GDC 2008, Day 2

Soren Johnson
I wanted to go to his talk, but I couldn't get up in time because I stayed up too late doing [can't talk about it for 2 years.]

Ray Kurzweil, Futurist
Oh my god. There is no possible way I can explain to you what he explained to us. A year or two ago, GDC had a "vision" track with a keynote speech from a scientist that worked on the mock-technology behind the movie Minority Report. Spielberg had him assemble a team of top scientists from different disciplines around the world to create a plan about how future technology like transportation, advertising, computer interfaces, etc, etc might feasibly work in the future. The goal was for all these super smart guys to figure out something reasonable in under a year about all this. At first, I wondered what this had to do with the game industry, and maybe this guy didn't belong here. By the end, he seemed like the smartest guy at the entire conference and *we* were the ones who didn't belong.

Ray Kurwzeil (the Futurist!) gave me that exact same feeling. He's like, on a level above another level that's above everyone else. I was trying to estimate how many times more data he used in his talk than the average, but the average is somewhere around zero so it's hard to compute. I think there were tens of thousands of pieces of data thrown at us, at minimum.

So what does this guy actually *do*? He predicts the future. He's a very old man and he's done this for a very long time and he now has a team of 10 data analysts in various different fields who help him. I think he won us over immediately by saying how unfortunate  it is that the thing we make is called "games" and it's the "game industry." As if none of it matters, it's all just a game. Doesn't count! He does a lot of work in AI and said he's stuck with the same unfortunate naming there: "artificial intelligence." No, it's real intelligence. "Virtual worlds?" No, they are just worlds. In Second Life, $2 million changes hands every single day, but maybe it doesn't count, because it's a "virtual world."

The telephone, he said, is a virtual world that's been around for a while. It was really magic when it fist came out, allowing people to share a virtual auditory space across great distances. With deadpan delivery he joked, "But what about that agreement you made with me yesterday? Oh that wasn't a real agreement, it was on the telephone, so it doesn't count." Ha.

Seriously...*seriously* save your criticisms of any of that. Let it go. I have so far told you about 0.001% of what he said, so any trash talk you have will be ill-informed and picking on things like whether AI is "real intelligence." If I could tell you the rest, you'd see how fucking real he means it.

Kurzweil explained that predicting things like Google's stock price in 2020 is very hard, predicting how a possible merger between Microsoft and Yahoo might go is very hard. The fortunes and misfortunes of individual people. The waging of wars and results of wars (that might happen 100 years from now) are hard to predict. But that is not what he predicts. What he predicts, he can do with amazing accuracy. While he can't answer any of that, he can tell you the price of a 1 billion transistors in the year 2014. The capacity of efficiency of solar cells in the year 2040. The size and power of nanite technology in whatever year you name. And so on and so on.

He showed his track record in pretty great detail. One example was his published predictions of the expansion of ARPANET in like 1970 or something. He predicted a doubling every X time that would continue indefinitely, resulting in certain numbers of users in specific years, reaching a tipping point in 1990s at which point he said the entire world would basically be connected through a "world wide web," so to speak. At the time he said this, they were struggling with signing up 1,000 new scientists, so to many, this idea was ludicrous.

There are many, many more examples, but I'll have to skip them. Kurzweil explained that he is not some magician here, he just uses the same trick every time: it's all about exponential growth. Hopefully you're all aware of the whole phenomena about how humans are horrible at understanding exponential growth (it's not intuitive, we can only handle thinking about linear growth). He has a ridiculous amount of data about everything from punch card computers to cell phones, to the nanites about how every kind of technology follows the same exponential growth curve. (He had at least one zillion examples of this, all with tons of data). He says his predictions are only surprising to people who have not looked at where we are on a certain technology's exponential curve.

For example, solar power cells are expensive, hard to make, not terribly efficient, only contribute a very small amount of all the power we generate today, etc. If you aren't looking closely, you'd think they're going nowhere. But they increase in efficiency exponentially over time and we're currently in the period where they are only doubling small numbers to begin with. You might think that even if we gathered all the energy that the sun sends to Earth, that it's still not enough to power the Earth. Well, the actual research shows that we need only 1/10,000th of the energy the sun sends to Earth and solar power planners will be able to capture that amount by the year 2029 at which point fossil fuel will be irrelevant.

The human genome was only very recently mapped, but we can already reprogram the software life runs on, he says. We know the exact gene that causes your cells to store energy in the form of fat. This was useful to hunter-gatherers (hard work to find food, need to store it if you get it), but today it leads to heart disease, diabetes, etc. We've already tried turning off this gene in rats and they live longer, have all the health benefits of being slim, yet eat whatever they want. This is one of like a dozen health technologies he covered, trying to demonstrate that human life span is also going to change on a radical scale. Every year, the human life expectancy increases by some amount. In 20 years (or 30 at the worst), the increase per year will be *greater* than one year! Think about that for a minute!

By using his graphs for exponential shrinking of computer technology, you can see how easy it was to predict that by the year 2006, we'd be able to insert a pea-sized computer into someone's head that has Parkinson's disease to do whatever it is you need done if you have that disease. The computer interacts with your neurons just fine and performs the missing brain function for you. Now in 2008, we can even upload new software to these computers inside people's heads.

It's totally foreseeable when computers will be small enough that instead of a pea-size, they are the size of a human blood cell. I got a little confused at this point about whether this next part is really being tested now or it's a prediction of the near future or what, but he was talking about blood-cell-sized nanites that perform the function of your red blood cells at radically better efficiency. Replacing 20% of your blood cells with these nanites would, for example, allow you to run a marathon without being winded or sit at the bottom of your pool for four hours without needing more air.

He also talked about the implications of having nanites like this inside your body, and how they could be used. Things we call HUDs would be sent directly to your visual cortex (and he had a ridiculous amount of information about how we have now deconstructed the human visual cortex and are implementing  in robots exactly how it works in humans). Anyway, you could have these computers take over your sensory system and make you feel as if you're really in a virtual world instead, perhaps with some kind of "picture in picture" thing of the real world so you are still alert to dangers or whatever.

I have now captured maybe 5% of his talk, so you'll have to live with that. Imagining games in the future he describes is completely mind-blowing.

Clint Hocking
Clint thinks really hard about what he says and that puts him far ahead of a lot of speakers at GDC, ha. His lectures from last year and the year before were incredibly, though this year I wasn't quite as deeply affected. His topic was immersion and he explored the difference between sensory immersion and immersion in thinking deeply about something. For example, you can really be taken over for a moment by the tast of great chocolate, the sight of a beautiful painting, the music of Miles Davis, etc. Movie theaters themselves are a way to try to immerse our senses in a movie (filling up your field of vision, darkening the rest of the room, everyone is supposed to be quiet).

Clint says movies are really pro at this type of immersion, and while games need to care about this, they probably shouldn't try to attack Hollywood on their home turf by completely relying on this type. Games create this feedback loop between actions players make and changes in the game-world, so this leads to the possibility of a different type of immersion. Playing an intense game of chess or even moving Mario around his world is are interactive loops, and create their own kind of immersion.

He also talked a great deal about Guitar Hero and how it offers a pretty deep type of immersion to hardcore players who really try to improve their skills, but it provides a different (shallower, but still enjoyable) type to Grandma, who just wants to fee like she's playing a guitar and rocking out a little, without learning the intricacies of a timing system with hard dexterity challenges.

Experimental Gameplay Workshop
Always great stuff here. Several games involving playing with time that are dangerously close to something that, if only I could tell you.

A few games demonstrated the idea of "obfuscation." One was a simple platform game where everything is made of static (like on a tv screen with no signal). Any screenshot of the game looks like pure static, but in motion you can see patterns in the static that show you where the ground is. Another obfuscation game lets you play an invisible monster. No predator-like shimmering though, you are really completely invisible. It has platforming challenges, dodging bullets, and you even fight a boss that's another invisible monster. The remarkable things about this game are that you can actually play it at all, and that while you are very intensely concentrating on something like jumping around and avoiding stuff, spectators see absolutely nothing happening! ha!

One game, I don't know the name, was probably the weirdest game I've ever seen. The entire point of it is that the rules themselves are obfuscated. Even moving your guy around has mysterious consequences. There's some strange low-rez alien that screams at you for a long time, but he somehow disappears if you walk into him for long enough, and then a heart falls from the sky, and you can enter it and ride it to the next area. Then you have to walk to the right for a while, which scrolls some star background behind you but nothing else. Ok, so it's like this gibberish game of what-the-fuck-is-happening. But it really does raise interesting questions. When I see one single screen of Super Mario Brothers, I know like 100 things. I know about jumping, about landing on the enemy's head (not the side which can hurt me). I know about going into pipes, hitting question mark blocks that might give me a powerup or a coin, about jumping on turtles that make turtle shells that I can kick around, etc. But *this* crazy game? I have zero, zero clue what's going on, and you basically never have that feeling in a game. It was actually very interesting to have the feeling that I knew nothing at all about what even the basic rules are.

Space Giraffe (XBLA) was another demonstration of obfuscation. It's a Tempest-style game, sort of, but the entire point is that it goes to more and more and more lengths to completely cover up everything with ridiculous special effects. It reaches the point where a spectator cannot even tell anything about what's going on. It looks like some kind of malfunctioning psychedelic explosion that can't possibly have any meaningful information in it, yet you as a player can actually play it. As you play each level, some mysterious part of your brain really can sort out the gameplay part from the crazy light show part, and that's kind of interesting.

There were more interesting experimental games, but I kind of forget the rest right now.

Positive Impact in Games Panel
This session had a simple message, and it was powerfully delivered. It was very similar to Jonathon Blow's message from the day before, but somehow delivered in a "good guy" way rather than Blow's "bad guy way." (No knock on you Mr Blow, I love your stuff!).

Anyway, the point of Molyneux, Chris Taylor, Louis Castle, and the rest was that our medium gives us incredible power over people's lives and it's totally irresponsible for us to shrug our shoulders at what messages we're sending. They were all very clear that they want to create commercially successful games (rather than games intended entirely to convey a certain message, regardless of profits). They want to make AAA games, but they want them to have some positive impact on the world, rather than negative.

I think their stance is totally reasonable, and they each gave several examples of how they are at least trying to do this. Molyneux said his favorite angle is to let the player do whatever they want (but with consequences) and then hope that the player will learn something about himself or herself through that. Others talked about how realistic World War 2 games gave their kids a better idea of what the Normandy Invasion was really like, how hard the odds were. They were all for Saving Private Ryan approaches that show people how horrible war is and remind people that 18 year old kids died on that beach.

Another panelist explained a game he designed about the French Resistance period of history, which he researched a lot and worked with a historian on. The theme of the game was that you cannot trust anyone, and yet you need to trust people to survive. As he said, this is an interesting facet of the human condition and exploring that is meaningful and worthwhile and does not preclude making a fun game, in fact it brings a new depth to the game that could be a great success. He also explained how he made the save/load screen into a calendar with pics and text about major events in the war. If you click on them, you get to read historical papers about those events. Of course, you can completely ignore this and it has no bearing on your progression of the game, but because it's on that save/load screen you see a lot, you might eventually get curious enough to actually learn something.

Other panelists had examples that were more mass market than that, but the general theme was that if you really want to make something that has *some* redeeming value, then can. There are many ways and opportunities to do so, rather than make a game that says how great and fun war is, for example. Note that they are at the level of Lead Designer or Creative Director or whatever, so they're assuming you're in charge of your project. Also note that in a proposal for a simple flying shooter about World War 2, I tried to show through an extremely short story sequence on each side of the battle, and through missions that were basically mirror images of each other where you got to play both sides, that war starts to feel futile. I could explain this in more detail so it doesn't sound stupid, but you get the idea. The publisher said "let's just take all that out." You see, they don't want to say anything because if you ever say anything about anything you might offend someone, and game publishers are generally very against that.

Will Wright
I won the lottery (literally) to go to Will Wright's talk. Will is just as much of a super genius today as he ever was, but I've seen him speak like 20 times now so I feel like I could almost give a Will Wright lecture at this point, or at least a parody of one.

He talked about fantasy worlds/franchises/IPs that are really successful (from tv shows to Godzilla to Disney to Starwars to Carebears). The bottom line from his talk is that he thinks the best stories/worlds have characters/settings/verbs that are the most easily deconstructable and separable because then you can play with them and recombine them in new ways in your head. And the flip side is that the best games, he says, comes from giving the player pieces they can put together in many ways, or at least give the player a way to generate their own stories based on what they do.

As an example to make more concrete, Darth Vader in Star Wars is very iconic and extreme and it's easy to separate him from everything else and just think about him. You could imagine him appearing in some other world or setting, and you can imagine what would happen. Spiderman has this great verb of swinging around and Harry Potter has great verbs about casting magic spells. You can deconstruct all those things into their components and imagine combining them in new ways. Properties that lend themselves to that lend themselves to "franchise" ubuquity. If something is going to be a movie, console game, phone game, lunch box, card game, t-shirt, etc, it's going to have a better shot if you can deconstruct and recombine the various parts easily.

That's it for today.

--Sirlin

Comments

[zem]

it's been two years since this was posted! (re: first sentence)

[Sirlin]

I have no idea what that sentence refers to. There was a game similar to Resident Evil: Mercenaries Mode that I was working on that didn't end up happening. That might have been it.