[sidyapa]

A tangent but, a lot of google's sites disobey browser standards and rules like for example sound autoplay on load. When you visit https://santatracker.google.com/ or youtube, it automatically plays sound without any user interaction, which is impossible for non Google sites to do

[andrenotgiant]

Google disobeys their own standards in MUCH worse ways. This year they are pushing a reduction in Cumulative Layout Shift (CLS). https://web.dev/cls/

But they purposefully use CLS in Search to increase clicks on Ads https://twitter.com/andyhattemer/status/1262564268890820609

[lima]

> But they purposefully use CLS in Search to increase clicks on Ads

You present this as a fact, but it would be absurd that Google would use such a cheap and easily detected trick to increase CTR. It would be bordering on ad fraud and I'm sure that Google, of all companies, knows better than that.

Occam's Razor says that this is a stupid async content loading bug, which they subsequently fixed. I've never seen this happen and when I just tried it without adblocker with that exact search term, it didn't - the page loaded with the ad.

[eitland]

> You present this as a fact, but it would be absurd that Google would use such a cheap and easily detected trick to increase CTR.

3 years ago and I wouldn't believed it at all but around 2 years ago I saw it happen consistently with a colleague at the desk next to me.

I cannot say for sure that it wasn't an extension in his browser but I can say for sure that I think Google has been really busy tearing down the mountains of trust they had before 2007 - 2009.

[miskin]

Similarly, thanks to async ad loading, gmail replaces first two items in my email list with ads with such a convinient delay that I accidentaly click on the ads more often than I would like to. Occam's razor would say that if it can bring more money, it is not accident.

[lima]

Accidental clicks are invalid clicks according to Google's own documentation[1][2].

For this to not be an accident, one would have to assume that Google actually makes more money from those invalid clicks, and that someone decided that yep, rendering ads asynchronously was a decent and legal approach at increasing advertising revenue, and requested the GMail team to implement it.

This kind of corporate misbehavior is not unheard of, but I just can't imagine it happening at Google.

It's much more likely that this is just unfortunate UX design to "improve" rendering performance without considering users on slow connections.

(I can reproduce this one just fine in desktop GMail - on the first render of the "Promotions" tab, the ads render asynchronously)

[1]: https://support.google.com/google-ads/answer/42995?hl=en

[2]: https://www.blog.google/products/ads/preventing-accidental-c...

[FridgeSeal]

"No you don't understand: invalid clicks are things that happen on other peoples properties. You definitely meant to click that ad in gmail. We know, we're google, you can definitely trust us about this"

'Unfortunate UX design to 'improve' rendering' is the plausible-deniability they can use to justify this.

> This kind of corporate misbehavior is not unheard of, but I just can't imagine it happening at Google.

I definitely can, I don't think anywhere is immune to this once you reach a certain scale. They have a profit-motive to make money, they will absolutely try and get away with as much as they possibly can.

[tsbertalan]

For example, the scourge of "people also ask" at the top of search results, that appears synchronously where the top result was a second ago, and has a randomly-generated container ID to prevent easy blocking. Not an ad, but, equivalently, content that I didn't ask for but that Google, for some reason, clearly really wants me to click on.

[hugey010]

Happens to me all the time, it's either complete UX heresy or ad fraud.

[andypants]

The android gmail app is horrible with this. They load a couple of your emails above ads so that the ads start on the second or third row.

And the re-ordering happens as your mails and the ads are loading! You might be about to tap your email, then the ads load in and you suddenly click on an ad. Or you want to tap the top row, but the app decides to put a different email above the ads and you end up tapping into the wrong mail because it was reordered just before the tap.

[tsbertalan]

I've also never seen ads in the Gmail app. Maybe it's because of G Suite.

[IndexPointer]

There are ads on the Gmail app??

[lima]

> [...] which is impossible for non Google sites to do

No, they don't. This is false. It's a mechanism called Media Engagement Index, Google properties have zero advantage, and any site can get a high score.

Chrome ships with a preloaded MEI assembled from global telemetry data, which is then trained locally:

https://www.chromium.org/audio-video/autoplay/autoplay-pre-s...

[hrktb]

You are technically true. It just happens that Youtube is the dominant video platform and gets pre-loaded in the default seed.

Would they have made the same choice of preloading a default seed if they had no properties in the seed ? who knows

[lima]

The whole point of Chrome is to push the web ecosystem forward such that Google can build better products on top of it.

[hrktb]

This is part of the plan, but I find this angle to give too much credit to Google.

Once they reached a dominant ad network position their whole strategy has been “advancing the web is advancing our revenue”, and it bled into mobile to the point where building and maintaining a whole ecosystem for free makes sense as long as they stay the search and ad engine of choice (that’s the only thing they’ll fight to impose).

Chrome is built in the same optics: push forward the web and webapps as long as search is theirs.

[userbinator]

The whole point of Chrome is to push the web ecosystem towards Google such that Google can exert more control over it.

[lima]

Exactly. They were tired of Microsoft doing it badly and realized they couldn't build on someone else's platform.

[robertlagrant]

When you visit a Netflix content URL it automatically plays sound and moving pictures without any user interaction! Evidence of Google owning Netflix?

[nosianu]

I read somewhere - I believe on HN actually, some time ago - that a number of high profile sites were exempted from this restriction, Netflix among them. Really, wasn't this a thread right here on HN, saying that this was anti-competitive, oligopoly essentially, making any other sites of smaller competitors and upstarts automatically worse off? I'm sure someone will be able to provide a link...

There are other examples where only the large sites benefit while everybody else has to play by stricter rules: "EU Parliament bans geoblocking, exempts Netflix and other streaming services" -- https://www.dw.com/en/eu-parliament-bans-geoblocking-exempts...

EDIT: User teraflop posted a link to the list of "sites that are allowed to autoplay video even without any prior media engagement" right here in this thread https://news.ycombinator.com/item?id=24818178

[geofft]

This is not actually true. There are no shortage of random news sites that auto-play sound. Reddit does too. Does Google own all of them?

[speedgoose]

Reddit and Twitter starts video with muted sound on my browser (Edge).

My guess from someone who had to develop a web video player at work, many websites will attempt to autoplay the video with sound and if it fails, it's easy to catch the failure event, they will mute the video and try again.

[geofft]

I'm talking specifically about Chrome. There's no web standard that says what a browser must do about autoplay requests, and Chrome permits a large number of sites to autoplay with sound on.

Web browsers are also capable of determining that autoplay on technically-not-load-but-automatic counts as autoplay. (There's even text in the spec about it.) In particular, they can tell whether it is in response to a user action/gesture on the site or not.

[speedgoose]

Chrome has some special logic about autoplay. The following page describes them, but I feel like it's a bit more complicated in more recent versions of Chrome.

https://developers.google.com/web/updates/2017/09/autoplay-p...

[silexia]

Wow, this is terrifying. I am a big supported of Google and dislike the recent attacks on FAANG, but this is shocking to me. If they are exempting themselves from this, what else could they be doing?!

[Hurdy]

The comment implies that this is somehow hardcoded just for Google sites, which is not true. Autoplay is allowed for sites with a high enough media engagement index. You can check chrome://media-engagement.

[teraflop]

The media engagement index is based on a user's past activity on a site, but Chrome has a special list of "preloaded" sites that are allowed to autoplay video even without any prior media engagement.

The preloaded list is in the source code (https://github.com/chromium/chromium/blob/master/chrome/brow...) but it's encoded as a finite state automaton that makes it a bit difficult to enumerate the list of whitelisted domains.

[neatmonster]

I made a small Python script to unpack the DAFSA in preloaded_data.pb.

Here is the code: https://gist.github.com/NeatMonster/e9cdb01441a3cd842e6a20fd...

And here is the plain-text list: https://gist.github.com/NeatMonster/e9cdb01441a3cd842e6a20fd...

[stingraycharles]

One has to wonder whether they intentionally obfuscate this list. It sounds like they “trained” a browser, and captured the resulting state. I’m sure you can argue this makes things more fair (we trained it using real world behavior!), but I really can’t give them the benefit of the doubt anymore.

[kaszanka]

It's generated by a Python script [0] from a list of URLs, but the input list doesn't seem to be included in the Chromium source (only the binary output of this tool).

[0] https://github.com/chromium/chromium/blob/615d5eed47c10d8890...

[lima]

> The pre-seeded site list is generated based on the global percentage of site visitors who train Chrome to allow autoplay for that site; a site will be included on the list if a sizable majority of site visitors permit autoplay on it. The list is algorithmically generated, rather than manually curated, and with no minimum traffic requirement. With the implementation of the autoplay policy for Web Audio in M71, Web Audio playback is also included in calculating the MEI score for a given site.

https://www.chromium.org/audio-video/autoplay/autoplay-pre-s...

[stingraycharles]

Will this not have some kind of self-reinforcing behavior, as the measurements are biased towards sites that are currently unmuted by default?

According to the MEI it actively measures user behavior and one of the most important measures is that a video is unmuted. From the document:

“The MEI is meant to allow media heavy websites (e.g. YouTube, Netflix) that rely on autoplay for their core experience. It is a non-goal to allow websites with a “good media behaviour” to autoplay without restrictions”

It doesn’t sound too good, and still doesn’t really explain how everything is seeded.

[ignoranceprior]

If it's a FSA can someone at least convert it to a regular expression or some other more readable format?

[moneywoes]

Is there no way to decode it

[cpeterso]

neatmonster wrote a script to decode the list and then shared links the results here:

https://news.ycombinator.com/item?id=24819473

[verroq]

Take a list of top X websites and enter it in every one.

The preimage space is finite and easily enumerated.

[drawfloat]

And media engagement is based on an opaque set of factors that just so happen to give top authority to Google sites.

[eznzt]

The source code is public.

[rurp]

That doesn't mean it's easy to parse.

[amadeuspagel]

Amazing. I once built a web app with autoplay, which worked for me, probably because I was using the app a lot which gave it high media engagement, but didn't work for others, and I never figured out the problem until now.

[kerng]

Well that's a nice way to say that its allowed for youtube and very few other sites... possibly none.

These are the kind of tricks a shady company would do. So disappointed what Google is doing to the web the last few years.

[YawningAngel]

I'm not so sure of that. My top sites by media engagement are: Spotify Twitch clips Youtube Twitch Eurosport Netflix The Independent Discord

It isn't obvious to me from this that Google are privileging their own sites above others here

[kretor]

Not "very few other sites", it's around 700 sites: https://news.ycombinator.com/item?id=24819473

[spencercw]

For what it's worth, Netflix has a higher score on my machine than YouTube.

[orf]

I loaded the page and went through a few actions, but I cannot see anything in chrome://media-engagement about it

[yreg]

I do see Santa Tracker in mine, it gave it a score of 0.05, the same as the web of my high school and less than say knowyourmeme.com which sits at 0.1

[liveoneggs]

they have always had whitelists for friends inside of chrome

[m463]

outlook (via web) also seems to be able to play sounds, like meeting notification sounds in firefox.

[vaccinator]

That site doesnt even load in my browser... I only see the Google wave (Firefox mobile v6x)... but on the other hand, there are Firefox extensions that make websites load as if you were using Chrome.

[circularfoyers]

Loads for me on Firefox 81.1.3 on Android. It did take a little bit to load, so might just be your internet connection.

[vaccinator]

It did load this time but no sound