Not impressed by PS4

[Sbrideau]

I see, but still, I thought this kind of DRM was already effective on the 360.

[Bubs]

I see, but still, I thought this kind of DRM was already effective on the 360.

 

You can play games on the 360 without having an internet connection.

[Giordano]

So pretty much whether they go with "Always online, but it's cool if you're not" or "Always online, and [bleep] you if you're not".

Yup, people worry about the latter.

 

Not everyone lives in 15MB/4G zones, has no caps, or willingly signs up to a reliable ISP.

[Estonian dude]

In Norway in summers I have no means for connecting my PS3 to the internet... And even now, half the time I have the internet connection on my PS3 off, since my internet is a bit too slow for good matchmaking.

I seriously will never buy a console or a game that requires online UNLESS it is a MULTIPLAYER ONLY game. I am not buying SimCity either, while I really much like the series and I would have bought it had it not that internet requirement.

[Riku3220]

Maybe it's all some sort of anti-hype thing. I don't think we'll ever get to the levels of hype that the Wii/PS3/360 had before they were launched so Microsoft is trying to get everybody to think the NextBox will suck and when it's finally announced, it will exceed their expectations.

[Powerfrog]

Maybe it's all some sort of anti-hype thing. I don't think we'll ever get to the levels of hype that the Wii/PS3/360 had before they were launched so Microsoft is trying to get everybody to think the NextBox will suck and when it's finally announced, it will exceed their expectations.

Advertising 101: It doesn't matter how awful your product is, make it look great. By the time they've bought it and realized it's awful you have their money.

 

Anti-hype is not a wise business move.

[Riku3220]

Maybe it's all some sort of anti-hype thing. I don't think we'll ever get to the levels of hype that the Wii/PS3/360 had before they were launched so Microsoft is trying to get everybody to think the NextBox will suck and when it's finally announced, it will exceed their expectations.

Advertising 101: It doesn't matter how awful your product is, make it look great. By the time they've bought it and realized it's awful you have their money.

 

Anti-hype is not a wise business move.

That only works for the short term. If you want a successful business that continues to generate money then you need people to actually enjoy your product.

[Powerfrog]

Maybe it's all some sort of anti-hype thing. I don't think we'll ever get to the levels of hype that the Wii/PS3/360 had before they were launched so Microsoft is trying to get everybody to think the NextBox will suck and when it's finally announced, it will exceed their expectations.

Advertising 101: It doesn't matter how awful your product is, make it look great. By the time they've bought it and realized it's awful you have their money.

 

Anti-hype is not a wise business move.

That only works for the short term. If you want a successful business that continues to generate money then you need people to actually enjoy your product.

I don't think consoles are really going to be around long term. We're reaching a limit to what we can do with games besides graphical improvements, the hardware requirements for top tier games are not increasing. Consoles will be redundant when the average "office" computer with decade old parts can support the most advanced games out there.

 

I'd be shocked if we saw anything after the PS5.

[jasignhagj]

I don't agree with the notion that games are reaching the limit of their potential. One of the major thing holding back PC gaming has been this ancient generation of consoles, which is the reason it seems like things are stagnating. But once the next gen comes out, I think we'll start to see a lot of improvement in PC gaming, especially in areas like game physics.

[Randox]

I'm holding out that we might someday start building hybrid neural/binary computer systems. Some interesting things could be done with a computer like that.

[Powerfrog]

It's just that we can already do the coolest things imaginable with existing hardware and some design tricks. Infinitely expanding universes? Random generation and skyboxes, no problem. You can even customize the "random" generation so that it gives you exactly what you want with a billionth of the power needed. The only limitation is how much time and effort the developers are willing to put into it.

 

The only thing we aspire to do in the future but can't do right now is improve the graphics of the same mechanics we currently have. And gimmicks like motion control and 3D.

 

Hit me with a possible game idea that would be impossible with a current gaming PC but might be possible in 20 years. Genuinely interested in being proven wrong.

[Alg]

The only thing we aspire to do in the future but can't do right now is improve the graphics of the same mechanics we currently have. And gimmicks like motion control and 3D.

Hit me with a possible game idea that would be impossible with a current gaming PC but might be possible in 20 years. Genuinely interested in being proven wrong.

I suppose the obvious "Make a better game" option is out of the question? Part of the reason I don't like this whole power arms-race is that it doesn't necessarily lead to better games. You could have something that's fun to play and genuinely immersive on basically any platform, but it really feels like a lot of the big names are more focused on showing off on a technical level.

[Powerfrog]

I suppose the obvious "Make a better game" option is out of the question? Part of the reason I don't like this whole power arms-race is that it doesn't necessarily lead to better games. You could have something that's fun to play and genuinely immersive on basically any platform, but it really feels like a lot of the big names are more focused on showing off on a technical level.

That's entirely my point. We don't need and indeed can't even utilize further hardware advancements for anything gameplay wise. There's nothing cool you can do with physics that can't already be done with current hardware. Thus, consoles are becoming less useful as the "average home" computer reaches this point without any optimization necessary.

[Ginger_Warrior]

The first half of your post makes sense. The second half does not, so long as Facebook apps and mobile phone games continue to be the real perpetrators of home console gaming's downfall. Neither of which require "average home computers", let alone gaming rigs.

 

Unless you want your game of Candy Crush to run really fast.

[Randox]

I think the coolest thing you could ever hope to do with neural based computing, likely working with traditional computer hardware, is self modifying programs. Why program a new biome into minecraft when you can just tell it what you want. You want a biome that forms a bowl shaped depression, with orange water forming a lake in the middle, and trees that are shaped like arches? Just tell the game that and it can make a custom biome.

 

This is ability for a program to make new content and assets that have never been defined for it in the program.

 

 

A perhaps more useful application of that ability would be in dealing with bugs. If you have a neural based program that understands how a program is supposed to work, and knows how to code or has some other means of modifying the program, then it should be able to observe when the program messes up, and either patch it, or for a game like Skyrim, perhaps use the concole commands to work around the problem. Autonomously self correcting software.

 

EDIT: One I forgot, because I like the idea of a computer that understand language. Take any RPG, and imagine what it would be like if instead of choosing what to say off a menu and listening to predefined responses, you just said what you wanted to say and the computer responded appropriately. Imagine the holodeck, minus the 3D virtual environment, where the charecters in your game are fully simulated personalities.

 

In the near term, none of this happens though. It's mostly useful for making existing computers better, since neural networks are many orders of magnitude more efficient in terms of power. By using them to handle the tasks they excel at, your reducing load on the other parts of the computer, allowing for computers that don't burn hundreds of watts, or just as a way to leapfrog what you can do with computers without needing to build faster processors. Everything above is simply to illustrate something you can't do with conventional home computers because no computer that could do that with transistor based technology would fit in your room.

[Powerfrog]

Self modifying programs. Why program a new biome into minecraft when you can just tell it what you want. You want a biome that forms a bowl shaped depression, with orange water forming a lake in the middle, and trees that are shaped like arches? Just tell the game that and it can make a custom biome.

 

This is ability for a program to make new content and assets that have never been defined for it in the program.

 

One I forgot, because I like the idea of a computer that understand language. Take any RPG, and imagine what it would be like if instead of choosing what to say off a menu and listening to predefined responses, you just said what you wanted to say and the computer responded appropriately. Imagine the holodeck, minus the 3D virtual environment, where the charecters in your game are fully simulated personalities.

This is more about developer time and effort than anything else, certainly not hardware limitations.

 

Custom biomes (From what I understand, you want the user to be able to edit the basic constants of something, which will then be randomly generated within the defined constants, resulting in a unique yet controlled creation.) already exists, both as a concept in many games and actually as mods in the game of minecraft.

 

These are in fact often used to develop games. Instead of spending years drawing every shard of grass on land, you'd set some parameters for what a shard of grass should roughly look like and then let it randomly create millions of shards within those parameters to get a full field of grass that looks like a lot of detail was put into it.

 

Understanding language has been done as early as text based adventure games. A more modern example of a computer understanding completely free text is the game Facade... Note it didn't do a very good job and is an awfully dull game, but it proves the concept.

[Estonian dude]

You know what is the problem with Siri and Google Voice and the such? Hardware they run on is way too weak to guarantee more than 80% accuracy in text understanding. There isn't hardware powerful enough to completely understand human voice recognition yet.

[Randox]

No, I think you misunderstand. To do what I am thinking of, you need an AI capable of understanding casual spoken language, and ideally pictures and hand sketches, and it needs the problem solving capacity to deal with abstract problems (which humans can't do until around teenage years).

 

You are essentially making possible a sandbox which is just made up of a very robust game engine and physics engine (in an effort to be able to deal with scenarios no one has thought of), and let the AI create the content to user specifications, bypassing things like modeling. Want grass in your world? Go take a picture of someone lawn and show it to the camera, and now you're computer knows what grass is and what it looks like, and you ahve something you can change and modify further at whim. Want a space ship? Show it some pictures of some federation ships from star trek so it can learn that artistic style. You want to do a cutscene with a hollywood style space fight? show it a few movie clips.

 

And I'm not saying that this can't be done on conventional hardware. By their nature, anything a neural computer, or quantum computer, or any other information processor can do, a conventional computer can do too as long as it is sufficiently powerful enough to simulate it. I am saying that if you want to fit it in a consumer desktop sized package, not cost millions of dollar, and not consume more electricity than a small town, you need to use the real deal.

 

Oh, and those text adventure games don't understand language. They only know the words and phrases specifically programmed into them, and have no ability to understand anything they were not coded for at creation. This would be one of those shortcuts I am trying to avoid, to allow you do things that were never imagined by the programmers.

 

@ Saq - Siri I think it depends on if it has web access or not. The phone isn't powerful enough to run the full voice recognition suite on board, so it normally uses an apple server for that I think.

[Powerfrog]

AI is software. Everything mentioned would technically be possible on a standard computer and doesn't exist already simply because of the time, effort and sheer understanding required to produce something so large and complex. Although certain hardware would be better, that isn't the limiting factor here. Optimization is just as important as raw power upgrades with this kind of thing anyway.

 

The text based games only recognize specific words or strings of text, but so do humans. A complicated AI would just be the same thing as the text based adventure ones but with a lot of work done on adding more words and phrases, and adding appropriate actions and reactions to words and phrases suitable to the characteristics of the AI and the previous conversations it's been in. This task can be aided by "learning" technology similar to cleverbot mixed with archives of actual conversations between humans (like these forums!), and it certainly doesn't require any fancy new hardware. Again, it's just an almost impossibly large task that hasn't been given much attention.

 

I just feel sorry for the poor voice actors.

 

Note I'm by no means saying there is nowhere for hardware to advance for stuff like energy conservation, power, and size/cost reduction, because there is. It's just the extra improvements isn't necessarily needed to run these seemingly elaborate and futuristic tasks. What we currently have is capable.

 

Just for the record there are devices that recognize images and intuitively learn things on their own, they're currently very simple and their improvement is very much based on improving the software and logic etc... Not the hardware.

[Randox]

...

And I'm not saying that this can't be done on conventional hardware.

...

 

Hardware is a limiting factor though, if you consider power consumption and space to be limiting factors. The transistor can not be made infinitely small, thus a computer with a certain transistor count (which limits computational power) can only ever be so small. To compete with the full power of a human brain, in real time, you're talking computers that use mega watts of power, and a computer running that many transistors is going to be pretty much the size of a small house.

 

You don't need quite that much AI for what I had in mind, but not that much less (maybe half?), and it's still way beyond what any desktop could handle conventionally. I'm talking about an AI on the level of Jarvis out of Tony Stark, minus the ability to talk back (though that would be nice, so there could be a back and fourth in the design process).

 

Anyway, you asked what a conventional desktop couldn't do. Artificial Human Intelligence is one of those things, and I've outlined a possible use for this AI in a gaming environment that a conventional desktop could not do as well (the AI would be able to design and render the models pretty much the moment you stopped telling it what to imagine, just like a human. A conventional based AI would take much longer to work through this).

[Randox]

But the whole point is to not script it at all, to give the user total freedom of creativity :lol:

 

you can actually do it with less. In fact, it could be done in a slightly limited fashion using all conventional hardware I'm sure, though I haven't a clue how to solve the issue of texturing. Computers suck major wazoo when it comes to pattern recognition, so while extrapolating a new ship or landscape from some samples is relatively easy with some awesome math and random numbers, it's quite a pain to do the details. You could perhaps treat an object as a whole bunch of objects in a relative formation, and with a library of texture patterns you would make something work.

 

Couple that to a a nice image processor (I'm not sure how fast a processor can do this, or how it's currently done. If a CPU isn't up for this, some modification to the GPU driver would allow access and use of a far superior set up for edge detection), and you have a good start. Some cool math that I am not versed in later, and a random number generator, and you could build new objects from samples, and also perhaps create areas from samples, using other objects as replacements.

 

Physics tweaks would be entirely different, and I'm not sure about designing gameplay. It's really more a fancy asset creation engine using the computer as a design assistant rather than using a CAD type program to make things. I don't have an idea for gameplay creation yet. As I learn more programming, I actually would love to impliment the last bit. A program to break down an image into profiles, and then recombine them. I think such programs do exist at least for 2D images, but I've never seen hint of one for 3D modelling (not that I have looked). If it works, I think the neatest application would be shape morphing, like corssing a fire truck with a car by linking the constituent parts to each other (front wheel to front wheel, rear to rear and so on) and stuff like that to see what it makes.

 

Not as awesome as having your own Jarvis (from Iron Man), but would be quite cool.

[Powerfrog]

As a general rule "I want to do something new" falls under software, while "I want to do something better" falls under hardware (and also software). Admittedly I'm really talking about generic computer hardware (CPU GPU RAM etc..) and not fancy new cameras that could possibly do cool new stuff with images.

 

Hardware could technically be a limiting factor for this kind of thing, but it is not THE limiting factor at the moment. You have a bucket that stores 1L, and you have 100ml of water. You should be focusing on getting more water before you look for a 10L bucket. Or, focus on creating an AI that would actually come close to needing better systems before looking for better systems.

 

Smart AI could be done in a game context with a current computer quite easily. Given the context of the game your AI's character would not be expected to know or respond appropriately to EVERYTHING EVER.

 

IE in a wild west cowboy game a sheriff asks you what you're doing in town. "Looking for a computer so I can play runescape of course!" The AI would understand everything but "computer" and "runescape" and could respond with a generic "Well we don't have none of these "computers" and certainly none of those "runescapes" so tell me boy, what are you doing in this town?" Forcing the user to give a suitable response the AI is expecting and can act appropriately to. A few 100 possibilities at most once it understands the meaning of words and can treat "I'm looking for work" and "I need a job" as the same thing. Humans really aren't as complicated or creative as we're given credit.

[Sbrideau]

Hardware could technically be a limiting factor for this kind of thing, but it is not THE limiting factor at the moment. You have a bucket that stores 1L, and you have 100ml of water. You should be focusing on getting more water before you look for a 10L bucket. Or, focus on creating an AI that would actually come close to needing better systems before looking for better systems.

 

 

With the current console gen, hardware IS a limiting factor. They're probably fitting 5L in your 1L bucket with the current games on consoles. Although with the fact that they do that kind of optimizing on the games means that there is a lot of space for performance improvements on computer games.

[Randox]

Your market is a limit. Games like Crysis are rare for a reason, the market wont support much of the practice of releasing games intended for consumer computers that don't yet exist.

 

We are probably running up against a human capital issue too, unless games want to get more expensive, which they might. If you want to be profitable, you can only spend so much effort on a game, and the top games are ever more demanding to create. Games like GTA4 already show that they are hitting that limit, where Rockstar obviously threw more money into the world than the game (and then used that mega expensive world in two more smaller games).

[Randox]

Was thinking about this again last night, and aside from realizing that I think we have both seen each other's point and are just arguing for the sake of it, I did also think of something that a computer can't do. It's a mathematical limitation to how they work, and to our own understanding of mathematics. I would consider it not unlike how it's not possible to accurately represent 1/3 as a base 10 decimal (or any fraction that causes a repeater), where a base 9 counting system can perfectly represent the same fraction as 0.3

 

Now, the random number generators we use are arguably a lot more useful for most applications, and there are work arounds if you need truly random data which can be better approximated by incorporating the closest thing to random phenomenon you can find (radio active decay tends to be a good one), but you can't do it mathematically. If you could, it wouldn't be random.

 

If you wanted to produce truly random outputs, your best bet is probably a quantum computer, because they don't operate on the strict mathematical rules we are used to. The quantum states that a quantum computer uses to represent bits (qubits) are probabilities, so I think you could run the exact same situation multiple times and get different answers, though in truth my understanding of quantum computing is superficial at best.