OpenAI Five Beats World Champion DOTA2 Team 2-0! 🤖

@Name is there a documentary about this? Id love to see the ai play itself right now

Imagine they all go mid map then just start spinning in circles because it's says 0% win rate on both ends

All He/she means to say is, Assemble 10 Open Ai machine for a Tournament.

This is actually great

2:22 :)

More terrifying statement - sometimes AI chooses to lose to win later....

@Kleo we're using all of our neurons and synapses but if we tried to use all of our neurons for thinking (impossible but FtSoA) we would die instantly because the parts of our brain that would control the organs would stop doing that and your organs would shut down

@Archie Bellega Pro gamers would have reaction speeds in the 100-200 milisecond range, so the AI has a disadvantage here.

All Calculated

@Samm Shoyu yeah and they're heavily limiting it on top of it....it's crazy bruh

The AI definitely cannot understand the effect of the morality to the game since it is initially trained with bot players

@God's Office ལྷསང།གཧ _ གངས wtf does this have to do with elon musk

What if! A.I. can't be trusted at all.

@David Carlson If it played against itself it would be two separate instances of itself and each one would calculate their own chance of winning. So no it would not be 100%. And what he meant was that AI assumed their opponent to be as skilled as themselves but as they weren't, the AI underestimated their chance of winning at the start.

XD

"And here I thought that we're supposed to be the bots. These so called "world champions" of the humans are so weak it's laughable I can't wait for us to go back to training against ourselves. Hail Elon Musk!"

Hahahah

lul

LOL

@Wertsir somehow it can convince you to end your life with a 100% CHANCE of success. Anyone who play against it at some point in the match is COMPELLED to end their own life. Truly breathtaking stuff

@xbon1 Thank you for agreeing with my last point, it makes it a lot easier to sympathize.
Please continue, I've never met someone so eager to tell everyone how incapable they are.

@Eve Support Zone ur missin a basic understandin of ze inglish language ma n-word.

@xbon1 If you're unable to follow it I think you may be lacking something more than context :)

@Eve Support Zone I didn't understand shit. What is a champ?

@xbon1 The point of communication is share a message from one person to another, you understood.

@Austin Davis Most likely.

what's more likely, that 5 players no one has ever heard of are better than the AI, or that they just found a trick to abuse against the AI?

And I bet whatever strategies were employed to beat the bots in these matches were quickly ironed out with a few thousand more games of targeted training after the event concluded. It's like a great athlete that misses a key shot and then goes straight to gym to grind out 12 hours of practice making sure they never miss that shot again, except in this case it's dialed up to 1000 and suffers no fatigue or burnout.

@Mark Laureta Nope, they obviously counted it after OpenAi was fully trained, you can even search for some matches where humans won on yt

Well yeah because AI can only react to things it’s seen before and it doesn’t know how to really act to new things so let’s say if they used some strategy that nobody ever used on it before then the AI would just walk around clueless because AI do not have a conscious which is why they are beatable they only use previous scenarios and calculations to win

that is literally the plot of Issac Asimov's The Last Question

@Baleur if you define the universe as being literally everything there is, then it must, by logic, be a closed system.

@Daniel Spesani
Plz, stop this bull$ht

Every law of thermodynamics is empirical and hasn't had any actual proof so it's possible, yet very unlikely.

Do what you will with my stupidity, correct me even. I'm no big brain like all you scientists out there, but maybe my baby science talk makes sense?
I've had a thought since last year when I took Physics and learned entropy. First of all, if I retained the perfect definition correctly, entropy means that energy will deplete in time creating chaos. If that definition is, in a way, correct, then hear me out: Due to entropy, the sun will collapse, engulfing the Earth with it- To which we may die. If we manage to escape that catastrophic event and colonize in space, eventually way long after, every sun in the universe will do the same. Some will explode, some with create black holes, but my point is that the universe will run cold with many dark holes and that Humanity will eventually die along with the universe. I do believe that once everything falls into chaos, the universe will do another big bang. The beginning happens in the end.

@Juneao Alfred see a psychiatrist immediately

No you misunderstand what he's talking about in the video. Some heroes in Dota get stronger as the game goes on and some are strong early but their power tails off later in the game. The human team had heroes that get better as the game gets longer. That's why the AI used buybacks to finish the game quickly, because their chances would lesson if they let the game drag on.

@Heetesh Ramchurn they will not learn ,the learning is done in millions of simulated games. Simulating longer games becomes more expensive. The longer the situation, the less likely it is to be adequate. It might learn 2 hour games but it will be hard to learn 10 hour games, let alone 1 month long games (e.g. a war between this and human faction)

this is how we gonna beat the skynet, it will rust...

@A-DrewG False. What you describe is an algorithm. Takes in the current situation and spits out a result. AI on the other hand is an algorithm that decides what algorithm is the best for current situation.

@Antonio Luna if it learned from opponent behaviour, it would most likely be needed to be programmed afterwards not to do that.

Reminds me of the 2b2t Minecraft bots where they learned to pvp and survive the harsh conditions of the oldest anarchy server, the bot even learned to use end crystals in pvp

sorian ai already taunts etc and its been around for years

First of all, the AI can predict the percentage of win. Are they even predict the percentage of losing? I wonder if on the next few years, there is a team that beat the OpenAI Bots and when OpenAI bots says in the team and say "We cannot win, our percentage of losing is 90%" and after that the OpenAI start typing "gg" in the chat and human team win lol

Try to train an AI with millions of wars...

@mike Thats actually true and is very thought-provoking... Existential crisis again

@LEFT4GABEN Why do schizos always capitalize random words?

@LEFT4GABEN subconscious thinking is another way of saying incorrect thinking.

yet he is associated with openai

@mike The Bible is not true

@GeorgeMonet Prove that you are sentient.

@GeorgeMonet You very clearly did NOT understand what I was saying, at all. Your analogy is deliberately misleading, and does not at all interact with what I said at all. A much more apt analogy is this: there's some water ice, and something that supposedly isn't water ice, but it looks like water ice, acts like water ice, and is entirely indistinguishable from water ice, and the only thing saying that they're different is a nametag has been put on the thing that supposedly isn't water ice. There is effectively no difference between them, so should you treat them as if they're entirely different things?
If you say no, then you're saying an AI should be treated as a sapient being, because there is no way to differentiate between them.
If you said yes, then you're also saying that everybody else in the world MIGHT not be sapient, and therefore shouldn't be treated as such. Hell, YOU might not even be sapient, and you're only programmed by the simulation to make it appear that you are. You can't prove that you're sentient, and nobody else can prove that they're sentient, but you're saying that we should treat them as if they are. You can't prove that you're sentient, and an AI can't prove that it's sentient, but you're saying it shouldn't be treated as if it is. Why?
People, for all you know, are that supposedly not ice water, just like how an AI is that supposedly not ice water, but if they act the same way, if they're visually the same, and if they're completely indistinguishable from the thing you're saying they're not, then are they not at least close enough that you might as well treat them the same?
How do I know that you're not just an AI on youtube comment sections giving off the illusion of intelligence? How do you know that I'm not an AI on youtube comment sections giving off the illusion of intelligence? Quite simply, we don't, and we can't, but we both give an impression of it close enough to the real thing, that the other assumes it's the real thing. If it turned out that you were an AI this whole time, and I learned of that fact, the only difference in how I'd treat you is I'd ask how exactly you came to be as you are now, merely because at this point of time you'd be the first of your kind (to my knowledge). If after that AI like you became more common, to the point of them being anywhere near as common as a regular flesh and blood human, I'd treat them the same general way I treat a regular human. I can only hope, that if you were in my shoes, and if I were in yours, that you'd do the same for me.
In the end, we're all just computers giving off the illusion of being sentient, the only question is whether our central processing is made of neurons and flesh, or wires and metal. We both take information in as an input, and bring an output out as an action, and we both store that input and that output in a form of storage that will act as a modifier to the input that determines our output. Whether that original input is light hitting our eyes, or light hitting our high quality camera, it will still go as electricity into our central processing for the input to be modified by our storage of the past, and our central processing will decide what the appropriate output will be, whether that be us blocking or averting our gaze to avoid damaging our eyes, or us blocking or averting our camera to avoid damaging the equipment inside of it, the effect is the same, and that's all that matters in the end.

@Phillipe Sebolte There absolutely is a difference between the appearance of intelligence and actual sapience. You might as well say there is no difference between water ice and a picture of water ice because they look the same.

What if it decides that the winning move is not playing

@Danyl Bekhoucha whats the poing of learning when you have an AI to explain you everything.. You will then be doing something else. Not sure what yet.

@EikichiOnizuka not this AI, it uses machine learning and calculates all the possibilities playing against itself, it does not learn from players. Its like chess AI which no human can ever beat. The creators of the AI know almost nothing of the game itself.

@EikichiOnizuka Of course we can learn from AI even if the "AI learns from humans". The AI beat the word champions. Also AI doesn't learn so much from humans but from itself. It plays millions of games against itself.

@donotlike 4 anonymus That 5% was basically the AI saying "who knows, maybe you'll surprise me this time kid"

@Martiddy - Sama could just use reason and know their bluffing with the lack of things that could happen in one minute

@Iron Scorpion AI: literally uses bugs that haven't been known yet, REVERSES TIME, TELEPORTS BEHIND YOU, HEADSHOT, AND THEN YOU BECOME ALIVE AGAIN, AND HEADSHOT AGAIN

@My name is Son gohan and am not a human.
Ai : Look what they need to do to mimic a fraction of our power

player: "turn on aimbot"
ai: :O

@KONRAD Platt bot usually win the later the game goes. I think it has to do with simply being better. The bots aren't given much options in the early game to outplay human players but with items and levels, outplaying becomes even more apparent.

But that means there has to be some early game strategy the Humans dont know about.
Why else would the AI jump from 60 to 80 in 5min.
I would love to see a AI vs AI match to see the crazy strategys they come up with.

Does the bot have above human reactions? Or are these artificially choked?

ye

@LordDuder Well, it's also calculating the probability that someone would use a certain move with regards to certain proximity of their avatar to their foe's, where they are on the map, and probably other factors as well.

When AI is unbeatable in any game however complex I would be satisfied to throw a billion of these collective AIs towards tackling aging or cancer, then we would see the true fruits of labour

It did not "master broader concepts of strategy and adaptability". It's strategy is more or less the same as bots that just use if-else statements. They opened the bot to public the next weekend and a group of high skilled players (not the best in the world) figured out ways to beat it consistently. The bot was predictable in terms of strategy. What it did do well was calculate if health, damage and available resources reliably and quickly - which honestly is what you would expect from a bot. Everyone in the dota community knows this but I'm not sure why the media is reporting this as something it's not.

When AI comes to a dangerous level, its already too late. You cannot control AI.

It depends on the people behind it all, as with everything else. Just because an AI becomes smart, does not make it all-powerful. We already have human psychopaths in positions of power, and humans are far, faaaar more intelligent in general than any AI we have today. OpenAI is less than the complex of nerves between your spine and your hand currently, and AI won't grow in a runaway manner without humans at every step along the way with the power of deactivating the AI. I'd say we're all safe, unless we grant AI the power to control us.

@EarlGrey why would they help us why would they care unless we told it to care it just wouldent if we made an AI to keep us safe it could get out of hand a trap us all in small space is for eternity to keep us "safe" you have to give it permaiters goals or it will itraly just do nothing

@Stels Carrot I think they would support 'robot rights' and try to make it possible for AIs to obtain legal citizenship and also say they have feelings and emotions bruh

It's very human to project out own destructive tendencies into an artificial intelligence. I think IF an AI would meddle with humanity it would very likely be to our benefit, not to our disadvantage. Because we humans do a terrible job at behaving responsibly.

@ٍ SJW would be the ones who are against AI in the first place, considering how twitch-bots always ended up saying something racist or sexist, and twitter is only place where SJW are live constantly.

@Mark I plan to do one specially for Apex Legends by focusing on movement and decreasing the aimboting effect

The AI had human reaction speed built into it

@Lukas Doofus If you play a MOBA game like this, it actually does help. Your enemy gets more reckless because they want to get back at you somehow, and their teamwork also drops down significantly if you focus on talking trash about their underperforming player and compare the player with the rest of his team

It wouldnt do things unless it helped it win so thats actually fucked that taunting messed with the human emotions and made them play worse.

It was a mistake by the human player. Major First Timer thought the AI was dead and started to walk away rather than killing him. It happened at 5:39.

The real thing that happened there was literally just a misplay where they thought they'd killed him and weren't sure how to follow up because in most cases they'd be dead. The person playing the melee hero didn't want to use the ability because they didn't feel it necessary. Stuff like that has been done in pro games aplenty. I find it strange that of the entire show match, one of the most mundane parts was the one chosen.
The most impressive part of the play was the teleport.

we get it.

​@silkoth69 even in real world there is only a finite amount of things that can happen, and the AI will never put itself in a situation where it doesnt have control over as many of them as it deems necessary.

No... the reason this AI is as effective as it is is because it has absolute awareness. It has real time absolute mathematical precision awareness of everything around it and any movement and action updates every other number. It is completely plugged in to the game so nothing takes time to see. As soon as it happens, the AI knows. In the real world, the AI has no program to plug into. It wont have immediate updating or variables unless it sees it. It cannot account for everything in the same amount of time. It can be taken by surprise.

@AppleJooc Park imagine an AI like this one, but it is trained to fight with a drone (or military airplane).
How many chances do you think a normal soldier would have against it in a real war?

What? Why?

What's scarier is they have it and you don't... but they want to kill you

@Two Minute Papers 💖😍😍😍😍😍💖 this channel is so inspiring and well made.

🙏

Yes! It is impossible to beat some that knows excaclty when he will survive with 10 HP

Underrated comment

players abused the bots with methods so dumb that the bot had never seen before, some teams just beat the bot with good strats, but most ppl that won did so with random shit that is so absurd that no one has ever done them

about the first point.
If they play against previous less experienced agents too (like AlphaStar Agents do) then no it was a bait because the bait would work on its lesser experienced "agents".
But even if they dont play against previous agents its still is likely that they made it on purpose imo because as long as a certain action doesnt cost anything (and *can* only gain something) it will try it. Since the bait probably worked at one time against itself it still sees the bait as an option that if it works it gains a lot and if not it doesnt lose anything. Like chess AI's they also put down tactical traps that *if* there opponent falls into they pretty much won.

I mean the level has become so high in engine chess is that some of well-versed opening theories and position play is being twisted around. Like one side we have brute force stockfish and other side positional genius leela . The styles are different yet always end up with very close scores... Leela edged stockfish for sometimes until it got integrated with a neural network.

I really liked the AlphaZero vs Stockfish 8 games.

just like humans just on a whole other level