The problem with that is that it creates a culture of complacency. If failure is acceptable and even expected to be the norm, then don't be surprised when lots of projects you thought would be sure-fire hits or are executed successfully by competitors end up failing.

This reads like a classic Dilbert strip.

Sure, we all know 9 out of 10 projects fail. Let's make sure we only ever begin the 1 that doesn't.

Found it: https://dilbert.com/strip/2000-03-08

“One out of ten research and development projects will succeed. I recommend cancelling the other nine.”

If you accept failure that easily, you will never try hard to turn anything around.

You need some sort of balance between the two.

What ever happened to "fail fast"

But don't fail stupid.

And don't fail the same way over and over again

Vine..

So how do you prevent mediocricy because every employee is afraid of failing because of the draconian consequences?

Why are you assuming the consequences should be draconian?

There might not even need to be any consequences in any given case, but there should be an appropriate and fair approach to rewards and possible consequences.

Because grand parents mentioned firing people...

Well to be fair Stadia was an idiotic project that was doomed to fail due to physical constraints(getting sub 80ms latency, and that's VERY generous for input latency that people are used to), which are unsolvable. No ML does not fix that, because if game predicts your inputs are you even playing it, not to mention the frustration when it mispredicts.

That's the sort of thing you have engineers for - so you don't sink money in something so idiotic. Not to mention that they picked the worst possible business model: pay price for title and pay a subscription at once.

The one who greenlit it, and proposed it should be under scrutiny.

The engineering side did a great job on a complex technical task. It's satisfactory, kind of works, not great, not terrible.

What boggles the mind is that nobody is able to formulate who is Stadia even meant for. Who are the target customers?

Draw a Venn diagram of people who:

  - care about gaming
  - are OK paying full price for their games
  - are OK paying 10 euros a month subscription
  - are OK buying an extra controller
  - have top-notch, 100% reliable internet connection
  - are willing to accept occasional technical glitches
  - are OK with Google's mercurial approach to product lifecycles

but do not:

  - have a computer
  - have a console
  - have a Steam library
  - mod their games
  - subscribe to Nvidia's cloud gaming service

Just giving some off-hand estimates, my estimate is 0.00015% of the world's population meets these criteria. The only reasonable business model must be "people will want to try it out and then forget to cancel their subscription".

Not only that, but also: what's the business model behind Stadia? Like how do they expect to ever make money off of it?

Right now they are still giving away the product for free if you just use 1080p, which the vast majority of their target audience will be perfectly happy with, if you care about 4K and 120hz you probably want to have optimal latency as well and just get a local box anyway. So where do they expect the money to make the service profitable will come from?

I've said the same thing so, so many times from even back when there was OnLive. Streaming games to paying customers is an interesting idea and some people will love it, but I just fail to see the business model. If you can pick up a cheap second hand console for less than 1 year of subscription service, who are the people who are going to keep paying for your service because they 'just want to play a game every now and then'? You have to build and maintain servers all around the world to make the service work, and you expect people who are not really into gaming enough to just buy a $400 console once every ~5 years, who just want to play every now and then, to be willing to spend enough on your service to cover the cost?

It boggles the mind how much money has been going into these kinds of services without any one of them ever seeming to have an obvious and solid long-term business plan...

Stadia could be a way into gaming for crowds who cant afford top-notch machines.. which is 5/6 of the population on this planet.

I think the person you’re replying to had the right analysis of why that doesn’t match the product they built: not owning a high-end PC, instant install/patching, etc. are a decent pitch but Stadia charges full price and doesn’t credit anything you already own.

I think it’d make a lot more sense if they didn’t charge full price, allowed rental/trial periods, or cut a deal to recognize your existing Steam/Epic/GOG purchases - making it a good deal for people who are traveling, don’t want to fire up a desktop at the moment, etc. – and worked on some unique features like hitting GCP strong points – AI at a new level or massive scaling for things like Sim-style games, etc. — so the value pitch wasn’t so narrow.

A top notch machine is only like 50% faster for 2-3 times the price. You don't need it. You can just get a console or a budget PC for $600. If you are crazy you can spend $300 on a PC and only play dota 2 and 2D games like Terraria, Don't Starve Together or Rimworld.

Also, what makes you think that 5/6 of the world has fast internet? I have good internet but if you asked me 3 years ago I would have said no.

Nah, that's what consoles are for. You don't even need the latest one sometimes. Consoles have existed before Stadia.

But Google makes all decisions based on data and research, we keep getting told.

The data is that previous directors that made huge gambles and got lucky are now VPs trying to figure out where they want their third vacation house to be and the unlucky ones got moved to new projects to try again.

Google's fundamental theory seems to be "users will always forgive us" or "there are always new users" and neither is true.

Disagree very strongly that it would be idiotic.

Nvidia's GeForce go is under the same physical constraints and Cyberpunk 2077 is very playable there (so I can play it on my IPad Pro using XBox bluetooth controller...) . There are lot's games where brief latency is not a problem. I agree for some games latency is a dealbreaker but hardly for all.

GeForce go also took down almost all of it's titles due to licensing issues where Nvidia were too afraid to upset their partners, which shows there is no business here and they should never have invested in the first place.

But at least Nvidia's initial conceit was that it would piggy back on someone else's content ecosystem rather than building their own - building your own gaming content ecosystem is a) incredibly hard b) a fifteen year plus project and c) costs a hundred billion dollars. Google weren't willing to do those but launched a project that needed them.

Do you people realise the there are whole generes of games,where beautiful graphics is a serious advantage, but low latency is not?

Like, 600 ms latency is still 100% acceptable and enjoyable.

And not only graphics. Turn-based strategy genre can also benefit from gigantic ram and cpu resources - all to make ai smarter, more enjoyable to beat the ever.

I think stadia and cloud gaming in general is a gold mine. But it is a different platform, games should be optimized for it.

You don’t need to stream video to run AI in the cloud for a game. Strangely given the existence of server infrastructure based games for decades we’ve never seen it. Maybe the idea isn’t that amazing after all?

That is quite true, the technology was here and nobody used it. That proves nothing, though.

You can apply the same logic to anything new. We had VR headsets before HL:Alyx and anyone could say "VR has no good games, idea is toast".

It proves quite a lot about the hype around cloud native and the things it enables.

And there were great VR games before HL:Alyx and Alyx is no killer app.

Oh there are games where input latency does not matter. You can usually run them on toaster too.

You can make amazing AI that beats humans 100% of time pretty easily, the hard part is making it challenging enough for people to have chance to beat it... It also just needs game state, without all the cruft of stadia(visual side, sounds, graphics, inputs etc).

Also did google partner with any of developers of such games?

Did they develop their own? or even formed any form of partnership for them?

No. They went for action titles, or titles with great graphics.

It was disaster from business perspective no matter how you look at it. Do not expect game developers to just bend over for google just because they release a new platform(and where are doubts about sustainability and quality of that platform from the beginning).

Toaster doesn't cut it when it comes to convenience, full hd graphics and depth of gameplay. Because, let's face it, toaster doest have that much compute resources and storage.

What kind of games will emerge 3 TB of assets will be totally fine amount? For a mobile turn-based game, no less. Or when 6 2080tis will render graphics for your phone no problem?

Modded Minecraft has deeper gameplay than 99% of all games ever made. If anything, high quality graphics make it harder to support deep gameplay because making high quality graphics for tens of thousands of items becomes very expensive.

Roguelikes like Cataclysm DDA also tend to be extremely heavy on the complexity side despite having "extremely poor" graphics by your standards.

6 2080TIs? How are you going to pay for that waste? Do you know how delusional you sound?

I have persistent feeling you are talking to some other person. I don't have standards you mention.

I am talking about future of gaming.

There are TVs with more than 100ms of latency, and there are strategies using state rollback and replay that can further reduce the impact of the latency besides. Latency isn't Stadia's hardest problem.

It is, as it is input latency stacked on top of display latency, stacked on top of internet latency!

Not to mention that 100ms of latency in TVs is constant usually, it isn't the case with network latency.

I play games on stadia, including shooters, and latency is not an issue. Latency is the reason you can't launch a Stadia competitor on AWS, but Google has a lot of infrastructure to build on. The latency issue is well handled at this point.

The problem with stadia for me is that gaming platforms are social networks and Stadia is relatively small. No one else I know has Stadia and for many games you can't even expect to find randos willing to play at all. Embr and GetPacked are basically useless.

You never heard of rollback netcode?

Why would anyone want this in a singleplayer experience? Or rather, why would anyone want two layers of rollback on their multiplayer experience?

How do you prevent mediocrity when every employee is nice, accepting of failure of others but absolutely useless?

Why is it that all of the most competent people I've worked with have been humble and empathetic towards failure, while most of the mediocre to not so good people have had harsh attitudes towards others? It makes me sad.

There is being human and there is being incompetent. Stadia did not fail because of bad luck. There were lots of obviously bad decisions made that went into this. If what you’re saying is true, then the decision makers moving on to other projects are just burning Google’s money without repercussions. Oh well, no one actually cares I guess.

> We are all human, we all make mistakes, in fact I bet 90% of the projects Google works on fail (public and private).

The fact that well over 90% of the products Google launches fail for very understandable reasons (I would argue that the last successful product Google launched was Android, and they technically acquired that) is why Google's structure and management need to change.