Slashdot Mirror
Elon Musk + AI + Microsoft = Awesome Dota 2 Player (theverge.com)
An anonymous reader quotes the Verge:
Tonight during Valve's yearly Dota 2 tournament, a surprise segment introduced what could be the best new player in the world -- a bot from Elon Musk-backed startup OpenAI. Engineers from the nonprofit say the bot learned enough to beat Dota 2 pros in just two weeks of real-time learning, though in that training period they say it amassed "lifetimes" of experience, likely using a neural network judging by the company's prior efforts. Musk is hailing the achievement as the first time artificial intelligence has been able to beat pros in competitive e-sports... Elon Musk founded OpenAI as a nonprofit venture to prevent AI from destroying the world -- something Musk has been beating the drum about for years.
"Nobody likes being regulated," Musk wrote on Twitter Friday, "but everything (cars, planes, food, drugs, etc) that's a danger to the public is regulated. AI should be too."
Musk also thanked Microsoft on Twitter "for use of their Azure cloud computing platform. This required massive processing power."
"Nobody likes being regulated," Musk wrote on Twitter Friday, "but everything (cars, planes, food, drugs, etc) that's a danger to the public is regulated. AI should be too."
Musk also thanked Microsoft on Twitter "for use of their Azure cloud computing platform. This required massive processing power."
Let me repeat: computers playing games is NOT AI. Computers love games. Games have strict rules and limited parameters. Computers love that. Computers excel at that. IT IS NOT AI.
"Nobody likes being regulated," Musk wrote on Twitter Friday, "but everything (cars, planes, food, drugs, etc) that's a danger to the public is regulated. AI should be too."
So should you, you hypocritical, delusional bag of gas. I just can't take this douche seriously anymore. And whoopty-do. AI can play games. Let's elect it mayor.
...I am beginning to ask myself, whether weak AI like this (no actual intelligence or understanding) may not actually be on-par with many humans, which fare not much better at understanding things.
Most ACs are not even worth the keystrokes to insult them. Be generically insulted by this and ignored otherwise.
I'm right here. What do you have to say to me, snowflake?
Ok, Elon. How about if we start with "don't teach your AI that it's primary objective is to destroy every other creature on the map".
That's our word. You can't use that.
Wah, sorry I triggered you snowflake. Maybe your hood is on too tight, cuck.
AI basically means tuned specific savant agents. They can do that 1 thing very very well. There is a place for those sort of programs. To call it 'artificial intelligence' completely distorts what they are. I can not tell the bot they created in this case to go learn SMB and drive a car then grab me a gallon of milk later and whatever else I want and it still be any good at the original task. Instead I have to spend tons of time training the net and I get very optimized decision trees. If you hand it something it does not know it does not say 'hmm what does that mean'. It instead gives a wrong result and goes along its merry way. It has 0 self awareness to even know it got the wrong result.
So like the term 'hacker' turned into maker we need a new term for AI. It has been co-opted by people who should know better and the news ate it up to spoon feed back to us.
like a 12 year old binging on red bull and hot pockets?
Did the bot have direct access to the game state or only the frame buffer and audio? I'm assuming the former, the latter would be much more impressive.
first time artificial intelligence has been able to beat pros in competitive e-sports
If Dota2 is going to be called a form of sport, surely Chess is too. Computers were besting the top pros in Chess in the 90's.
When words lose their meanings, every statement is wrong.
That's because MightyMartian and 01010101010 are worthless SJW wimps who are always trying to invent problems and feign outrage over everything.
Yes you are otherwise you wouldn't have replied.
Sincerely,
Not the other guy but still cracking up about stupid right wing snowflakes
How about we start by teaching it the difference between the side of a truck and a sign on an overpass?
(multiplayer online battle game called...) Defense of the Ancients 2
/. summary and linked article.
... Because some of us aren't au fait with computer game trends but still want to understand the
... giving computers full tactical control of potentially dangerous equipment/systems. The game is changing, no matter what name it's given.
It's not that the equipment is any different, nor that computers are already in the loop, nor even that software can be subverted, but that the job given to the computers is a whole level up in decision making.
We've repeatedly seen what happens to a chat bot that is left to learn on its own. It has no compass and goes bizarre.
BILLIONAIRE Elon Musk has claimed that people should be more worried about Artificial Intelligence (AI) than the threat posed by North Korea. http://www.express.co.uk/news/...
It was obvious that e-sports will be short-lived because bots are going to beat us all within a short time. Soon you'll be able to run them on your home machine and nobody will be the wiser. Anti-cheat mechanisms will work for a short time, and then go to the dustbin of history.
But what I'd find even more interesting is the output of those learning algorithms. If it can amass lifetimes of experience, can it answer the question which heroes are over- or underpowered? All we'd need to check is its picking preferences. Can it figure out if there is one optimum loadout? One optimum skillset? Will it find an optimum strategy, especially once the whole team is bots (who know that the others are bots, or even a network of bots that communicate) ?
Using such bots not at tournaments but during game development will do miracles for balancing.
Assorted stuff I do sometimes: Lemuria.org
"Musk is hailing the achievement as the first time artificial intelligence has been able to beat pros in competitive e-sports"
I've been playing multiplayer games for decades, and cheat bots have always been able to beat humans. There is no way you can compete with something that is programmed to headshot you the millisecond any pixel of your player appears in their field of view (which also happens to be 360 degrees all the time).
Maybe this claim only applies to DOTA, because for any FPS this has always been the case.
The player has to use a mouse and keyboard for output which is not the ideal interface between a brain and a computer game. For input the player use a screen which shows him a limited part of the play area whilst the machine gets to "see" the entire play area at all times and lets it do it in a way that is native to it.
Let the machine use a physical keyboard / mouse and an actual screen and I might be impressed.
I think the question must always be asked when people call for regulation is can they profit from it? I can't help but wonder if Musk is actually not doing good here but setting in motion the slow train of regulation so that when he's ready with some uber-AI he'll be in the position to get the rules of the game altered to his advantage with bureaucracy, regulation and licensing.
I have about 5000 hours on this game so its fair to say I have a decent grasp of the mechanics. Dota is a team based game where you have to destroy enemy buildings to win the game. Its a 5v5 game and it has a pretty diverse strategic pool with every game being different to the previous.
This feat whilst impressive is still being sold as more than it actually is. The bot was created by coding in the rules of the game (killing monsters or destroying towers gives you gold, win conditions etc. that sort of stuff) and having a bunch of bots compete against each other and applying machine learning to the events that happened to be able to find strategies that are successful.
The problems arise with the fact that they only implemented rules for a subset of a subset of the game (a specific 1v1 matchup instead of any of 100^10 5v5 matchups). Also, due to the nature of how the bot works, if you do anything that is weird (like ambushing the bot at the start of the game but not doing anything to it) then it will pretty much just kill itself for you. There is also an extremely unfair advantage to the bot in terms of perception, the bot doesn't use the rendered screen as input, it reads all the data out of memory. A human can only see what the camera is looking at whilst the bot can see you across the map as long as you aren't in fog of war.
If the bot had to play using the rendered screen it would never win, 100% guarantee that it loses every game even to some of the worst players of the game.
From a facebook post about AI "I still wouldn't trust it. It's like cloning. It's something we shouldn't be messing with." and the classic " If you were a computer, and had the ability to, why wouldn't you exterminate the human race? We are hip deep in fucktards".
Fat, drunk, and stupid is no way to go through life, son.
If the AI learned anything it would be how to use racial and sexual slurs in Peruvian whilst smack talking the entire time ultimately quitting the match should any of the other "parameters" upset them...
A computer learning how to win a computer game doesn't impress me much. Had it learned gamer behavior that would be something to see! :)