of possibilities. However Dota has nearly infinitely more possibilities.
    all the different heroes and all their different amount of damage and range
    all the different items and the actives they provide
    all the stats
    all the ways a fight can play out
    there are so many different things in dota that there just isn't in go

  2. I'll just say that it wasn't a fair fight. bot knew all the possible moves of his opponent, while dendi didn't have a clue what he is going against. so, whoever says that AI is better than human yadda yadda is very VERY wrong.

  3. and i will say elon, leave video gaming alone. you don't need these things in order to improve your AI strategies for the future of whatever your ideas are. such things where others interfere with gaming and specially politics, aren't even on the list of things we don't want and need. let alone the last thing we want and need.

  4. I believe we are creating a god like civilization. Earthlings will one day rule a part of the universe but I'm sure it won't be us humans. Let's hope we don't land in a zoo cage.

  5. 5:02 In every second in a video game you can make 200 choices. Per player. Then raise that by the duration of the game. That's how many different variations a video game has.
    Lets say an average match is 10 minutes. (200^600)^2 = To much for calculator to count. (decisions^seconds)^players. Atoms in universe amprox 10^78. Variations in a 10 minute video game 30^3121 (for 2 players, otherwise it's more)

    Not to be rude but. You are seriously wrong here.

  6. I think it's harmful that we interchange the terms deep learning and AI.

    Deep learning us not harmful, as it must be trained and is tak specific.

    AI is scary, it trains itself and develops it's own goals, it can do just about anything at superhuman competence.

    Deep learning is used to create AI, but they are not one in the same.

  7. How did ColdFusion disagree that Dota is more complicated than Go?

    I read about an article that OpenAi is the best AI up to date. Although it doesn't innovate like Alpha-Go Zero . It actually use their concept but on a much MUCH bigger scale in order to play dota.

    Go is turn based, Dota is Real time . Go has no item build . Dota has. Dota has heroes, has neutral , has roshan , has many objectives from tower to hero kill to capture rune to place ward .
    Dota is team based game . Dota has 100++ heroes. Dota is a game possible because of advancement in technology (Computer and graphic card) . Go is on a 2d plane board.

    Howww? Howwww? Cold fusion How….????

  8. Technology wise, what OpenAi and the Alpha Go bot have done is incredible. From a futuristic stand point, I'm with Musk. We really are opening the gates of hell with these Ai

  9. Ai as strong as it looks is like a cheat as soon as you figure out a weakness the rest falls down in pieces that is until a new upgrade

  10. It's so awesome… but at the same time, I can't avoid thinking about such learning power being injected on an armed robotic system (such as those from DARPA or Boston Dynamics) that becomes a proficient, morally neutral and totally efficient killer.

    This is not sci-fi.

  11. Imagine dedicating your entire life to a board game just to be beaten at it by a bunch of transistors

  12. The reality is that science moves forward relentlessly, and money drives science. Assuming most people who can fund science are inherently Greedy, then Greed will become an inherent (and even intrinsic) aspect of AI. What it does with humans after the point of technological Singularity is anyone's guess. The best might be for it to go on internally calculating until mankind ruins the planet or blows each other to bits… (The answer is: "42")…

  13. Technically it's not Elon Musk's anymore, he rescinded from his position. He was the cofounder though. It would be a mislabel to call it his.

  14. Interviewer: "What do you think about AI?"
    Musk: "AI is really dangerous."
    Interviewer: "Then how do we deal with it?"
    Musk: "Let's see if we can make money from AI."

  15. The scariest idea for me is an AI with human training, followed by self-improvement. It would already learn from the best and then make itself even better.
    Goodbye humans

  16. What I take away from a vid showing AIs doing better than humans in so many areas :
    I feel incredibly stupid.

  17. If a super intelligent AI found out that Earth would be a better place without humans, would it be wrong?

  18. There are definitely more possible configurations of DOTA 2 than there are of go. Things move around with floating point precision not just in large discrete steps. So by that argument DOTA 2 is vastly more complex because there are far more possible configurations. Unfortunately this argument is wrong, and the conclusion is actually somewhat subjective. The point was that the techniques of calculating every possible combination of moves each player can make for the next 15 moves ahead of time, and weighing the best outcome of them with a simple heuristic, is just as possible or useful with go or DOTA 2 as it was with chess. Both games require a different approach to play at a professional level.

    Basically the neural network takes a configuration wherein it takes a lot of data, and spits out an action based on the data. Different data, gives different actions, but it goes through a huge tree of nodes that weight and process the data and affect the end result. To start with the configuration is completely random. Then it "evolves," by pitting millions of these random configurations against each other, taking a few of the best results, and using them as the basis for millions more configurations, then repeating for millions of generations. The end result is a tree configuration that crunches that data into very effective actions. When they say the AI has many human lifetimes of experience, this is how they are getting that number. They don't require the sort of intuition or innovation a human has to get good at a task quickly if you have tremendous processing power to parallelize all of this learning.

    This is great for games because you can afford to have billions of AIs failing billions of times. Supposing you want to teach such an AI to perform a real world task like surgery, you first have to give it a perfect simulation to learn against, complete with every sort of thing that could go wrong during a surgery. That is more difficult than making a neural network AI in the first place. Either that or find a billion volunteers ready to make the ultimate sacrifice for some AI to learn.


  19. The AI can operate directly interfaced as software to the game software, and beat a single player.
    The competition isn't equal unless the AI operates from a robotic self-contained interface using the same physical controls as the human player.
    When that happens we will have much to be concerned with.

  20. I'd love to see these kind of AI for video games. Different versions of the AI have different ranks (e.g. bronze, gold, diamond) and can be used to train a player how to play a ranked game more effectively.

  21. Yep, games definitely need better AI. Not for competitiveness but as companions to enjoy single player games with.

  22. So, I have AI running in my GPU. So using that as a benchmark I would predict that its only a matter of a few years before we see everyday applications of computer learning and AI neural networks in society.

  23. The number of combinations possible in the Dota game-state i'm certain vastly outnumbers Go.

    1. Dota has many world positions with many heroes with varying stats.
    2. From what I can tell Go is just a simple checkered board with 2 colors 1 for each player.

  24. And even we see where we are going with this catastrophe we are not stopping or fighting to stop this right now?!!!!!!!
    How is this not a world wide decision debate for all humans to vote?!
    Why they are allowed to prepare consequences that the world gonna suffer in the near future but we have no vote in ?!

  25. He is screwing the planet more than anyone else, he talks about the people deciding, when do we have a decision? We all know how we would vote, Fuck Space, Fuck Wars, Fuck racism, Fuck the Rich fucking the poor. Peoples lives are fucked by paying taxes for all this political bullshit. As AiI takes over there will not be jobs of worth, minimum wage on universal credit is what is comming. George Olwel's 1984. @ and Adam Curtis's Hypernormalism @ We are all just told lyes. the next world war will be the people uprising againced governments and the rich (populism) and the army and police you supposedly pay to protect you will be fighting you.

  26. Update on OpenAI Dota2: They managed to beat 99.4% of teams*

    *in a tailored version of the game that suits the AI needs and shrunk sufficiently to allow them to focus. 15% of heroes with reduced game mechanics is where AI managed to beat the humans. Humans go into the game at 100% learn, develop and get better managing 100% complexity. After 2 or 3 years of training and development, AI managed to get a grasp on 15% of the game. Most pro teams would beat the AI in 5 games, win consistently in 15.

    Here is the issue explained more in depth:
    AI does not make long term strategy planning – OpenAI can only learn through past 5 min choices but game was tailored to a set of heroes that can pursue short term strategies. They cut the complexity to 17/110+ heroes, they promised 25 heroes but admitted, they could not deliver top 1% player performance for about 20% of the game complexity. The AI also slashed the entire illusion mechanic. The AI bots were MUCH poorer performers in the laning stage than the 1v1 single hero AI bot. Consequently, the increasing complexity lowers their overall skill and at 20% of the game is where AI can beat humans for now, it may grow into a higher percentage but it doesnt seem the current technology will reach the 100% capacity. OpenAI team realized they will not be able to deliver much more out of their AI and acknowledged they will stop pursuing further improvement.

  27. Eventually AI will discover an entirely unknown glitch in GO that allows it to freeze the other player and rack up infinite points.

  28. Go you manipulate variables by hand. In StarCraft 2 you have entire assets across the map under hotkey that can be manipulated almost simultaneously, and this is the case when considering both players are moving units literally at the same time. Then comes the dimension of ground based and airborne means to war. Also, talk about a lazy floating point with Go. The incrementalism of what would otherwise be a square on a board is far FAR more insane. No wonder humans are functionally retarded VS a.i., which inherently can reduce a game like go and chess to numbers. Once you start expanding the complexity, the displacement of the best human talent slows down. You are clearly wrong and musk is actually correct. Dota 2 isn't going to be that far from StarCraft 2. There is a reason the team defeated humans at go first.

  29. If the AI plays against itself, can it ever win? Oh yes, and also the thing bout the tree in a forrest…..

  30. One day someone will be able to steal that AI on that flash drive. A hacker will then edit it and turn it into a virus. And release it to the world. Imagine a virus that quickly learns and adopts.

  31. AI is a double edged sword. On the one hand it could be the next logical step in an evolutionary process. And on the other hand the human race could be writing it's own obituary.

  32. The number of variations doesn't make a game harder or necessarily more complicated (after a certain point), at least from a human perspective, because humans are incapable of calculating any statistically significant number of those variations anyway and so it is, therefore, irrelevant to the human game.

  33. Chess grandmasters never ever said that alpha go seemed like playing a human but the AI machine not only beat a human but the human said it was like playing a human! so much for the go series og machines.

  34. THE ROBOTS ARE COMING THE ROBOTS ARE COMING…….! , just think there's going to come a time when we as a society, will look back and say how did we coup, with out A I !? . The future or one of the potential future's of the world is here.

  35. Humans : Where is GOD ?
    Artificial Intelligence : Well, there is NOW 😈.

    Me : I promise you guys in just 50-100 year we'll have AI. If we and government didn't create AI , then some bad people will create this illegally and use for human destruction. This is better if we create first and take care of this properly.

    Take a screenshot of this comment because this is going happen in near future and you can see this happening today.

    I don't know I am going to be alive until 2080 or more to see what happened, but I don't believe anymore because my brother died in accident in 2018 at the age of 20 year. So, anything could happen to me before 2080. Wish me luck and sorry for my bad english.

  36. Isn't Dota just a simplified version of Warcraft 3? I would like to see programming an AI to play that game against the best of the best. . . Because lets be honest, Warcraft 3 was so micro intensive and difficult many casual players just stuck with Starcraft.

  37. It can’t be bargained with. It can’t be reasoned with. It doesn’t feel pity, or remorse, or fear. And it absolutely will not stop, ever, until you are dead.

  38. DOTA is a FAR more complex game than go. 1) It's played in real time as opposed to turn based, and 2) There are many more variables and degrees of freedom. This increased complexity makes it a more difficult problem for a computer to tackle. The significance of this achievement is that we are starting to see the gap being bridged between what humans are good at and what computers are good at.

Leave a Reply

Your email address will not be published. Required fields are marked *