Cyberpunk 2077 Game (E3 trailer video)

[Chief Blur Buster]

I think this is one of the games I'm most looking forward to now.

Play this YouTube video in 4K full screen mode for the full wow.

Single player role playing game. No in-app purchases and no lootbox!

Time for another GTX 1080 Ti to SLI my existing one in my box.

phpBB [video]

(But better yet, click here for the 4K version and go full screen, pronto.)

All that pretty graphics that would look great on the 4K 144Hz GSYNC HDR displays of sweetness!

[RealNC]

After the big letdown of Fallout 76, this is what I'm looking forward to the most now. I was afraid Cyberpunk was going to be an FPS, but fortunately they made it clear it's an RPG.

[mello]

Time for another GTX 1080 Ti to SLI my existing one in my box.
...
All that pretty graphics that would look great on the 4K 144Hz GSYNC HDR displays of sweetness!

First announcement was in 2012 and first trailer was released 5 years ago, so i suspect that there is still a long way till Cyberpunk 2077 will see its final release. If i had to guess, i would say 2020 or 2021. So until then, we will get both, better GPUs and better monitors

Also, a little off topic... do you guys think that AI in the future can help game developers create better games and make them in a shorter ammount of time ? What i mean is:

- bigger open worlds (with a complexity level close to a real big city or a small country)
- large number and complex NPC's (both meaningless interractions and side missions) based on the game narrative
- a multi-player open world with long-term end goal, where you might or you might not encounter other players ?

[Chief Blur Buster]

Also, a little off topic... do you guys think that AI in the future can help game developers create better games and make them in a shorter ammount of time ? What i mean is:

- bigger open worlds (with a complexity level close to a real big city or a small country)
- large number and complex NPC's (both meaningless interractions and side missions) based on the game narrative
- a multi-player open world with long-term end goal, where you might or you might not encounter other players ?

Yes, I anticipate a role for AI assistance of game development.

It has already begun at least at the library(ish) level, in some algorithmic generation of game worlds, to other things like that.

AI is already helping out. For example, the skin textures of a face, or the hair on the top of a human -- amazingly rendered in Cyberpunk 2077 -- may actually offload developer creativity to focus on more general things.

Who knows -- eventually 3D world programming may become more like directing a movie -- in some game world programming workflows:

"[Alexa|Siri|Holodeck Computer|Whatever], please put a futuristic convenience store in that corner. No, make it grimier. Pause all motion. Ok, show me a library of blue and red colored logos of convenience store. Now, add a 2176 flair to that. Okay, make that signpost more rusty looking. Great! Now generate about 10 color variants of this street corner and display it in a matrix. Oh wow, I love idea #3, use that instead. Except, make the sidewalk much grimier. More. More. Ok, that's perfect. Now put some cyan ambient lighting from that electronic advertisement. Brighter glow. No, too much. Perfect. Add one rusty robocourier. Ok, populate with people of the 2176 era, similar to System Shock 6 and Cyerpunk 2101, mixed in with all the existing NPC art I have. No, randomize again. Ok, crowd looks perfect. Play. Ok, crowd animates nicely with the rain. Pause. Add plot device, I need the rusty robocourier to bump into that punk. Use light depth of field effect. No no! No lens flare, don't ever use that again with this program. Just depth field. Okay, when robocourier bumps, the..."

While beginner programmers may do things like:

"[Alexa|Siri|Holodeck Computer|Whatever], please produce a 1930s replica of Paris, centering around the Eiffel Tower area. [Holdeck: Daytime or nightime? Summer or winter?] Yes, sunset in the summer. [Holodeck: Populated or unpopulated?] Populated. Play. Ohhhhh, beautiful!"

With the AI doing a lot of fill-ins, obviously.

Who knows?

[RealNC]

Meh, I can't see that happening. The results would most probably be bland and generic. Where I can see AIs being useful is in rapid prototyping. You know, the boring boilerplate stuff that happens before you can get to the actual real work. And in optimization.

[Chief Blur Buster]

Meh, I can't see that happening. The results would most probably be bland and generic. Where I can see AIs being useful is in rapid prototyping. You know, the boring boilerplate stuff that happens before you can get to the actual real work. And in optimization.

That's exactly what I said. In twenty years, anyway.

You can go advanced or as simple as you need to.

In the interim, you'll need to help the AI a lot more, but eventually it'd fill in a lot more -- and it is the skills of the person to get the best effects out. And if you still wanted to do traditional artwork, the AI can still integrate that too, and mix-it up on your command. The real professional stuff would look kick ass, while the generic stuff would be actually very useful.

Other spinoff elements can occur:
-- It can be as simple as an AI for assisting your personal artwork that can generate an amazing unlimited texture library knowing your likes and dislikes, for your own art.
-- An AI for converting ordinary photographs into 3D world. A much smarter and more intelligent version of this ... You'd then customize the resulting world.
-- It'd understand all the nuances that you need done, based on your existing styles of work, as well as ability to mix them up, or generate 100 "AI brianstorms" of a specific single piece of art that you do. You'd need excellent skillz to make sure all of it meshes properly. It'd accelerate amount of quantity of artwork that you can do for a game, allowing you to focus on the game rather than trying to design every single individual strand of hair, or the natural movements of a tumbling object, etc.
-- And hundreds of different examples

There's a lot of improvements concurrently in many areas that can occur, but -- later -- about a decade or so perhaps after the subtask AIs -- eventually it will integrate into a single AI that controls all the sub-AIs, in my examples.

A beginner would then be able to whip together a simple game -- in one day -- that looks far more advanced than Doom -- and people would still buy it for the experience. Tomorrow's stars equivalent to YouTube or Twitch. Some recreational glider pilot using a 3D camera could convert it into a funky dogfighting game. A cyclist riding through a forest with a 360 GoPro could convert their ride into a VR environment suitable for virtual drone racing or star wars speeder style racing after a few funky mods. And so on.

On the other hand -- it can be brought to way another level (Atari Pong versus Doom -- except it could be Crysis Original versus Cyberpunk 2077 ---- tomorrow it's another comparison between beginner capability versus professional capability). A professional or studio would produce an amazing immersive single player game story of your own dreams (Like a proper Star Wars world, etc) -- and people would pay a mint for the right professionally designed & curated experience that are so amazing far above and beyond what an everyman can do.

Take a look at humans generated by AI and converting doodles into realistic looking arts/photos ....

Zoom 20 years later, you can just show an AI several hundred historical black-and-white photos of Great Exhibition 1851 World's Fair -- and you will have an accurate full color 3D rendering of The Crystal Palace building full of the exhibits. You might have to enter the world and do some minor corrections (especially of the unphotographed areas) but that would be quick and easy verbal job rather than a laborious CAD job. The AI would have an "imagination" necessary to avoid photogrammetry, recognizing everything in the photos.

Such an game world creation AI of the year 2040 or 2050 knows the context, it's a world's fair, its around year 1851, the period clothing, and all the clues in the photos (e.g. planks of wood in black and white can still be recognized to be birch, oak, etc -- and the AI knows what birch looks like -- and then replaces it with ultra-Retina realistic birch in full color. Rinse and repeat for everything in the photo, and the thousands of photos of the historical venue seen. And then AI would intelligently know if something looks "off" and automatically fix all the seams with common sense & what a human artist would do). Then you can pay for admission to Great Exhibition 1851 in virtual reality -- the world's first world's fair, and experience the day super realistically in virtual reality. Or turn it into a game, with an early "Sherlock Holmes" style plot of an espionage between two participating countries. Or heck, turn it into a Fortnite-style Battle Royale complete with exploding and shattering glass of the Crystal Palace in the mayhem. The imagination at tap.

In another senses, an architect who might ordinarily designs a Las Vegas resort, could potentially also design a virtual vacation resort 100 times fancier for VR use, as an example -- the AI would understand blueprints, construction-style instruction, correctly intelligently interpret photos and artwork concepts into perfect 3D objects via logic & deduction (rather than simple "photogrammetry" tricks), etc.

Even today, games are very different from what 1980s were.

In 30 years, the technique of making games will be very different.

Certainly, programming will still be involved, but a lot of things probably will eventually move into AI-assisted design, while programming will shift into other territories.

Even today, you no longer need to necessarily know trigonometry to generate 3D graphics like you needed to in the early days -- all you have to do is design something in a graphics program, or CAD style application, drag it around, use your favourite engine (e.g. Unity, Unreal, whatever) and drop in all the beauties to your game world.

Yes, the professionals still go deep into the advanced stuff, but many mega blockbuster games of ultra-graphics-quality were made by people who no longer know the low-level programming stuff anymore -- this includes Cyberpunk 2077 too -- though a few key people do have to know the low level stuff like shaders.

But a lot of that is quickly becoming drag-drop stuff too -- you want simulated hair, here's a library of hair [automatically selects custom hair shader] -- as tools are being created, and eventually some shaders are algorithmically (AI) generated -- AI-generated software -- in drag drop libraries.

Eventually, hundreds of millions of modules (whether be opensource shaders or standard textures -- human generated and AI generated) would be on automatic seamless tap to a Holodeck assistant, to have perfect seamless realism on instant easy tap. Big studios would afford in-house "Holodeck assistants", while small indies would use a cloud assistant (ala Siri-style extrapolated 20 years into the future) balanced with doing things a little more manually after the AI helped generate the skeleton/template that the artists now take over. And mixing would become much easier in future -- "I want hair that looks between this and that" -- and the AI automatically writes a new hair shader program based on two different opensource programs. Small programming modules of these kinds would be AI written. It's already being done now (to a limited extent) and is rapidly advancing. While there is Skynet worries, the small time module writing would be massively simplified allowing you to focus on the really specific stuff that's hard to do by AI or all kinds of customizations. You want something unique, you actually have to do extra work to create a funky never-seen-before Mars City -- or a brand new Avatar-style world -- but it would be child's play to create a historical venue or an existing city, and/or make minor modifications to it, allowing developer/artist/whomever to really focus on designing a great plot.

It's an incremental path to this in the meantime, but eventually "Holodeck assistant" style scenario are one choice of artistic workflows that will be made possible.

It's a sign of the times...

[haanuman]

I am so impatient man, I can't wait for the gameplay video to be released, I have watched the impression video on Youtube for like 10 times and I just can't get enough. I just wish that it beats our expectations in a good way. I might need a new rig lol, using a rig since BF3 was launched.

[Chief Blur Buster]

....and the gameplay video.

48 minutes long, but more entertaining than most Netflix episodes!

phpBB [video]

Get popcorn, sit back and play this full screen in 4K on your HDTV. I guess I'll need a 2080 SLI to keep framerates at the high Hz. Hopefully strobed OLED monitors have arrived by the time this game does.

[mello]

I guess I'll need a 2080 SLI to keep framerates at the high Hz. Hopefully strobed OLED monitors have arrived by the time this game does.

Game was supposedly running at 30fps@1080 with i7/1080Ti, video was then upscaled to 4k and uploaded to youtube. And current rumors about release date are the end of 2019 or 2020, and the end of 2020 being most probable at this point. So, based on that GPUs will be produced at 7nm at that time (NVIDIA - 2080Ti+ or early rumors about 3080 series & AMD Navi+ or their next gen GPU).

Also a message from CD PROJEKT RED from the live stream:



This basically means anything can change and the final game will look different from what we saw.