Google owns the chrome (top navigation) on all AMP pages. This makes it easy to navigate back to Google (why would you ever want to leave?) and other publishers in the AMP ecosystem, but much harder to navigate to the publisher that created the page in the first place
In essence, this means that what was once a publisher-owned page is now shared property: between the Google and the publisher. By controlling the top navigation, Google more easily controls the content the visitor sees, keeps visitors on Google longer, provides greater opportunity to track visitors, and perhaps most importantly has the opportunity to earn more ad revenue.
Now imagine if this was a requirement for ALL pages served in Google search results. You publish a page and it appears in Google, but when the user clicks on it Google has pasted a new navigation on the top of your page. This is exactly what is happening with AMP.
This is especially troubling in light of all the anti-trust controversies Google is finding itself in, both in the US and abroad. A recent study showed that 49% of all Google clicks go to Google properties of one kind or another (Maps, YouTube, Ads, etc) http://www.slideshare.net/randfish/intro-to-mozcon-2016/24-L...
Does AMP count as another Google property that will push more than 49% of clicks their way? Hard to say, but it's a disturbing trend for a monopoly and a hard pill for publishers to swallow.
I'd already say that in terms of publishing online that Google is the only search engine publishers/marketers care about.
No other search engine comes close to the quality of Google, at least for certain topics.
Outside of Russia, China, Japan and some other countries Google is the dominant search with a big lead.
And that is troubling indeed! Competition is good. And necessary for a healthy market.
It also means that Google has a huge amount of control over what information can be discovered.
What if MS decides Bing isn't worth their while anymore in 5 years? Yahoo will have dwindled away by then, leaving only Google. In Europe, Google has +90% everywhere (except Russia).
www.yandex.com is a really good alternative, I found it even offers better search results for most topic I look at.
Yahoo did switch to using Bing as its back end, some years ago.
The motto, the glibly pedantic color schema, the focus on simplicity on their home site, and cleanly (yet viciously) integrated ad revenue, even the language of the letter in the original link, all of these things point to a naivete that ignores how corrupt the ethos of what Google's central argument is: meritocracy above all else, in other words: down with the dumb, down with the lesser, all hail the perfectly manufactured data stream that contains what seems like childlike wonder, but is instead an amalgam of people's minds. They are sucking the intelligence out of us and selling it back at a markup so high it begs disbelief.
The fact that there is Bing is good, adn Bing's search isn't all that far below Google's in terms of getting good data from the web, but the point is: aren't we forgetting that there was a loooong time when we derived a lot of good data without their services? I find the aggressiveness of Google's ignorance to be quite frightening. I find that they are so far removed from the effects of their casual relationship with owning what happens to the data they are "curating" that it seems anathema to good governance, and governance it is because we are subject to the whims of their algorithm's ability to provide us with data that goes beyond what the mind can handle and handle sanely.
There is a monumental schism between what the web is comprised of and the first page of Google's query results, and that is what I think is at issue. What about all of the rest of the data that we never see? How do we know it's not interesting? How could we ever spend the time to sift through it to make certain we aren't missing data that is far easier to comprehend?
Long post, but a quick example: I program a fair amount. I like C++ for various reasons, but the C++ reference page is dry as the Red Sea after Moses played the downbeat of the 40 year cadence the slaves danced to after raising one too many pyramids.
I would wager that there might be a great C++ reference page out there somewhere that has easy on the eyes detail, a fairly conversational tone and nice examples that aren't at the very end of the page. But, because Google's algorithm (and Bing's, Yahoo's, Duck Duck Go, etc) all get tons of traffic from the cppreference.com page, it or some Stack response is the go to answer for essentially every question for which I need an answer. How is that well crafted page ever going to get the right recognition if everyone is being diverted to the cppreference page? It's like the old trope, we do it that way because that's the way we've always done it. Sometimes this means we drive the speed limit because that's the safest way to drive according the engineers who calculated the road conditions when they were built, but it's a little different in this instance because it's the _quality_ of the road that matters, not merely the velocity we drive that dictates the "goodness" of the experience.
Google (and the other big bots out there) are ignoring the quality in their algorithm(s) and merely settling for popularity as the guiding rubric, and I think we are all too aware of how that mentality can lead to unsalient choices.
If you think the megacorp that doesn't do customer support is going to be any easier to dislodge at other levels in the stack, once they manage to squeeze themselves in by undercutting the market, I'm afraid you're hopelessly naive.
1. Introducing incompatibilities, forcing closed standards etc. which block any potential competition.
2. Stop spending as much on product development.
While yes, you can install CyanogenMod (without Google Play Services), or others along those lines that remove the "Google" out of Android, it's true value (for most common users) is Google Play Services.
While you can say that you can get apps from apkmirror or F Droid, most apps still won't work without Play Services (case & point - Snapchat)
my entire point being - yes, it's very easy to have alternatives (and their importance can't be ignored), Google services (be it Play, or Gmail or others) are what people are looking for.
I was more thinking of other operating systems: we are fare from lacking functioning alternatives both on mobile and on desktop.
How do you replace search ?
How do you get competitive against mail, youtube without the amount of data google uses to sell ad on them ?
There was NO ads! You can actually read your email - and NO ads! That is amazing!
Its device-dependence exists because it's a unique dual-provider MVNO (Sprint and T-Mobile). This allows them to broker the best deal for the end consumer. But the bar is high -- devices must support dual-SIM, which is rare. Devices without explicit support aren't barred from Google Fi but may not get the full features and you can't get help with those devices if you run into issues.
Just like Facebook doesn't care about ripped video's on their platform. They know that publishers won't bite the hand that feeds them. They know that they can use their power as the defacto gateway to the internet as there is no competition and noone powerful enough to stand up to landgrabs like this.
Maybe it's even worse. They don't stop and think about the power and influence they have, because over the years it has become entirely natural to them to just take whatever they want, repackage it and call it their own.
What comes first? Hen or egg? Google isn't the internet, the web was there long before Google made things findable. You had to know how and where to look. And yes Google made things back then way better.
Non the less without content on the web a search engine is just blank. Unnecessary imho.
So Google making my content a shared property is something of concern to me. Therefore I am not using AMP.
Yandex is surprisingly good, take some time to try it out, whenever you think "oh shit results" and want to switch back to Google search, remind yourself how often you automatically rephrase your search when you dont find anything on Google either. Rephrase search, try again, you will find better results than what you had with Google.
There is less spam/ads, and you will discover whole new websites, and network of websites, some even foreign sites but in English.
It feels like using Google from 2004.
www.yandex.com and no Im not payed by them.
In fact, DuckDuckGo gets its results from over four hundred sources. These include hundreds of vertical sources delivering niche Instant Answers, DuckDuckBot (our crawler) and crowd-sourced sites (like Wikipedia, stored in our answer indexes). We also of course have more traditional links in the search results, which we source from Bing, Yahoo, and Yandex.
Still, duckduckgo is an US company, better stay away from those given the history of last decade.
Did they bring back the + operator? I still miss it even if you can use quotes instead...
This is modus operandi for many users, hence why Google inserting a big 'go back to Google' banner at the top, is a bad thing.
 Or from the standard user point of view, a good thing.
For the browser that is back, and if you opened the page in a new window then back means close it and reveal the previous page.
Maybe not everybody realizes that and uses it. I didn't at first but now it's a happy habit.
Also, given that it's hiding during scrolling now (or soon, depending on your platform) and you'll supposedly soon have the ability to define your own AMP menu, I think it mostly should calm the GGP's nerves.
On the desktop, I will open a bunch of tabs and close the ones that don't pan out. Not just after a Google search, but after anything resembling a search.
Mobile browsers allow this (though slightly awkwardly), but most mobile apps which display web content don't allow anything like this.
Given that landscape, what is added by a fat page wrapper that makes it easier to go back to the search engine?
Also, in regards to your up-thread comment, on mobile browsers I will sometimes use that technique because (but only because it can be slow to wait for multiple to load), but doing so is generally more cumbersome. Hold on link, select to open in new window. If I could just quickly visit each as needed, returning if it wasn't what I was looking for, I would much prefer that (as opposed to on desktop, which while I would like quicker loading, I do still occasionally want to open multiple tabs for results as you suggest). That is, I don't think this is a half-baked solution to a problem caused by making tabbed browsing hard, I think it's a solution that addresses some of the reasons we use tabs in the first place (specifically, to parallelize loading times for possibly slow sites).
What makes it hard? People click on links to navigate around the publisher's web site, and you can still have links in the AMP landing page.
I wasn't expecting the examples of its usage to involve altering the UI on AMP pages!
Is it just part of the blessed AMP WordPress plugin? Or is google actually injecting their own UI into AMP pages by default?
If the latter, that would seem like a power grab. Sure, it's "just a toolbar." And search results and adwords ads are "just links."
Yes. Google doesn’t actually link to AMP articles, but caches them on their own servers, modifies the markup, or removes non-compliant ad networks, and serves that to the actual user.
The web is supposed to be a platform, and AMP is supposed to be an open standard ... but the end result of this is that Google Search will modify your web application with their own navigation UI?
Yikes. I mean, injecting your stuff into other people's web pages is bad when a cable provider uses it for ads. Is it not bad for the web platform if it's google that does it?
Or is it linked to your usage of their CDN, and not something that necessarily comes with being an AMP site? (Ie, google will cache it for you, but first you have to agree to let them inject your page). That wouldn't be so bad, because it'd mean that you could play in the AMP playground without agreeing to Google having their fingers in your users' navigation.
But if AMP means "Google-injected UI" right now, then that just seems bad.
If Google indexes a page with AMP, and returns is at as a search result, it NEVER actually returns the AMP page, but always the modified one from the cache.
Maybe they'll make an exception for Germany and France, where the old image search is still live.
I still don't understand why it's mandatory to use the Google CDN, but I probably don't want to know.
The one upside I see is that Google-moderated AMP pages might push ad networks and library developers to develop less bloaty solutions to stay compliant. Those solutions might make it easier to develop regular non-bloaty pages as well.
I wonder whether we can get to a sane level now. Consumers do not seem to be prioritising sites with less bloat (otherwise there would already be a trend for leaner pages). Those who care probably use ad blockers and aren't the target demographic anyway.
the main goal of amp is to monopolize it further and make that last one they couldn't buy stop working.
It's like spying on everyone so we can make world safer.
I kinda just stopped reading Google news rather than try to fix it (:
* back button is broken 1/2 the time
* the bar wastes 1/3 of my screen
* I can no longer see what site I am on in the url
* it's hard to navigate to the / of the site
* I can't forward the link
* being on a good network in US it solves no problem that I have
What I'd really like to see is a way to opt-out of seeing AMP'ed pages in my search results. Or at least a way to navigate from AMP page to its native version.
Further, I noticed that AMP is a signal for low quality content. I am guessing sophisticated publishers are conservative enough to wait and see. And individuals haven't bothered dealing with it. So you get low-end publishers in between.
The AMP is really quite solid for the news use case, in my experience.
And bigger issues. My Galaxy S3 takes 4 secs to open Firefox without cache compared to 3 secs to open Chrome.
Non-AMP sites are frequently improved by the reader button. That's how I prefer to read news sites.
The only valid criticism appears to be how google displays search results using its own URL and this toolbar – it seems to break rather fundamental assumption about http and has the potential to break all sorts of tools that rely on the established structure of the web and open standards, as has already happened with the refer(r)er as mentioned in this response.
I wonder if there's a way to get the same result without rehosting content on their own URL. Couldn't they allow publishers to achieve the same result with a CNAME, possibly for amp.<hostname>.<tld>? And do the google servers add anything beyond being distributed caches? Because if not, it seems this level of indirection is redundant for websites already hosted on CDNs.
Regarding the toolbar: yeah, that's a terrible idea. I have no sympathy for publishers who object to it because if it reduces your retention rates there's probably more wrong with the content than the presentation. But as a user, it's the sort of "assisted browsing" that feels intrusive, like resizing the window or a "you need flash" popup (I don't).
Considering their market share in browsers isn't far behind the in search, I wonder why that function isn't just a chrome feature. Funny thing is: it's a feature that exists in Safari ("Search results snapback").
Since Google search is most likely discovering the AMP page via the <link rel="amphtml"> tag, they really should just use that URL. The Google servers do add a couple minor things beyond being a distributed cache though. First, there's swipe navigation to other articles if the link was located in the top carousel in search results. Second, if you are on a laptop, the Google servers will redirect you to the full version rather than use the AMP cache.
Congrats humanity. Your best invention ever and it only 20 years or so to completely fuck it up. This is why we can't have nice things.
I'm amazed that there is no federated (not owned by a single entity), anonymous search engine to this day and we're still pouring our brains into Google. Google knows more about a person than his spouse, doctor and lawyer put together!
People/powers-to-be would fuck it up, there would be "copyrighted material" and eventually it would be declared an illegal activity.
Only companies such as Google and Facebook prevail not due to technical reasons, but due to political power. They are selected by the likes of CIA, invested in and nourished to their position, and do their bidding. And alternatives are extinguished.
Google should not have existed had copyright laws been enforced as they are towards another search engine, thepiratebay.org
It can still be a good or a bad thing for the web as a platform.
Hopefully adoption is slow enough that the default behavior of any "blessed" libraries like this WP plugin can be encouraged to be less ... well, not evil, of course ... but really, why does that toolbar and 'X' need to be there? Browsers have a back button.
My take is simply that this is not a good look for Google, and hopefully they realize that. The article featuring the WP plugin, and these discussions on HN that give a lot of attention to the carousel, and how it competes with non-AMP results, don't build goodwill for the project.
Many users have no need for 'real internet'.
>completely fuck it up
No. 'Real internet' is still there, it's not getting absolutely smaller, only relatively to 'rubber doll' one
> Guess what happens when the "close" button is clicked inside the AMP view?
And the amount of disingenuous on this just rubs me the wrong way:
> If you are not comfortable with traffic on your AMP pages, please do not publish AMP pages.
You are on a search page, you go on layer deep by clicking. The close button goes back. I don't see what else it should do. Mind you that I do understand that a link to the origin server would be a good improvement (and I've stated that in the post).
I'm not sure what you find disingenuous. AMP are just web pages saying "HEY I A AMP, please use treat me as AMP". You can publish the same bytes minus the "HEY I A AMP" and get the old behavior.
No. Just no. The back button goes back. Period.
The close button closes modal windows overlaying the original page and originating from the same site.
I can already hear you countering "But.. since the content is served from google.com due to caching, this pattern is technically correct".
Which is tantamount to saying "We serve the content. So we can add whatever the heck we want"
Very frigging clever Google. Well done. /s
... or is it?
If I hit the search bar on my Android device, and it opens Google Search in an app interface, is it an app, is it a search engine, or is it both? What about with their new strategy of integrating Google Search more closely into apps? How about then?
The line is so blurry as to be irrelevant. The wording on the responses seem to be on a different side of that philosophical argument, but I'm not sure it is that black and white.
Yes but then you won't get the AMP icon next to your page when it comes up in the search results. Over time users will become conditioned to prefer clicking on AMP enabled links. With the icon you are basically categorizing the search results into "really fast sites" and "regular speed or slow sites", even though it is perfectly possible to achieve the same AMP speeds without using AMP.
Why are you introducing a new and totally unnecessary display element when the back button is built right in to the browser and everyone already knows how it works?
No matter how hard I try, I can't think of any good reason for this banner.
But hey, to the right of that X there sure is a lot of room for placing an adsense ad! What a strange co-incidence.
I can't give you a 100% reason for the X. I know multiple UI versions were tested and this did well.
What an amazing coincidence that this great new tech requires users to remain on google.com "for technical reasons".
We can do it in our native apps, and we do.
I guess I could use AMP but intentionally break my page's validation so it wouldn't be cached, but then I wouldn't get the special flag on SERPs, which effectively penalizes non-AMP sites.
I think it's a pretty fair position to take. If you don't like amp don't use it. We'll see if amp catches on over a period of time.
The users browsing the AMP results may never feel compelled to scroll down to see non AMP results.
So a publisher's choice isn't: "standard size for faster delivery", it's: "use AMP or receive less traffic than you're currently receiving"
That said, I find myself clicking AMP articles 5x more than non-AMP news links because I know the AMP articles will be fast and clean.
I'm still surprised so many people here are using google and not duckduckgo. I use the duck all day long. I used to revert to google when I couldn't find when I want, but I've found in the last six months, if it's not on ddg, it's not going to be on google, either!
We don't have to support google. Let's just stop giving them everything for free. People sneer at using Yahoo for mail, but what is the problem, really? It's almost all aesthetic.
Why give away our power and beg google to throw away profit and treat us right?
Persuade your friends to use other email and search providers. Design your websites so they have Links sections your visitors can navigate instead of going back to google. That's what I do and it works.
I'm hoping that Amp is really just Google training publishers how to make web pages that aren't terrible for users.
It is truly fascinating to watch.
At the time Amp was announced, Facebook's Instant Articles was really taking off. It was by far the best place to read news online and arguably still is.
To get a similar experience outside of Facebook (before Amp) you had to use an ad blocker. Right around that time, Apple added support for ad blocking in iOS and suddenly ad supported content outside of Facebook was looking unsustainable.
File an antitrust complaint against Facebook, sue Facebook, and get a warrant to get Instant Articles instantly shut down?
> suddenly ad supported content outside of Facebook was looking unsustainable
And just pivot regarding that?
AMP works on many platforms, Bing just announced support of AMP.
So, what's the big deal?
"The original idea behind AMP was to allow content to be distributed to platforms (such as Google, Twitter and Pinterest) in a way that retains [...] control for the publisher."
So Google sees itself as a platform in the tradition of Twitter and Pinterest - i.e. a controlled space on which content (with more or less control by the authors) is published. That's a significant difference to the gateway to the open web that they primarily still are at the moment.
That's why clicking "X" means "go back to Google", because in this model, you never left. You just temporarily viewed some content in a pane which you are now closing.
> We’re looking at ways to make the source link more discoverable and will update once that is done.
If Google was actually going to fix the issue, they would have said "we will make the close button direct users to the original site and will update once that is done" OR "we are changing the x (close) button to a ← (back) button".
"x" means "close" and "←" means "back". This is confusing UX at the least and arguably a dark pattern.
As a user, I hope Google do not implement either of your suggestions.
When I click the 'X' I don't expect to get a copy of the page that loads a bit slower, and I'd be surprised less than 95% of Google's users agreed with me (but by all means do a test).
And I don't want the close button changed to a back button. When I click on a search result I might then click one or two links on the publisher's web site before wanting to check out the next result on the Google results page. I want a single button that takes me quickly back to that search page - I don't want to have to click a 'back' multiple times (and maybe overshoot or undershoot my desired history position).
My current feature request is to link the "publisher domain" to the origin article, so that one can always click that to get there and use it to copy the link.
I've also sometimes been on a non AMP version, and Chrome has suggested "a version optimized for mobile". Clicking that has opened the AMP version and after reading the article, clicking the X took me back to the non AMP version.
This is working as expected and you need to do more work to make AMP work for you?
Why can't it work with templates, why does content that was created need to be created again for AMP? This doesn't seem very scalable.
Also, the attribution for ads and analytics basically means you need to reimplement your entire tracking code within the schema and spec of AMP's analytics attributes which only supports a subset of existing providers rather than allowing an abstract interface.
Also to the point that AMP doesn't affect search position, is this true if someone serves a shitty AMP page? Or is it only true that it won't boost position?
Perhaps you don't use it either?
So far I haven't thought of Google as a platform like Twitter or Pinterest. I thought there was a free web with Google and Bing and good old Altavista searching and indexing it. Is making a website now similar to posting something on Facebook? Or why would they have to assure you that your traffic is your traffic, "period"?
If that's the case we need an urgent change in direction!
Note, It has been suggested that a speed rating on Google would be equivalent to the amp experience with less Google control.
Page render speed is just as important as server speed. A website bloated with ads is what I object to the most, or any type of pop over that obstructs the content.
I believe that page speed is also important, google provides PageSpeed Insights to help you improve your site. 
 PageSpeed Insights https://developers.google.com/speed/pagespeed/insights/
What matters is if one gets the results he was searching for. And somehow those pages that have "amp" written nearby open faster, so let's click more on those.
Market pressure will drive decisions on whether it's better have amped pages (with those claimed drawbacks) or try to capture attention to the whole site with navigation and lighter pages.
And as a prisoner's dilemma, if sites that don't have amp are as fast as the amped, having "amp" near your link won't matter. If they are slower, the distinction will matter, independently of your specific optimization.
You just admitted the purpose of AMP being moot.
TL;DR Google AMP also hijacks app deep links
> The Android users might have already noticed that it is now scrolling out of the way
It doesn't scroll away. Just checked on Google news with both Chrome and Opera. Android 6. Am I missing something?
An external contributor is currently building first class support for "My simple ad network that just wants to put an image with a link on the page" https://github.com/ampproject/amphtml/issues/5541
There are other ways to do it, though.
I think with the numbers of ad tech companies coming on board every week one can safely say that the process is working for them.
Since I normally use Firefox Mobile (has to try it with Chrome), I don't get the AMP icon and I'm taken to the requested site. I'm guessing this functionality is limited to Chrome and (maybe?) Safari on mobile.
AMP is attempting to solving a challenging problem, and although I don't personally agree with their solution, I've gotta recognize that opening the embedded AMP version of the page from the result of "git tips" was faster on Chrome than on Firefox. I'm hopeful that the lessons learned from this will be pushed upstream and help improve the web.
There’s been discussion about adding Google Search to the list of sites for which Firefox Mobile always serves a faked User Agent because of this hostile and anticompetitive behaviour before.
This is deceiving.
If Google's going to cache the page, then there should be a "cache" link so the user knows that they're clicking on a cached version of the page. Otherwise, deliver the user to the site that they think they're clicking on.
Why exactly is it permissible for Google to "cache" my page and then serve it up to users?
That's literally them copying my page to their own servers, which is my content, my intellectual property, and then serving it up however they like.
I'm a little confused how that is legal.
Of all his 725-words response only 18 words directly address the problem:
>We’re looking at ways to make the source link more discoverable and will update once that is done.
That's 2.5%. Is he really a tech lead?
Then you go on to imply that he's unqualified to be a TL because you don't like his polite and conversational tone in an article that criticized his project?
I just don't understand HN anymore.
A appreciate the fact of quick reply, even if half of the credit for it should go to HN discussion.