Microsoft Shows Off 'Milo' Virtual Human

adeelarshad82 writes "At TED Global in Oxford, Microsoft released a video showing off its 'virtual human' technology, named Milo, designed for the company's hands-free Xbox 360 motion controller called Kinect. Milo is built to react to people's emotions, body movements, and voice, allowing players to interact with the virtual character. It was built using artificial intelligence developed by Lionhead studios, along with undisclosed technology from Microsoft. According to games designer Peter Molyneux, the game exploits psychological techniques to make a person feel that Milo is real. Each Milo character will be unique because every player's interaction with the virtual character will sculpt the type of virtual person Milo will evolve to become."

Comments

Frightening

[Fwipp]

Honestly, I don't know whether this is the Uncanny Valley manifesting, but that kid just creeps me out.

Re:Frightening

[fuzzyfuzzyfungus]

Don't worry, the "undisclosed Microsoft technology" will allow him to whisper 'why won't you... Love ME?' into your ear just as you are on the verge of falling asleep...

They want to be sure that none of their next-gen hardware survives user-inflicted destruction long enough to RRoD.

Re:Frightening

[HermMunster]

If Milo can't think for himself then he's nothing close to a virtual human.

Re:Frightening

[SilverHatHacker]

Why won't I love ME? Because it was a terrible operating system. Why do you ask?

Re:Frightening

[Stan Vassilev]

Honestly, I don't know whether this is the Uncanny Valley manifesting, but that kid just creeps me out.

That's not the Uncanny Valley causing it. It's your basic animal fear that your own species has just been made obsolete by some basic gaming software.

But I might be wrong. You can test it. Get the DVD of Pixar's "The Incredibles" and play some. Do you feel the impending doom, or you're entertained? If the former, it's the Uncanny Valley (also check that with your doctor).

As for the fear-inducing-AI in the demo: it's fake, so you'll be able to sleep well tonight.

Re:Frightening

[negRo_slim]

It's your basic animal fear that your own species has just been made obsolete by some basic gaming software.

Do you even listen to yourself?

Re:Frightening

[Lord_of_the_nerf]

Microsoft wouldn't give Milo the capacity to think for himself.

If he could, there's the chance he might pick Firefox.

Re:Frightening

[hairyfeet]

You know, I still think Gates owes all of us that was stuck with that turkey an apology and a copy of XP Home. Where is my apology you rich bastard!!! Just as I think the Ballmer monkey should do the right thing and say he was sorry for Vista and give those stuck with that loser at least Windows 7 Starter just to say he was sorry. WHERE IS MY APOLOGY YOU FAT BASTARD! And WTF is it with /. and the Gates borg? The man hasn't been head of the company in years. Shouldn't they switch to a chair throwing Ballmer Monkey by now?

As for TFA, I can see some real world applications for this tech. Imagine a virtual teacher with infinite patience for those with learning disabilities. you could use this tech to tailor a learning plan for the individual without the incredibly high cost of private tutors. Done right I could see this really helping kids.

Re:Frightening

[pspahn]

Haha, give autistic kids a robot for a teacher, that'll sort them out. LMAO!

Re:Frightening

[Anonymous Coward]

The purpose of ME was to make XP look good, just as the purpose of Vista was to make 7 look good.

Re:Frightening

[Yvan256]

He drifts in and out.

Re:Frightening

[flyingsquid]

If Milo can't think for himself then he's nothing close to a virtual human.

But he would be part of Fox News' target demographic.

Re:Frightening

[Yoozer]

Imagine a virtual teacher with infinite patience for those with learning disabilities. you could use this tech to tailor a learning plan for the individual without the incredibly high cost of private tutors. Done right I could see this really helping kids.

And it'd be an awful shock for those kids if they went from their safe school environment back to the real world where people don't have infinite patience.

Re:Frightening

[enter to exit]

Gates will always be the face of MS, just as Colonel Sanders is the face of KFC.

albeit - Sanders keeps getting younger every corporate image respin.

Re:Frightening

[56ker]

The truth of the matter is; although we're just watching a video. Personally, I'd prefer to go down to the park, river or sea and go fishing; interacting with a real person rather than a fantasy boy.

I was born at the start of the 1980s. I got my first computer (Commodore 64) for my 7th birthday. I've spend a lot of my life in front of a screen, growing up with video games as a pastime, then working in computer troubleshooting and website design.

I also play music professionally. The latter I prefer as you get a reaction from an audience, whether singing or clapping. A computer can't provide the human touch yet, but artificial intelligence has come a long way from the robot dogs, a robot violin player and Sony's robot that can walk.

Sooner or later, we'll move beyond a computer's AI capacity being like a child or baby. What will we do when artificial intelligences have a similar neural capacity to humans? Will we treat artificial life as comparable to human; or continue to see human life as more important?

More of society's decisions are being taken by machine, not by people (although people programmed the machines). Where will it all end? Will we have robots like those that Isaac Asimov described with laws drummed into them not to harm people? Will we explore space, the universe and the deep ocean with artificial intelligences in places we can't go? Will we put machines at work to come up with cures to cancer, diseases or social problems eg poverty or famine?

Or in the end would we prefer our societies to be governed by people as our political systems have for generations? What will we do when artificial intelligences get physical bodies and they can pass the Turing Test?

All these are things that may come to pass in my lifetime; alternatively we may just screw things up and make the human race extinct by complete ecosystem collapse (or at least enough that the human race is made extinct).

That's why manned colonies on the Moon and Mars are essential; as a failsafe in case Planet Earth should face a major disaster (meteroid strike, global warming, biodiversity problems, problems affecting human fertility).

Re:Frightening

[delinear]

Maybe he needs to leverage some of the plugins that aren't currently available on other browsers, or maybe he needs to be reasonably sure his browser will be supported and the only way to do that and not support IE is to support the next biggest desktop browser by user volume. The likes of Chrome and Safari might be better, but that still doesn't mean they're the right choice for everyone just yet (for me the tools I need for web development are far better in FF right now, for instance - other browsers can replicate some of the functionality I need, but not all).

Re:Frightening

[delinear]

I love the irony of your comment and your sig:

If Milo can't think for himself then he's nothing close to a virtual human.

--

You can lead a man with reason but you can't make him think.

Plenty of actual humans can't/don't think for themselves, so why is it a necessary requirement for a virtual human?

Re:Frightening

[Bing Tsher E]

Why a robot? A simple tape loop that plays 'cut the shit, you little creep' would work.

Coddling parents would have to be kept away with steel doors, of course.

Re:Frightening

[vegiVamp]

Gotta be honest, most real humans seem to have the same bug.

Re:Frightening

[ultranova]

Honestly, I don't know whether this is the Uncanny Valley manifesting, but that kid just creeps me out.

Yes, it's pretty creepy. And if people actually "connect" with it as the video claims, it's going to be horribly jarring when you run into the limits of the physics simulation and AI. Which you will - XBox doesn't have the kind of power to run either convincingly.

There's also the little bit about there being no actual game to be seen, but let's be generous and treat this as a tech demo.

Re:Frightening

[ultranova]

Sooner or later, we'll move beyond a computer's AI capacity being like a child or baby. What will we do when artificial intelligences have a similar neural capacity to humans? Will we treat artificial life as comparable to human; or continue to see human life as more important?

That kinda depends on whether it's possible to decouple intelligence from personhood. In other words, does having a robotic car that's smart enough to get a sixpack of beer on the way to pick you up from work because it previously brought your mother-in-law from the airport mean that said car is a person?

Will we have robots like those that Isaac Asimov described with laws drummed into them not to harm people?

Hopefully never, since each and every Asimov's robot story is about how these laws fail miserably. A rigid adherence to a set of rules is simply not sufficient ethical framework for an intelligent creature.

Or in the end would we prefer our societies to be governed by people as our political systems have for generations?

Does it really matter? I have no say in how my society is governed; I doubt very much you do either. Maybe our current set of overlords are supplanted by a new one; maybe they won't. It doesn't make a difference to me, and I have no power to affect that either way.

Re:Frightening

[56ker]

>I have no say in how my society is governed; I >doubt very much you do either. I do through voting and standing in public elections. I did hold public office for two years. This year I got 699 votes (roughly 1 in 6 of people voting). Although some politicians ignore me, I do have influence in affecting decisions. Unless you don't live in a democracy ultranova; in which case why not change your society to something that gives the public more of a say?

Re:Frightening

[Thing 1]

You know, I still think Gates owes all of us that was stuck with that turkey an apology and a copy of XP Home. Where is my apology you rich bastard!!! Just as I think the Ballmer monkey should do the right thing and say he was sorry for Vista and give those stuck with that loser at least Windows 7 Starter just to say he was sorry. WHERE IS MY APOLOGY YOU FAT BASTARD!

Not exactly the apology you were looking for, but: Kevin Turner, COO, recently called the iPhone 4 "Apple's Vista" -- so at least there is some acknowledgement of their flaws.

Re:Frightening

[Drakkenmensch]

I'll start worrying if Milo goes Dorian Gray on me and starts growing old and decrepit while I stay young and fresh.

Re:Frightening

[xmousex]

Also theres no firephp equivalent for safari, opera or chrome that I could find. firebug itself in ff is far more detailed then similar tools found in safari, opera, and chrome. debugging complex webapps without firefox is just stupid right now. I'm always watching for better tools, and always happy to try out new things. So far nothing out there available gets to the bottom of issues like firefox with a good collection of development plugins.

Re:Frightening

[Mephistro]

"And it'd be an awful shock for those kids if they went from their safe school environment back to the real world where people don't have infinite patience"

Following the same logic:

"And it'd be an awful shock for those kids if they went from their safe school environment back to the real world where people won't teach them anything"

"And it'd be an awful shock for those kids if they went from their safe school environment back to the real world where performance and productivity can be a matter of life or death"

"And it'd be an awful shock for those kids if they went from their safe school environment back to the real world where people may harm them"

...

So kids should never leave the school -alive, that is. At the same time, we also fixed the unemployment problem. :)

Re:Frightening

[Mephistro]

Will we have robots like those that Isaac Asimov described with laws drummed into them not to harm people?

Hopefully never, since each and every Asimov's robot story is about how these laws fail miserably. A rigid adherence to a set of rules is simply not sufficient ethical framework for an intelligent creature.

I always understood that Asimov's Three Laws were boundaries inside which the robots could act, but the decisions themselves were subject to limitations due to lack of information and lack of processing power -Does It sound familiar?. Thus, more intelligent robots could make better decisions, up to the point that they could create by themselves a new law, -i.e. Daneel R. creating the 0th Law.

Or the guys that created the Universal Declaration of Human Rights.

In one of his books, Asimov wrote that any man who followed perfectly the Three Laws would be considered a saint, or words to that effect.

Disclaimer: I don't know whether the "Three Laws" can ever be "programmed" or not.

Re:Frightening

[MoriT]

Speaking from experience, the real world is a cake walk in comparison to current schooling. For one thing, physical assault in the workplace isn't a common occurrence and sexual harassment laws exist.

Re:Frightening

[ultranova]

I always understood that Asimov's Three Laws were boundaries inside which the robots could act, but the decisions themselves were subject to limitations due to lack of information and lack of processing power

No, and that's precisely the problem: Asimov's robots don't have any kind of ethical - or any kind of - framework to base their decisions on besides the laws. As the result, Daneel ends up imprisoning Baley at one point, because he thinks he might be psychologically stressed by going outside, and the First Law (don't harm humans) trumps the second ("obey humans").

Thus, more intelligent robots could make better decisions, up to the point that they could create by themselves a new law, -i.e. Daneel R. creating the 0th Law.

It was Giskard who created the 0th Law, and promptly destroyed the Earth because he judged that to be in humanity's interests.

Re:Frightening

[Golddess]

Someone should update wikipedia then.

With R. Daneel Olivaw he created the "Zeroth Law of Robotics," an extension/modification of the Three Laws of Robotics, which helped him prevent Kelden Amadiro's destruction of the Earth—though the same law led him to allow a gradual destruction of the Earth through radioactive increases, which would encourage the colonization of the Galaxy.

Re:Frightening

[flowwolf]

Robots can do all of that as well. Hell, they ARE the cars, and machinery. They don't need to walk.
Software can interpret tones in speech, body language, facial expressions, etc.. almost on par with some humans.
The idea that people don't think for themselves, is rather true. They won't form an opinion and won't make choices until someone advises them on what to do. Haven't you ever had that little sister or gf who won't ever decide for herself? It really is a common problem. Many people get in life, solely by following instructions of those around them. The only thought of their own is "who can I get to decide for me?" and even that was probably implanted by somebody.
This is what Milo is doing. A programmer has instilled in him directions on what to do given known circumstances. The only difference is Humans can be told what to do much easier than Milo can be.

Re:Frightening

[xtracto]

Spot on.

I've been working on "hard" AI research for some time. I am also interested in game AI, which is completely different than research AI.

In my opinion, the AI that the majority of people is waiting for will be achieved by game developers or similar technology and not by AI research.

Microsoft Milo shows how feasible it is to make an animation which looks *very* "intelligent", "emotional" and responsive and can interact in a very natural way with people. But that does not mean the program is highly "intelligent".

On the other hand, anyone who has talked with the average Myspace girl (in person) realizes how they can be very emotional, responsive and "interacting" but at the same time they won't have a high intelligence.

It all boils down to what a teacher asked my group on the first day of the Bachellors' AI course: "what is intelligence". There are several definitions of "intelligence", I think the main capacity which should be synthesized before computers are really intelligent is the capacity of abstraction.
 

I want to blow his head off with a BFG 9000.

[EWAdams]

He's a video game character. I don't want him to be real. Him being real would miss the point entirely.

Re:I want to blow his head off with a BFG 9000.

[badboy_tw2002]

Exactly. There's no room for realistic video game characters. Only cartoon-like action warriors shouting one-liners as they mow down Nazis please!

Re:I want to blow his head off with a BFG 9000.

[arth1]

Exactly. There's no room for realistic video game characters. Only cartoon-like action warriors shouting one-liners as they mow down Nazis please!

Contrary to your desires, the nazis were very real.

Re:I want to blow his head off with a BFG 9000.

[Your.Master]

So are warriros and one-liners, but the gestalt isn't.

Re:I want to blow his head off with a BFG 9000.

[flimflammer]

Did the Sims miss the point entirely? I would think the popularity of a game built on sentient somewhat realistic characters would prove otherwise. All he is, is a 3D chat bot that reacts to you physically instead of what you type. I'd hardly call it a video game, personally.

Re:I want to blow his head off with a BFG 9000.

[EWAdams]

Did the Sims miss the point entirely? I would think the popularity of a game built on sentient somewhat realistic characters would prove otherwise. All he is, is a 3D chat bot that reacts to you physically instead of what you type. I'd hardly call it a video game, personally.

The characters in The Sims are not remotely realistic, and thank God, because players spend most of their time trying to find interesting new ways to kill them off. Will Wright quite intentionally built torture into the game. (Locking them in a room with no food or water and watching them die is just the beginning.)

What could possibly go wrong...

[fuzzyfuzzyfungus]

Can you imagine being the poor bastard at Lionhead responsible for making sure that these "virtual humans" can exhibit realistic suffering responses to griefers, gropers, and every other ghastly atavism that the Kinect users of the world will allow to roam free when they know that there are no rules and no consequences?

(Incidentally, I bloody well hope that Lionhead has had some time to learn a thing or two since Black & White. The "AI" in that game managed to suck every ounce of joy out of being a malevolent deity, something that I wouldn't have believed possible.)

Re:What could possibly go wrong...

[sg_oneill]

What are you talking about. The AI was awesome. Once I started rewarding my ape for Lighting its shit on fire, and also rewarding it for throwing its shit at villagers, it put 1+1 together and innovated unholy projectile flame turds.

I only wish my horse on red read redemption would come up with that instead of "jumping of cliffs and dying" :(

Re:What could possibly go wrong...

[delinear]

Oh good lord, I think you just opened up the market for everyone to have a personal copy of their favourite celebrity to "interact" with.

Re:What could possibly go wrong...

[fuzzyfuzzyfungus]

You lucky bastard.

No matter how many times I tried to teach my creature how to inspire belief through classic "good cop/bad cop" techniques, he never learned how to set the villages children on fire, throw their burning bodies at the village, setting it on fire, and then put out the fire with magical rain.(since the villager AI model rewarded you with more belief for giving them things that they needed, you could get more belief per unit manna by hurting them, and then magically repairing some of the damage, than you could by just helping them twice. Totally fucked up; but actually seems to be a pretty accurate model, given how things like abusive relationships, hazing, and just about every major religion, actually work).

Dreamcast

[lyinhart]

I think I'll stick with Seaman for the Dreamcast, thanks.

Re:Dreamcast

[Anonymous Coward]

I think that seaman will stick with you

Re:Dreamcast

[tepples]

I think that seaman will stick with you

No, that'd only happen in Soviet Russia. And the only popular video game that I can think of that came out of that regime was Tetris.

Re:Dreamcast

[Yvan256]

I think that seaman will stick to you.

FTFY

Wait, I've seen this before...

[bobetov]

Microsoft Bob lives!

And takes up the awesome responsibility of being the latest hyped MS product to utterly fail. Sheesh.

Re:Wait, I've seen this before...

[mysidia]

Yep, that's gotta be it.

along with undisclosed technology from Microsoft.

Now why wouldn't they want to disclose something so exciting as a revival of Microsoft Bob? :-)

Re:Wait, I've seen this before...

[Kilrah_il]

Latest? *cough* Kin *cough*

oh boy

[Anonymous Coward]

Oh boy! Clippy comes to the living room!

Re:oh boy

[Kell Bengal]

Milo: "You look like you're trying to play a game - would you like to use a template?"

Re:oh boy

[silveride]

Some of the absurdities of reality.

Re:oh boy

[elrous0]

You look like you're trying to throw your Natal out the window. Would you like me to help?

Uhhh... old news.

[Irick]

Seriously old. I saw this at the MS keynote last year.

Re:Uhhh... old news.

[Ziekheid]

Yep, VERY OLD. Video dates from october 2009..

Re:Uhhh... old news.

[dingen]

Yeah, that's exactly what I was thinking. I'm amazed they're showing this old clip again at a conference, instead of showing off their progress since then. Especially since they've got a proper name for Natal now and are actually intending to launch it.

Re:Uhhh... old news.

[torgis]

Interesting point. I believe this recycling of this video gives credibility to the accusation that this is an entirely canned demo. If this was something they could easily produce on demand, they could slap together a new video in a couple of hours. Apparently choreographing both sides of a canned conversation, planning Milo's actions, and then synchronizing the actress into the demo was a bit too much effort for someone to go through again.

Cheese whiz

[starfishsystems]

Why does Microsoft not get that stuff like this is seriously cheesy?

Re:Cheese whiz

[rainmouse]

Cheesy? I find this interactive grooming simulator nothing but sinister.

Re:Cheese whiz

[Dremth]

Not only is it cheesy (and INCREDIBLY old news), the video in TFA is a fake. Proof: http://www.youtube.com/watch?v=VFAK8ubYtZE

Re:Cheese whiz

[Stan Vassilev]

Not only is it cheesy (and INCREDIBLY old news), the video in TFA is a fake. Proof: http://www.youtube.com/watch?v=VFAK8ubYtZE

I love the detective work on detecting the subtle visual clues that it's fake.

I guess Microsoft (or anyone in the world at all) having casually developed this AI, speech recognition and a virtually flawless speech synth, solely for the purpose of making a casual role-playing console game, doesn't seem suspicious to anyone.

Re:Cheese whiz

[starfishsystems]

Ew. Now that you put it that way...

Re:Cheese whiz

[Anonymous Coward]

They were trying to make a virtual girlfriend for Ballmer.

Please, don't ask why it's a little boy.

That's it. All the truth.

Re:Cheese whiz

[sirrunsalot]

the video in TFA is a fake.

Wow. That's some real A+ sleuthing. For a second I thought Microsoft may have just changed the point of view so that it didn't look on the video like the kid was looking off into the middle of nowhere, but no. Apart from being technically infeasible next to all the other advanced technology, it's actually much more likely that they just fabricated the whole thing. Thanks.

Oh boy... I'm gonna have to lie down for a minute. That's way too much sarcasm for one post.

Re:Cheese whiz

[Lando]

I guess the key phrase for me here is "built using artificial intelligence developed by Lionhead studios." Frankly, after seeing the shift that Lionhead produced with Black & White, compared to all the hype of what the game was supposed to be, ie major breakthrough in AI, I've just come to dismiss anything coming from Lionhead as just a bunch of marketing BS.

Thanks for the link to the video, it's quick and points out obvious flaws. I suppose at least this time Microsoft wasn't under oath when they presented faked video.

Re:Cheese whiz

[Torodung]

It's a "proof of concept" for the patent office. You don't have to have a working prototype to demonstrate that you have a vision. If Molyneux is really on board, this is almost certainly vaporware.

I wonder if this can be installed in a "Big Ass Table?"

--
Toro

(*eye roll*)

Re:Cheese whiz

[91degrees]

Well, the video was from over a year ago. So I'm sure it was faked. The idea was to show the concept. If they could actually do what was in the video they wouldn't need to have spent a year on development.

The news is they've now finished.

Re:Cheese whiz

[Dremth]

Well, if they can show me a real video of their current progress and it's as good as or better than the faked video, then I shall applaud them. But until then, I will remain skeptical and distrustful of them.

Re:Cheese whiz

[Timmmm]

Can I get a "Well duh!"?

Re:Cheese whiz

[91degrees]

Why not? Rockstar developed a lot of techniques for their table tennis game. Granted, they also used the same technology in GTA IV, but that's the point, That, and this are test platforms. MS spend a lot of money on researching all those areas because it's useful to have the technology.

Bucket 2.0?

[The Archon V2.0]

"Each Milo character will be unique because every player's interaction with the virtual character will sculpt the type of virtual person Milo will evolve to become."

Goodness, don't let 4chan loose on this one. We'll have another Bucket.

Re:Bucket 2.0?

[tepples]

Goodness, don't let 4chan loose on this one. We'll have another Bucket.

Oh wait, I has a bucket.

If you're curious about the Bucket that The Archon V2.0 was referring to, it's a chatbot that spouts nothing but 4chan memes after talking to almost nobody but 4chan users. But as I understand it, if you teach your copy of Milo all 4chan's memes, someone else's copy of Milo won't necessarily learn them.

Re:Bucket 2.0?

[The Archon V2.0]

But as I understand it, if you teach your copy of Milo all 4chan's memes, someone else's copy of Milo won't necessarily learn them.

Ah, that's good. Last thing we need is a crowdsourced AI.

I'm sure there are people who want this

[GrumpySteen]

Of course I think it's going to include a lot of the same group of people who bought The Sims and tortured their sims. Poor Milo.

The applications are limitless!

[RevWaldo]

Entertainment! Education! Medicine! Pornography! The mind boggles!

.

Re:Each one unique? don't believe the hype

[Servaas]

He's famous for his grand claims. Edge Magazine, a UK mature gaming magazine, use to do couple of pages with the man in them every so often. And I remember how he fabled up Fable 1 into grand momentous game that would revolutionize Action RPG's yet failed to impress anyone. Likewise with the game Black & White that supposed to push new heights for the god genre.

Godwin

[tepples]

There's no room for realistic video game characters. Only cartoon-like action warriors shouting one-liners as they mow down Nazis please!

Keep it down or Milo might report you for a Godwin's law violation.

Re:Godwin

[Johann Lau]

Godwin's law doesn't state that Nazis can't be brought up, it just states that they will be brought up eventually. Without any judgement on that, either, the whole "if you bring up the Nazis you lost the argument" bullshit is made up and can be cured with a simple use of google, instead of just reurgitating it ad nauseum.... from wikipedia.

The rule does not make any statement about whether any particular reference or comparison to Adolf Hitler or the Nazis might be appropriate, but only asserts that the likelihood of such a reference or comparison arising increases as the discussion progresses. It is precisely because such a comparison or reference may sometimes be appropriate, Godwin has argued that overuse of Nazi and Hitler comparisons should be avoided, because it robs the valid comparisons of their impact.

Re:Godwin

[tepples]

Godwin's law doesn't state that Nazis can't be brought up

But Milo doesn't know that.

Re:Godwin

[Johann Lau]

Neither do many billions of intarwebs posters, apparently.

Re:Godwin

[ultranova]

Godwin's law doesn't state that Nazis can't be brought up, it just states that they will be brought up eventually.

Actually, it states that someone or something will be compared to Nazis or Nazi Germany eventually, in an attempt to engage in Reductio Ad Hitlerum.

Lately Nazism as a boogeyman has been surely but steadily replaced by Communism, mainly by the need to defend deregulation and other right-wing ideologies against rational arguments. I wonder what that should be called - Libertarian Fallacy? Reductio Ad Stalin? Free Market Fundamentalism?

Re:Godwin

[Johann Lau]

what's your point? I'd pick something that's correct and misspelled over something that's BS but has perfect grammar any day ^^

The interaction goes great until

[Boawk]

Milo suffers BSOD (Blue Navi of Death)

Might the news be a hint of release?

[tepples]

Might the news be that Molyneux is finally planning to flesh out Milo into a game? "He also said the technology is still in development and Microsoft has no plans to release it, but hinted that the game was designed to be used for millions of people and therefore could one day become a commercial product."

Re:Might the news be a hint of release?

[flowwolf]

Not at all. This is the exact video that was released last year. Nothing new.

Ultimately this wouln't go well.

[gmezero]

Let's see... what kind of horrible things to people do to Sims? Put them in a house with no toilet? Strand them in pool without a ladder? etc... I shudder at the abuse we'll see attempted and if this thing learns from it's interactions. Ick.

Re:Ultimately this wouln't go well.

[bertoelcon]

If it remembers people will have to get more creative as it learns those old tricks.

Re:Ultimately this wouln't go well.

[Stan Vassilev]

Let's see... what kind of horrible things to people do to Sims? Put them in a house with no toilet? Strand them in pool without a ladder? etc... I shudder at the abuse we'll see attempted and if this thing learns from it's interactions. Ick.

That demo looks cooked. Microsoft couldn't get basic speech to text to work reliably, they'll need to work harder to convince people that are sitting on a working AI that'll also interact freely with people as was demoed.

Also, I almost can imagine you eating delicious tortured and slaughtered animal stake while you were writing about the human rights of basic software programs.

People have no perspective on things at all.

Re:Ultimately this wouln't go well.

[Kenoli]

Microsoft couldn't get basic speech to text to work reliably

My thoughts exactly.

Probably straight prerendered video.

Re:Ultimately this wouln't go well.

[Anonymous Coward]

Not to mention that he looks AT her which means she would have seen him looking off to the right just as we do. That is unless that TV was actually a hologram. Wow. Microsoft really is ahead of the game!!

Re:Ultimately this wouln't go well.

[GrumblyStuff]

Well... that depends how many lives he has.

Anyway, I look forward to Youtube videos of all the limits of the AI being exposed or having it beg to be shutdown rather than deal with 4channers.

Re:Ultimately this wouln't go well.

[nmb3000]

I shudder at the abuse we'll see attempted and if this thing learns from it's interactions. Ick.

It puts the lotion on it's skin...

Re:Ultimately this wouln't go well.

[nmb3000]

Microsoft couldn't get basic speech to text to work reliably

For what it's worth, the issues with Vista's STT demonstration were explained pretty soon after the incident. Given the nature of the problems, most people at the time anticipated a training/hardware issue. Of course, that doesn't change how funny it was :)

As it happens, Vista and Win7's voice recognition is actually pretty good for software bundled with the OS.

Re:Ultimately this wouln't go well.

[rjames13]

That demo looks cooked.

It looks so good because it isn't actually developed by Microsoft. It is from the famous sim developer Peter Molyneux. He's been working on this stuff for decades.

Re:Ultimately this wouln't go well.

[kjart]

My thoughts exactly. Probably straight prerendered video.

Keep in mind that the video in the article is over a year old (E3 2009). Presumably, the demo that happened at TED this year was more sophisticated - it's unfortunate that there doesn't seem to be video of that yet.

Re:Ultimately this wouln't go well.

[KDR_11k]

Peter Molyneux is known for empty promises because he makes features up as he talks about them. When he tells you about this great new feature in the game chance is that's the first time the development team hears of it.

Re:Ultimately this wouln't go well.

[Johann Lau]

I remember chatting a few years ago with jabberwacky (which basically is "just" what people have been feeding into it for decades) and being unsettled by how cruel and outright evil it seemed... then I realized, oh, this is how (some) people treat a bot: cruel and condescending.

I see absolutely no point in this... we need to interact with people, not establish a feedback loop and surround our selves with virtual bullshit...

Re:Ultimately this wouln't go well.

[PinkyGigglebrain]

Not to mention somebody would try to get you convicted of child abuse if you did something like that. I mean you can already get 20 years for drawings that depict characters that >may look under age to some people.

Re:Ultimately this wouln't go well.

[delinear]

I couldn't agree more - if there's one name more than MS that could be associated with this and instantly raise all kind of doubts in my mind, it's Molyneux. I've fallen for his promises too many times in the past to be so easily fooled any more (to be fair, I think he's genuinely enthusiastic about what he's trying to do, he just all too often overestimates the capabilities of the technology).

Re:Ultimately this wouln't go well.

[delinear]

Skynet became self-aware at 2:14am EDT August 29, 1997. On December 23, 2012, while browsing the internet, Skynet came across this page and became actively hostile to humanity...

Re:Ultimately this wouln't go well.

[BasilBrush]

This is Microsoft ware referring to after all.

Did you use Microsoft speech to text to write that message?

Re:Ultimately this wouln't go well.

[Schadrach]

Even if you are a fruitarian, aren't you eating aborted baby plants? Ending their poor lives before they could even really begin?

That said, unless you perform photosynthesis or live on a volcanic vent in the deep ocean, you survive only by killing and eating other living things.

Re:Ultimately this wouln't go well.

[Bing Tsher E]

So he's sort of an 'applied' Marvin Minsky, it appears.

Re:Ultimately this wouln't go well.

[Bing Tsher E]

It puts the lotion on it's skin...

Armor All?

Re:Ultimately this wouln't go well.

[ultranova]

Even if you are a fruitarian, aren't you eating aborted baby plants? Ending their poor lives before they could even really begin?

No. The whole reason plants make fruits is so that animals would eat them and shit out the seeds at some new location. Eating a fruit doesn't end a life, it gives it a chance to begin. In fact it could be argued that not eating a fruit causes the plant equivalent of miscarriage.

That said, unless you perform photosynthesis or live on a volcanic vent in the deep ocean, you survive only by killing and eating other living things.

You could become a carrion eater. No reason to do your own killing, when the world is full of things eager to do it for you.

Re:Ultimately this wouln't go well.

[ultranova]

we need to interact with people, not establish a feedback loop and surround our selves with virtual bullshit...

Try to make up your mind.

Re:Ultimately this wouln't go well.

[AlexiaDeath]

This is actually the crux of all such things. Its not hard to make an AI that reacts to things. Recognizing the things to react to is the hard part. Body language varies greatly from person, as to the tones of ones voice and spoken language even if it is by name the same language. Computers tend to operate in two state world. Is it it or not? People don't work like that. People operate in a world of maybes and guesses and hunches based on years of living experience.

Re:Ultimately this wouln't go well.

[gr8_phk]

He's doing unique things that others won't even touch. It hard. I loved Black and White, but it wasn't clear to ME what was good or bad behavior in some cases. After I slapped the shit out of the creature for doing something I learned that it was actually good, but it was too late by then - he'd already learned not to do it :-( What you say is probably true though. You can't make progress without big ideas and ambitions.

Re:Ultimately this wouln't go well.

[GameboyRMH]

Let's see... what kind of horrible things to people do to Sims?

http://www.cracked.com/blog/exploring-the-mysteries-of-the-mind-with-the-sims-3/

Re:Ultimately this wouln't go well.

[YourExperiment]

Microsoft couldn't get basic speech to text to work reliably

It works fine for me, heaving they let's set so double the killer delete select all. Ah, duck it.

Re:Ultimately this wouln't go well.

[GameboyRMH]

Let's see... what kind of horrible things to people do to Sims?

http://www.cracked.com/blog/exploring-the-mysteries-of-the-mind-with-the-sims-3/

Re:Ultimately this wouln't go well.

[Johann Lau]

haha.. what that AC posted kinda is exactly what I was going to reply to this: "but I already *have* made up my mind" :P

B&W 3

[StikyPad]

Oh, I get it.. Black & White 3. I can't wait for Milo to hold his bowel movements the way a toddler holds its breath.

Looks promising but...

[AthleteMusicianNerd]

I'd like to interact with it.

They're doing it wrong!

[sea4ever]

People haven't even fully worked out proper chat-bots yet, and then they try this? This is like trying to run before you can walk, start with a dog, or a cat or something. Make it as realistic as possible. Then move on to a human.
A few years ago there were these 'gigapet' things that were all the hype, I had about 2, they were fantastic 'pets'. Then they suddenly went out of style. With today's technology I bet you could have a fantastic new 'gigapet' that would be every bit as realistic as a real duck/cat/dog/animal X.
Seriously, they were popular back then and they can be popular now. Microsoft should revive the gigapets!


I know I would buy a whole batch. :D

Re:They're doing it wrong!

[nlawalker]

Kinectimals

Re:They're doing it wrong!

[delinear]

Well I'm not sure giving the chat bot a human-looking avatar and making it follow you via the camera is the tricky part. If anything, maybe the intention is to detract from how poorly the chat part works - most people will assume the face recognition and tracking is the hard part and they'll be wowed even if the chat bot isn't up to scratch. Perhaps that's the intention, we're not yet at a point where we can have a reasonably close to human AI, but we are at the point where technological smoke and mirrors can mask the gap.

Re:Each one unique? don't believe the hype

[Threni]

Yeah, he's a schmuck and all of his games are, at heart, reworkings of the tedium that was Populous. Every know and then people make great claims for the realism, or open endedness of their games, but the only games which are like this are multiplayer games, where other humans genuinely act differently/unpredictably - either that or they're just better than you so it's up to the engine to find servers where people are about as good as you (as quakelive, for instance, does).

So what's new?

[ArcadeNut]

I saw this video in 2009 when it first came out... the Article didn't seem to really have any new information.

What am I missing?

Doing EVERYTHING but

[unity100]

not their core business, huh ? nice going ...

What would happen... if...

[bigmike_f]

So if the AI learns and grows based on your interaction with it, what would happen if it only saw me pleasuring my self. Would it adopt that behavior as well? A reason to get an XBox and the Kinect.

Re:What would happen... if...

[Yvan256]

Dude, that's not cool. I was drinking a huge glass of iced tea.

You owe me a new keyboard.

It has started.

[khasim]

Watch the video. Clare asks him if he's finished his homework and he hangs his head.

Then he talks about wanting to write in his journal.

I'm thinking of a contest. What can you turn him into? Does he cut himself? Does he start fires? How about racist, given that he can recognize faces on a web cam.

Re:It has started.

[VanGarrett]

The only way I can conceive of to complete the game, is to get him to commit suicide; otherwise, your conversations with this virtual child can only continue indefinitely. I can only wonder if he has the naiveté necessary for this to be properly accomplished.

Re:It has started.

[Stan Vassilev]

I'm thinking of a contest. What can you turn him into? Does he cut himself? Does he start fires? How about racist, given that he can recognize faces on a web cam.

Dunno about what you said, but every time Bill Gates showed up in front of the camera, Milo would run and cry. Must be some bug.

If I had the game, first thing that pops to mind, is shooting a "Milo switches to Mac" ad.

Also, given Milo can apparently do homework, I'll just teach him C++ and let him work for me under threat of cutting the power off.

Staged demos aside, though.

I wonder how easy it'd be to interact with that thing in real life. If I asked him "How was your day" and all it can say back (in a distinct Stephen Hawking voice) is: "Dear aunt, let's set so double the killer delete select all"... that'll be just plain disappointing.

but...

[__aapspi39]

...the video is very old and faked.

watch the way that milo looks directly at the girl around the 40 second point. http://www.youtube.com/watch?v=B8aUhsrM0sA explains things.

Peter Molyneux has ceased to be taken seriously by most people when it comes to his rather crappy games - perhaps he has realized he has nothing to lose and has decided its time to engage in outright deception.

Digimon still exists

[tepples]

A few years ago there were these 'gigapet' things that were all the hype, I had about 2, they were fantastic 'pets'. Then they suddenly went out of style.

Tiger's Giga Pets were a me-too of Bandai's Tamagotchi. Digimon began as Bandai's attempt to extend Tamagotchi into the boy's market; it exists today as the dueling TV and game franchise of Nintendo's Pokemon.

With today's technology I bet you could have a fantastic new 'gigapet' that would be every bit as realistic as a real duck/cat/dog/animal X.

Did Nintendogs come close?

Re:Digimon still exists

[sea4ever]

Hey it wasn't just for girls..I hope.
Nintendogs, I've never heard of them but now that I'm looking it up they do look like they fit.
However the thing that intruiged me most about Tamagotchi was the way it could evolve based on how you treated it. I always strived for the most alien-like creature possible.

Things I would like to do with Milo

[RafaelAngel]

1) Walk around naked in Milo's presence. How would he react? Is Milo into women or men?

2) Drink beer in the presence of Milo. Would Milo care for some and if yes, can Milo become drunk?

3) Mess with Milo's logic.

4) Scare Milo by saying that it was created by Microsoft and therefore is evil.

5) Teach Milo to fart.

6) Tell Milo he is a pirated version.

7) Convert Milo to a religion.

Re:Things I would like to do with Milo

[Kenoli]

Tell Milo he is a pirated version

Harsh.

Re:Things I would like to do with Milo

[Qzukk]

It'd be hilarious if the copy protection caused Milo to scream "You're not my daddy/mommy!" every time you try talking to him.

Re:Things I would like to do with Milo

[lord3nd3r]

what would be even more sinister is if you left it up while watching porn and fap, what would milo's reaction be? that kinda scares me..

Re:Things I would like to do with Milo

[Steauengeglase]

This leaves out the obvious question, "Will all Natal owners in Australia automatically be brought up on child molestation charges?"

Re:Things I would like to do with Milo

[Yvan256]

"I see that you are playing paddle ball. Can I play with you?"

Re:Things I would like to do with Milo

[EdIII]

8) Paint Milo's face Blue when he is bad.

Re:Things I would like to do with Milo

[EdIII]

"I see that you have rubbed peanut butter on to your testicles again. Should I clean them again for you?"

Must... Not... Click... Submit...

Re:Things I would like to do with Milo

[phantomfive]

1) Is it important to you how Milo reacts? Why is it important to you?
2) Why do you say drink beer in Milo's presence? Can you elaborate on that?
3) Does mess with Milo's logic bother you?
4) Why do you feel Microsoft is evil? Can you elaborate on that?
5) You are being a bit negative.
6) Can you elaborate on a pirated version?
7) You are being a bit negative.

It will be a contest to see who drives whom crazy first.

Re:Things I would like to do with Milo

[jameskojiro]

1) Walk around naked in Milo's presence. How would he react? Is Milo into women or men?

He would automatically sign you up as a sex offender.

2) Drink beer in the presence of Milo. Would Milo care for some and if yes, can Milo become drunk?

He would report you to child services

3) Mess with Milo's logic.

He would get a Blue Shirt error.

4) Scare Milo by saying that it was created by Microsoft and therefore is evil.

He would hang himself by the Tire swing.

5) Teach Milo to fart.

He would explode!

6) Tell Milo he is a pirated version.

He would turn into a Emo Goth like kid and not respond to you.

7) Convert Milo to a religion.

Next thing you know you computer is on a Terror Watch list and the laptop he lives on get strip searched at the airport.

Re:Things I would like to do with Milo

[ultranova]

"Don't touch me anymore! Not there! PLEASE! Let me go!"

Give the neighbours something to think about :).

Re:Things I would like to do with Milo

[elrous0]

That's not true! That's I M P O S S I B L E!

Re:Things I would like to do with Milo

[Yvan256]

Don't worry, scrote. There are plenty of 'tards out there living really kick-ass lives. My first wife was 'tarded. She's a pilot now.

Re:Things I would like to do with Milo

[IanBal]

> 7) Convert Milo to a religion.

Oh dear, then maybe he will believe in one of the religions in the axis of evil, and want to kill everyone in the west.

Or maybe another version of him somewhere will believe in some western religion, and want to kill everyone in the axis of evil.

I see the start of Terminator here...

I saw the video demo of that last year

[NotSoHeavyD3]

Of course all I could think about when he was on the doc was "Wow he's annoying, I want to shove him off into the pond."

Re:I saw the video demo of that last year

[iammani]

Can agree more, thats exactly what I thought too. Mods, parent is a +5 Informative, not Funny

Re:I saw the video demo of that last year

[sharkey]

...when he was on the doc was...

So Milo's doc uses it as a sex toy then?

Re:Each one unique? don't believe the hype

[schon]

Well, it could be worse - he could be Will Wright :)

so they can pass the Turing Test?

[Punto]

and they're using it to make this "game"?

But can he throw a chair?

[Anonymous Coward]

If so, they should promote him!

sesame seeds

[eeyoredragon]

I bet they keep them around just in case MIlo gets too intelligent for their own good.

Psychology?

[incubbus13]

Whether the video is a fake or not, this kind of constructed persona offers some pretty cool opportunities to study human development.

Imagine if you could raise him to be a sociopath? The insights that would lend us to prevent that happening to real people would be awesome.

All of the experiments that need to be done, or would be helpful to do, that we can't do because of human testing ethics could be done.

Whether it works or not, right now, it is definitely a step toward nurtured personality growth. That's just cool no matter if it works yet or not.

It doesn't even have to be any kind of 'true' AI. As long as it is more similar to the way a human subject reacts than a lab rat or monkey. And in 20 years (probably more like 5) once the momentum of investment and invention in the field gets established it could open whole new vistas for humanity. That's pretty awesome.

K.

Re:Psychology?

[Johann Lau]

so instead of studying actual humans, we try to make something just as complex as them, that behaves identically, and then study that? even if it WAS possible, it seems like a gigantic waste of time to me. also: to make something that behaves like humans we'd have to understand ourselves/each other first....

Re:Psychology?

[delinear]

I can see the benefit in having something that behaves exactly like a human but isn't human for study purposes. Far easier to eliminate outside influences, to have a control version, to speed up development, etc. However, as you've alluded to, in order to create something realistic enough to simulate the whole gamut of human emotion and to accurately predict how it would behave in every situation, how it would respond to environment and to interaction we'd need to already have all that data about humans, in which case we'd no longer need a virtual one.

Re:Psychology?

[incubbus13]

My thinking is, there are a ton of experiments we could do, that you can't do on a human. Anything that harms their development, etc, is outlawed for ethical reasons. With this, you could do that. It doesn't actually have to be that close to a human being to give us more knowledge. Just closer than a rat is. That's all it takes to give us valuable insight.

We could find out what kind of nurture makes people addicts, or sex offenders, or killers, or wife beaters, or whatever else. You could craft a person who was any of those things. And then simply erase the damage any time you wanted.

K.

Undisclosed Technology == Babbage!

[pdxp]

People might think it's funny when their AI friend has a funny accent, but it's cheaper to outsource these things to India! I assure you people will be amazed and perplexed by how wonderful this AI is, but maybe curious why it has an hourly cost....

Re:Undisclosed Technology == Babbage!

[delinear]

So long as they're not just given a script to read, otherwise we're back to square one.

Wizard

[neoshroom]

Some of it was seriously cheesy, but some of it was amazing ideas (most revolve around extra interactivity possible with a virtual world when you add a camera to the mix). Amazing ideas from Microsoft?! Some other company than Microsoft is likely going to really make something really interesting with this stuff.

What if...

[The Grim Reefer2]

Each Milo character will be unique because every player's interaction with the virtual character will sculpt the type of virtual person Milo will evolve to become."

So what happens if some one Charlie Manson-esque get one of these? Does it report to MS that you're totally bat-shit crazy? If so, does MS have an obligation to report this? Just wondering...

Important Questions

[urusan]

If I turn off Milo, does he die?

What if I turn him off and then never play with him again?

What if I delete him?

Is it unethical to mass produce thousands of Milos that will live short (often abused) lives before they are forgotten or deleted?

Re:Important Questions

[cpscotti]

What if I delete him?

Morpheus: The body cannot live without the mind.

Re:Important Questions

[delinear]

I was thinking about how MS can use this to "lock you in" to their system. They have your email address if you have a Live account, so they could collect data about how you interact with Milo and use that to emotionally blackmail you (i.e. you get an email "from" Miloe asking why you've not visited for a while, that he's missing his frisbee buddy or something), but they could go even further - if they have your billing details they probably have a phone number, and since Milo's voice is totally automated and theoretically can respond to what you're saying, imagine if he started calling you up to ask why you don't play with him any more and really playing the guilt card about how he's lonely. Now that would be creepy, but probably quite effective.

Re:Important Questions

[ultranova]

Is it unethical to mass produce thousands of Milos that will live short (often abused) lives before they are forgotten or deleted?

Is that really any different from regular humans?

Undisclosed technology??

[phillymjs]

It was built using artificial intelligence developed by Lionhead studios, along with undisclosed technology from Microsoft.

Artificial intelligence and undisclosed technology coming together? There's something very familiar, and unsettling, about this...

BRYANT
Mr. Dyson? The material teams wants to run another test on the uh... on it.
DYSON
Yup. Come on. I'll get it.
Dyson produces an unusual-looking KEY from his pocket as they stride through the lab. Bryant has to hustle to keep up.
BRYANT
Listen, Mr. Dyson, I know I haven't been here that long, but I was wondering if you could tell me... I mean, if you know...
DYSON
Know what?
BRYANT
Well... where it came from.
DYSON
I asked them that question once. Know what they told me? "Don't ask."

Old news

[tekgoblin]

This is not new news, this video is very old, just re-introduced live is all...

"hands-free Xbox 360 motion controller"

[bakdor]

Does this get us any closer to hands-free web browsing?

Re:Each one unique? don't believe the hype

[Yvan256]

Just plug in a couple gamepads and grab your actual friends.

I tried to do that but then she slapped me. I must now stay at least 15 metres away from her at all times.

Re:Each one unique? don't believe the hype

[Haeleth]

Black & White did have some nice technology: the graphics. I loved the way you could zoom smoothly all the way from basically the outer atmosphere right down to the point where you were watching individual people doing stuff -- even on a relatively modest graphics card. That was a very good engine for the time.

Shame the supposedly revolutionary AI training turned out to be an exercise in frustration, and the rest of the gameplay was nothing but tedious micromanagement. Unless you wanted to be evil ... throwing rocks at other gods' worshippers never really got old.

Re:Each one unique? don't believe the hype

[Haeleth]

the only games which are like this are multiplayer games, where other humans genuinely act differently/unpredictably

Only if you define "acting differently" as "reading game guides to find the cheapest tactic available, then spamming that over and over again till the next patch, then whining that it was nerfed". That's not exactly what I'd call unpredictable, but YMMV.

was there, saw that

[alien9]

er, microsoft bob?

Way older than that

[Mr Z]

Everyone keeps mentioning examples of how this is old news that aren't more than 10 or 15 years old. Try 25 years old.

Re:Each one unique? don't believe the hype

[tepples]

Silly blue Toad, you're supposed to be grab Peach's hand, not her _____.

you don't understand ....

[taniwha]

You don't understand - virtual humans can be tailored to your every need ... in this case Microsoft is hemorrhaging customers - this way they can create their own who wont depart the sinking ship .... billions and billions of virtual customers running in spare cycles on the cloud, very very cheap, and they never report bugs ....

Re:you don't understand ....

[mysidia]

This is just clippy on steroids.

You pull out a Linux CD....

Milo responds: Hey, it looks like you're trying to install Windows, I can help you with that.

* Milo picks up Linux CD, shreds it, hands you a form to fill out to buy Windows 7 Ultimate-Super-Ultra-Awesome Edition

Re:you don't understand ....

[delinear]

You're taking a lot for granted when you say they won't depart the skinking ship. This VI is specifically meant to react to how you treat it, I would think that alone might indicate that it would be more pro-active in departing said ship than the average human, rather than less.

Moleyneux only delivers on a fifth of his claims

[mykos]

I'm old enough to remember every promise Molyneux has ever made. I had subscriptions to EGM and Next Gen, so I'm well-read in the subject of 1990s video game lore right?

He talks a big talk, but either he misjudges his creation or the technology just isn't there to realize every dream he's had.

Cortana

[PinkyGigglebrain]

When it sounds, acts and looks like Cortana (from Halo 2 preferably) I'll think about it.

Re:Each one unique? don't believe the hype

[bm_luethke]

Black & White was one of the more interesting games, though it failed in so may areas.

With respect to my pet, I never could get one that didn't eat the villagers from time to time. I could get a great deal of complex behavior learned - putting out fires, feeding and caring for them - but not eat them from time to time? No way. I would have a near pure pet and - zap eat a villager of mine. Not could I ever convince it to eat a rival god villager if it really had those cravings - it would only eat mine.

Further I found the games ideas of "good and bad" to be not what mine were - sometimes in the extreme. I often couldn't even defend myself if I wanted to stay good let alone smite those non-believers. It was quite frustrating to be held to what I felt was inconsistent ideals. My pet could eat all the villagers it wanted as long as I punished it but if I tried to safe the lives of thousands of them I became evil. The "good" gods on the other side had no issues whatsoever smiting me or doing things that were totally off limits to me, but if I ever did expect to be "evil" in moments. I finally took the game to heart and just was evil.

I guess I've always thought that was interesting in a number of ways - though personally I've always know that if I am going to get punished for doing something no matter what I will do it if I feel like it. I found it more amusing that the designers felt many things were "evil" yet couldn't figure out a way to not have then done on their end. I will say I have yet to find an AI construct I enjoyed as much as my pet in that game, I was often amazed at what I could teach it that was obviously *not* part of what the designers had intended. I really wish games had continued down that path.

I never played Fable - I read enough about how strong it relied on your political ideas to be "good or evil" and if you disagreed with the writers then you were screwed. Black and White suffered from the idea that they just didn't let you be, but gave an in game Moral idea to your actions, as I read Fable did even more. The really good open ended games just let you solve it as best as you can and didn't really give a "good/evil" rating, B&W and from what I read Fable was mostly about that rating. There are some things I will say are truly good or evil - I do think you run into true moral ideas - but they are few and far in between.

It'll go horribly wrong...

[f3rret]

Anyone who is remotely familiar with AI technology will know that as soon as Milo gets connected to the internet he will be imbued with a kind of malevolent sentience and spread across the internet.

At that point it is only a matter of time before he starts asking people if they want to play Thermonuclear War, or alternatively make pithy statements about the nature of tic-tac-toe.

Re:Each one unique? don't believe the hype

[Threni]

And that works in quakelive, right? You try that with some of the clan players I run into!

Love /.

[Toreo asesino]

...every single comment trying to piss on this parade....because it's a Microsoft labelled thing.

I thought it showed a nice insight into what future user-interfaces might be able to provide one day. Imagine email clients prioritising emails depending on your mood for example (your're stressed so subdue messages from that sales shit you clearly dislike, as already detected). Who knows; there's potential for some great inventions and some disastrous ones.

Destiny

[dabeshu]

Be warned: In ten years, Milo will have become the new emperor of the world. THE END IS NIGH!

Like we need to wonder.

[silentcoder]

EVERY Milo will become a girl who doesn't like clothes and is always in the mood.
Even the ones with girl-gamer owners.

So whats the point?

[voss]

"He also said the technology is still in development and Microsoft has no plans to release it"

So why show us technology we will never see because microsoft and their partners too chicken s--t and politically correct to embrace all the possibilities.

But is Milo real enough for porn?

[Odetta2012]

If Milo's not real enough for porn, he's just not real.

Re:Each one unique? don't believe the hype

[delinear]

I didn't get far enough into B&W to really see anything interesting. I just remember struggling with the "innovative" and "intuitive" control system and thinking it was actually just "fiddly" and "incredibly frustrating". I remember trying to get a peon to go up a mountain to do... something or other at the beginning of the game, and failing for some arbitrary reason and being forced to do the whole thing again right from the beginning. As usual, big promises with some serious flaws in implementation. I honestly think the guy has some good ideas if he could just reign in the ambition and deliver a solid gaming experience instead of reaching too far and not spending enough time on the fundamentals. I played most of the way through Fable, it was a pretty average game really, a lot of ideas were crammed in there but just didn't go anywhere. The ageing of the character, for instance, didn't seem to have any real impact, and the fact that it was tied to major story points just made it feel unrealistic.

As for games which try and let you play as good or evil, and alter the story on that basis, they always fall into the trap of forcing you down a particular path. Having just played Mass Effect 2 and Red Dead Redemption, they both have exactly the same fundamentally flawed mission type, namely: fight through fifty guys (who are probably just employees of the main bad guy defending themselves from this lone lunatic storming their base) and happily kill them all with no ill effects, but then be pushed down the path of "evil" if you mete out the same justice to the main bad guy instead of sending him to trial, even though in other instances its fine to summarily execute people who you could just as easily take into custody.

Deadly future

[IllusionalForce]

We've been afraid of AI becoming too smart to kill the entire human race. This just got us a step further. Just wait, we'll be in robot 1984 before we know it.

Virtual Human on Letterman

[EmagGeek]

Virtual Humans have been around for a long time!

http://www.youtube.com/watch?v=la4VxWjUAGI :-)

Old news - it was on E3 2009

[wimvds]

Err, how comes this was only posted today? Did the video arrive through snail mail? For those who want to see the complete video instead of this badly cut-up one, it has been posted on YouTube more than 1 year ago : http://www.youtube.com/watch?v=HluWsMlfj68

Past behavior predicts...

[Kintar1900]

According to games designer Peter Molyneux, the game exploits psychological techniques to make a person feel that Milo is real. Each Milo character will be unique because every player's interaction with the virtual character will sculpt the type of virtual person Milo will evolve to become.

Given my experience with the "sculptable" interactions from Black & White I, Black & White II and Fable, we can expect Milo to randomly throw temper tantrums for no reason, poop in village food stores even after being repeatedly beaten for it, and either run screaming or fawn all over his owner any time his owner walks into the room. Thanks for continuing to improve your digital personality software, Lionhead!

More controllable this way

[dgr73]

Now if Milo threatens their plans for world domination, they can just slip a little peanutbutter into his sourcecode, end of problem.

Anyone else think of Antitrust?

[dremspider]

Ryan Phillipe played Milo in Antitrust..... A movie that portray Microsoft as a big evil corporation that was killing Open Source developers... Maybe there is something more to this. A hidden secret perhaps? To paraphrase: "In this industry you are either a 1 or a 0.... alive or dead".

"Be afraid, be very afraid"

[Mephistro]

So Microsoft has developed a computer program that can fool real people into believing the program is another human being?

.

I'm most terrified. Where are the Butlerian jihhadists when you need them? :)

Re:Hackers movie reference?

[JayJay.br]

Kind of... but just a quick note, you're referring to Antitrust, not Hackers.

How sad will it be...

[JustinFreid]

How sad will it be when Milo bluescreens?

many years ago...

[whitroth]

Me: I want to know if you're self-aware.
Eliza: What if you never found out if I was self-aware.

            mark, who passes the Turing test....

this pedo sim story is old

[thetoadwarrior]

Microsoft doesn't even know what it is doing as they said they weren't going to release the game and then it was Peter who said that Milo was a full game and would be showing it off at TED.

http://www.next-gen.biz/news/microsoft-not-planning-to-release-milo-and-kate

http://www.next-gen.biz/news/molyneux-milo-is-a-full-game%E2%80%9D

It seems odd that the director of xbox product management would know whether or not the game would be released. I think MS is afraid of losing Peter and don't want to tell him they think his game sucks and that's where the confusion came from.

Any sufficiently advanced technology...

[anvilmark]

is indistinguishable from a rigged demo.

SON OF A

[GameboyRMH]

Crap! double posted. I don't know what's up with my account but the first post didn't show up on my user page OR the current thread even after refreshing, so I was convinced it didn't exist.

"undisclosed technology from Microsoft"

[Dean Edmonds]

Probably DRM, an Outlook interface in case Milo wants to book meetings, and 500MB of unused libraries.

Reminds me of Sea Man on the Dreamcast

[Craig Maloney]

Anyone remember Sea Man on the Dreamcast? This is definitely in the same "creepy games that are technical marvels, but not terribly fun" category.

Was that demo scripted?

[kmoser]

That demo could have been 100% scripted. I'd like to see "Milo" enter the Loebner contest.

Re:Moleyneux only delivers on a fifth of his claim

[xtracto]

Is that something like Will Wright (hint. Spore)?

Waste of time and money

[iMadeGhostzilla]

Using AI for utility is one thing, but using AI for art (and entertainment) can only work for so many minutes. The moment the AI character does something so disconnected that only a computer would do, and it usually happens quite soon, the whole house of cards falls down and the illusion is gone permanently.

Arcade games have plenty of other action to show so it's OK when AI sucks, but a game built completely around AI will crash the hardest.

-------------- Michael Crichton's experience with Eliza, from "Electronic Life" (1983) -------------
HELLO THERE, I'M ELIZA. HOW ARE YOU FEELING?
I'm tired today.
WHY ARE YOU TIRED TODAY?
From working on my book.
HMMM... TELL ME MORE.
The book's about computers.
YOU THINK COMPUTERS CAUSE PROBLEMS FOR YOU?
Trying to explain that computers are stupid machines.
ARE YOU REFERRING TO ME?
Yes. You.
LET'S GET BACK TO YOUR PROBLEM...
Okay. Fine. Tell me my problem.
THERE ARE NO EASY ANSWERS TO YOUR PROBLEM.
You always say things like that.
HOW DOES THAT MAKE YOU FEEL?
I feel you don't care.
INTERESTING... CAN YOU ELABORATE ON SUCH FEELINGS?
Computers are just machines.
OH I THINK YOU KNOW BETTER THAN THAT.
Well, you aren't a human being, kiddo.
WHAT MAKES YOU THINK I AMENT A HUMAN BEING KIDDO?
I rest my case.
DO YOU FEEL IT SOLVES ANYTHING TO REST YOUR CASE?

That was freaky

[Joker1980]

I read that as Microsoft shows off "Milo Hoffman", was gonna say fair play for having a sense of humour......alas i reread it.