Ask HN: What do you think of the Google Stadia? I watched the conference yesterday and it all sounds great at the surface, but I also don't know enough about all the things talked about to understand. Is it good, is it bad? Are they targeting Steam as well as the consoles with Stadia? |
- Adoption will depend based on Internet speeds
- Cable companies and their data caps could result in a demand for high speed and unlimited data but not sure whether the tyranny of Comcast can be broken. Maybe a new lease of life to Google fiber? Just own everything from network to server to platform
Good things about Stadia besides no console, multiple devices
- Sharing state, youtube shortcut to get you a walkthrough video from youtube is god sent, lobby feature is also epic
Some really amazing things about this
- People working on the engine or 'hidden' console do not need to care about packaging it in a sleek way or ensure it weighs less. E.g. the hidden console could be ugly, and occupy a lot for space
- Google can build it in a way that is scale-able from a hardware perspective without worrying for form - just pure focus on function and performance
- For consoles, the constraints are form, power consumed, heat management and yet they are required to deliver high performance
- Add a VR device and unleash power of version 2 or 3 of the hidden console on it and you swallow the VR market as well -> Streaming VR games is the future
Speculation of the future landscape of gaming (2030) if everything goes well for streaming based gaming
- Microsoft and Google leverage their cloud infrastructure to capture significant chunk of the market
- NVIDIA also enters the market but either exits or ends up having a very small % of the market, because existing server and cloud infrastructure will play a big role in determining the winner and no one can beat Google and Microsoft in this regard. Doubt whether that gap can be closed by newcomers
- Sony is a little late to the market and grapple, they realize that their strengths lie in good narrative games and yet delivering great graphics, good gameplay and go back to that.
- There are a lot of multiplayer games ranging from high graphics to small games developed by Indies (making this industry literally the youtube of games), and most of the people gravitate to streaming games slowly
- There is still a set of console loyal gamers who buy the PS4 and they join Nintendo in a niche category
- Microsoft dabbles in both console and stream, doesn't capture the people's imagination with its work but still comes out with solid content in both places. They hold the 2nd place in market for both styles of gaming.
Stadia is not just targeting consoles, it is also looking at freeing up some developers from being forced to tie up with massive publishers. It will also result in smaller dev shops working on simple games where Google compete with Steam
Firstly, gaming hardware is expensive, especially if you want the latest stuff. My desktop computer is quite powerful, but I haven't had time to turn it on even once these past couple weeks. Making a big pool of hardware available to anyone on a time-sharing or seat-sharing basis makes gaming a lot more affordable. Of course, this only scales to hardware, not licenses. This is a big plus for just about everything, from environmental consideration to GPU demand. Except maybe investment in consumer-oriented hardware.
Then, there are the new possibilities offered by both, more powerful hardware, and centralized servers. As well as anti-cheat measures, and new financial models that can be explored. Bigger servers, fair latency, better splitscreen, etc. Are just a couple.
Since it runs on Linux+Vulkan, this combination will get a lot of favourable treatment over the next months, which might mean more native games, and at the very least better engine support for those open technologies (which is great for future-proofing, portability, etc).
AMD will likely benefit a lot from this, and virtualisation tech can be improved further as a byproduct.
--
On the other hand, it comes with a lot of disadvantages, and endangers the status quo:
Dematerialization is very dangerous for game preservation: cloud-exclusive games or features, games that can no longer be played after the license expired, or they became unprofitable to maintain, etc. This is a very powerful form of DRM.
Modding community will also likely suffer from this, as there is no way to modify a game that runs on a remote server (ideally).
Of course, you will need a decent internet connection to play.
This will increase user monetization (and tracking). I am afraid that with google's reach, we could see a lot of Stadia exclusives.
And of course, Google gets to say what's allowed on their platform and what's not.
Gaming might actually be one of the reasons why we still regularly see powerful computers around us, and not just smartphones/tablets/mass consumption devices. It's a limiting factor in the migration to an "all-cloud" life, anyway. So google wins on every front with this move, and could see the "desktop marketshare" start dipping again, and/or make powerful hardware harder to come by for a consumer.
--
On the bright side, VR might be the redeeming actor, as Stadia will never be able to play those games, which will maintain a baseline of Gamer PCs and traditional game publishing. On a less bright take, it looks like VR is going to be mostly used in professional environments in the next few years, and might end up confined in VR arcades.
For now, I personally think the possible drawbacks outweigh the advantages, and will refrain from using it. I think it would globally beneficial if it finds itself a niche, with no exclusives. I would even be willing to pay for that. But the prospect of another Embrace, Extend, Extinguish tactics is way to dangerous, so I would try not to add myself to the number of customers they can leverage.
As an aside, I would really like to be able to share my untapped computing power within my circles, and have my friends do so as well. We would likely need far less powerful hardware if we could pool this together, as a federated cloud. I guess the blocker here is trust, and algorithms to distribute workloads. Working on idempotent encryption algorithms and proof-of-work-like verifiable computing (check that the output is a valid compilation result, without recompiling ought to be doable) could alleviate some of the trust issues, and the second part will likely come naturally anyway.
1. Google doesn't have gaming in their corporate DNA. They will commit a bunch of faux pas that will alienate the gaming community.
2. Streaming games adds latency. Latency problems ruin the experience.
3. Gamers are a very fickle community and if they don't feel your company 'gets' them, they will not engage with you.
As for 2, Digital Foundry tested how bad the latency of Google's services was (though not in a home setting), and were able to show that it's actually on par with playing on an Xbox One X locally.
I'm not thrilled about this development, but I don't think these are reasons they'll be unsuccessful.
What percent of mobile game sales do you think go to "gamers?"
Conversely, gamers are probably a high percent of the medium-high end hardware community. Those companies manufacturing physical products do need to cater to gamers.
A pretty sizable subset of AAA games are heavily effected by reaction/latency time. It could be a solution for some kinds of MMO games though, where high latency is already designed for, and not having to distribute content updates would both be a plus for the players and allow the developers to keep "secret" content/events secured before they were intended to be made available.
Stadia doesn't have to steal all of steam's userbase to be able to reach critical mass for their own userbase and videogame streaming seems like a logical next step (like music and video streaming were) if a good UX can be provided (and I think they're aware of latency concerns and would not enter this market if they thought it unfeasible to provide a good gaming experience).
Disclaimer: I work for Google but am otherwise unaffiliated with the Stadia project
2) I can't imagine this going out to the public with latency issues. Netflix can stream video smooth as silk to millions very nicely. Streaming the game will be the same thing, we are just getting the video. Sending keyboard and mouse input should be trivial in respects to how intense video streaming is, so I'm not sure why there would be any latency.
3) These particularly picky gamers are a small subset of the community. You can choose to listen to them or not. There are millions of gamers who don't stream, who don't watch streams, who game on their own, who don't care if the files for the games are on their computer, etc. And these gamers want popular titles on their Chromebooks. Google is about to make that happen it seems.
2. Streaming a film and streaming the full content of Halo is a whole different ballgame. Do you know the kind of optimizations devs have to do to achieve the performance they do on pcs and consoles?
3. And how many Chrome book users is there?
I think at best this will be more like a mobile game platform. Won't put a dent on AAA games on PCs and Consoles.
The biggest issue I've had is internet. Outside of my own home the internet has been slow, laggy, or compressed the image to much. It's generally unplayable.
In between being able to afford a gaming device. I used parsec, and rented a machine to play games. FPS online multi player games were very difficult. The latency was to much for that. But other games, strategy, sports, etc. Worked fine. It was also great being able to connect anywhere and have my full game suite.
Right now I'm trying something similar in my home setup. I switched mobile gaming to old emulators and indies. With a bluetooth controller. (Ryzen 2700u). My home server is running a windows vm for gaming, and I stream everything over gigabit lan. That is working wonderfully so far.
I see this going two routes. For the casual gamer. As long as we have good internet. Then this can be a big win. Discarding the drm and ownership concerns.
For the more enthusiast / pro gamer. With the spark of 144hz freesync and gsync monitors. I don't see these streaming services being able to keep up. I still see a number of gamers wanting to build a gaming rig. I want to because I think it's fun. I just hope that this doesn't harm/limit any of the other gaming stores.
I'm still a fan of the service. I'm using their software for hosting my in home network gaming setup.
The thing that's horrific to me is the quoted 20GB of data transferred per hour during playtime. I don't have data limits so it doesn't affect me too hard but my Internet connection hovers right around 5mbips per second. I live in a smallish city in Maine. Google is alienating those in rural areas.
People said the same about Microsoft in 2001 or whenever it was that they launched the Xbox.
I like the that they don’t have any load time. It doesn’t look like it’s going to support most games any time soon. It’s going to be mostly hype for a year , I feel.
I don't think that future exists, and I don't think Google thinks it does either considering they abandoned the effort to make it happen.
without client and server side input prediction you get huge input lag - way above tolerable ranges.
with input prediction you get annoyed players that their character doesn't do what they want. Plus with multiplayer games you get extra lag on top of that. And lets not forget about local input lag.
Transfer speed requirements are sky-high.
Also - can i just own something? i don't want to perpetually rent things.
I would really hate it if google attained yet another 'monopoly'.
So you'll see a visual glitch as the world jumps to "what should have been", but that's a lot better than you missing your shot / jump / whatever due to latency.
The fact that their controller is wifi and not connected to the computer gives a hint that Google is not doing this. Since the controller is independently connected, it can have a different latency than the computer.
There is zero chance of a fighting game like Tekken ever being popular on Stadia due to lag.
Distance from LA to New York: ~4500km.
Time of light speed to travel 4500km: ~30ms
Time of 1 frame at 60fps: 16.7ms
There are moves in Tekken ('just frame') that require 1 frame precision to pull off. For a while it was popular to hook up old CRTs to playstations to reduce lag as much as possible.
People might think the network will get better eventually, but the speed of light isn't getting any faster. The network is no match for the local bus, and it never will be due to physics.
Use AI to determine the "most likely futures" and process those branch further.
It could dynamically increase and decrease the "possible futures" sent depending on how fast a person is going and how many input choices they have at any given moment. Compression (combining like elements of different future frames) can happen strongly on frames further into the future that have more time to be decoded.
"lol" is the best reply.
If you’re in LA and want to play against someone in NY, your example makes sense, but it would be technically infeasible even without Stadia, so...
However, if you’re talking about local split-screen through Stadia, what does the RTT to NY have to do with anything?
Google/Stadia would definitely have a dedicated West coast datacenter, if not one in LA itself.
Let’s say the nearest local DC is in SF. 600 km from SF to LA converts to ~10 ms RTT.
Assuming you have a good local connection, you should be able to get ~15 ms max response time playing splitscreen against your friend at home. That sounds very reasonable to me.
[0] https://schedule.gdconf.com/session/project-xcloud-the-futur...
The big advantage AMD has that no one really talks about is their ability to do client virtualization at the hardware level, not software.
[0] - https://www.amd.com/Documents/Multiuser-GPU-White-Paper.pdf
It's also plausible for games where latency isn't a big issue. Something like EVE Online seems like a good fit (I've never played, but I get the impression it isn't a quick reaction type game).
At the conference they focused a lot on first person action games and streaming to Youtube. Both of these aspirations are deluded. Action games are plausible if you're on a network with super low latency to Google's data centers, like say Google Fiber. Since Google just discontinued Fiber, this isn't going to happen.
Streaming to Youtube is nice, but in no way needs a dedicated hardware button.
There are definitely interesting things they could do with massively multiplayer worlds with low barriers to entry. There's real potential here, if Google manages to focus on this product long term.
If Google can stay focused. There's the rub.
For most games, 100ms latency between action and response is simply not acceptable. Even games like Rocksmith, Guitar Hero, and Crypt of the Necrodancer will offer calibration to overcome mere 10ms delays between the controller and the AV equipment.
I'm not a gamer. The only modern video console I own is a Nintendo Switch that I bought to play BoTW. My previous console was a Super Nintendo. I only own low-powered laptops and will probably never build a gaming PC.
But, if this remote-gaming takes off, and I can play any game from any console at any time by paying a monthly fee, count me in.
It is my understanding that this is limited to PC-gaming only. Heck, I'll even say that console manufacturers, who sell money by, you know, selling consoles, won't be very happy about this. But, in my opinion, it is an inevitable future. It is music streaming over owning CDs. It is Netflix over renting DVDs.
Everyone is hung up on latency, which is true. So if Google deploys their game rendering boxes at edge networking locations, like internet exchanges, that will result in lower latency to end users. Once Google has their rendering boxes there, they might as well sell other computing services as well. Why not sell to the game makers themselves, too - if they put their matchmaking/host servers in the same place, then latency is lowered even more.
IIRC Google already has the largest private network (most fiber-miles) with the most POPs compared to other cloud providers. The more POPs you have, the better you can do CDN and other geographic-sensitive computing since you are physically closer to your end users.
Having mini-datacenters at edge network locations could also be an interesting bet on future/emerging technologies that are also compute-heavy and latency sensitive, say VR or AR. Imagine being able to deploy code to thousands of locations (for a price) compared to the traditional couple dozen regions * a handful of availability zones where the huge datacenters are.
Hardcore gamers, and by that I mean people that spend a lot of time playing graphically intense games, have their own gaming PC's and keep up with hardware. So they aren't the customer, at least immediately, maybe in 2-4 years.
Anyone that has a current-gen console already has access to play most any AAA title game that comes out, unless it is exclusive to the "other" platform, which means it won't be on Stadia either. They aren't really the customer, at least immediately.
The boasting of Stadia GPU vs current gen consoles doesn't mean much to people playing on consoles. They already bought into a average common denominator of graphics capabilities. So they aren't the customer immediately, maybe when another generation of console comes out.
So who is the customer? Gamers playing non-graphically intense games? Why would they care? Maybe the killer feature here is being able to change devices (tablet, phone, Chromecast) and pick up where you left off. At what price point though is that worth it to someone?
"You mean it will be 100% impossible for piracy to exist?"
It's a great business idea, it's going to be terrible for consumers.
...except all the latency between the player and the game of course.
If Google can keep latency low and graphics as close as possible to original, then it has the possibility of being a Netflix for gaming. IMHO, Google need to focus on 3 things: latency, price and game library. Without those, it will fail or be a niche product considering that for 600ish$ you can build a pretty decent gaming PC.
I think Public Cloudx from Microsoft and NVIDIA Geforce is doing similar things. Not sure why when a Google does it, it's such a huge deal.
Either way, last thing I want is to have to have a Gmail account to play GTA5.
Not necessarily. Chromecast and the like already does side-channel communication to devices on the same network. Its how my phone can know that my TV is streaming something sent by my PC.
The controller could be sending input commands directly to the server, while also talking directly to the video streaming device.
I've gone the other way. I used to download and buy movies and music. I never enjoyed having to sort it all in my computer. Thanks to spotify and streaming services I really don't get have to worry about, "owning" these things anymore. Most of the time I didnt actually want to own the movie after watching it once.
I realized this recently about steam. I've got 500ish games. I only buy them because they are on sale and I know I'll play them eventually but don't want to pay full price for them. I'd happily pay a monthly price to be able to rent/stream/play any game in the steam library that I want. It would also solve the near constant issue with multiplayer games where I want to play with friends but they don't want to make a $60 investment for a game they may not enjoy.
20GB in an hour is about 44Mbps, so a 100Mbps connection is probably a good goal if you want Stadia to work well.
I'd expect high-speed connection uptake to be quite a bit higher amongst their target market, so over here betting than the majority of gaming users have 100Mbps+ at home isn't that crazy at all imo.
to the same effect, plenty of people want 4k and hdr, but arent film critics, nor videophiles, and dont actually care about data compression, bitrates, or banding.
My point was less about demographic and more about the network-latency concerns for specific common game-types. You cite correctly that they want to commoditize high performance hardware for a mass market, and note games like Halo/COD/Fortnight seem to have a very broad appeal. My question then becomes, can their solution commoditize high performance for the masses for these kinds of games? I'm curious to see how their solutions pan out, but there are a lot of road-blocks to achieving their goals. Graphical quality is only part of what qualifies as high performance in a game. In any FPS games like these game-play/graphical latency is a huge issue for play-ability. Developers go as far to take specific monitor hardware into consideration to shave down the latency. Stadia effectively adds a network loop connection between the keyboard and screen -- that's A LOT of a extra latency. Also, if this is going to target the mass market, then it has to work for the mass market and I think (in the US at least) there is going to be a large percentage of the population who's internet will lack either the necessary bandwidth, or network latency to Stadia's servers, in order to achieve a playable/desirable experience.
I think this segment will be a primary target of Stadia. If you want uncompressed 4K, 144hz, HDR with extremly low latency then your not currently playing on a console anyway and probably have a $2000+ machine sitting on your desk. For the enthusiast the tech is not there yet in streaming. But make no mistake. It will be there soon(ish). The day is not that far off when local machines will disappear for almost anything. And on a technical level nobody will be able to tell the difference.
Can you think of a AAA, "hardcore gamer only" game that is significantly more expensive, precisely because they know sales will be lower?
Latency can be fixed by moving the processing closer to the edge, like a CDN. Cloudflare and Netflix have boxes everywhere, google just needs to pop some mammoth consoles colocated inside isp data centers, and group people together connected to the same console. They can get there with 5G faster than itll make sense to keep adding gfx power to phones (battery life.)
Google is looking towards the future, where half the population lives in a city with cheap 5G.