It appears Safari implemented a special case. We'd prefer a more generic solution where browsers would share the canonical link by default, but this works for us.
We had that solution it was called 'the web'.
I'd prefer a more generic solution too: get rid of AMP.
It really isn't about what 'works for you' it is about what is good for Google versus what is good for the web.
If you want to incentivise not delivering multi-megabyte "experiences" to people, I could see a case for feeding page weight into the ranking algorithm. Then people would be less inclined to have tag managers loading tag managers loading a zillion random js trackers.
I have personally saved at least one decent laptop from landfill by installing uBlock Origin on it, and making the web usable again.
That would get rid of the cruft on the web in a heartbeat.
Every SEO consultant would suddenly have an actual job, playing code golf on CSS and HTML to deliver the content in the minimum number of bytes possible.
I mean, kinda, but not really. You have to cater to it specifically
But your average news page would weigh in roughly the same no matter what the source was. Assuming you'd strip out all the cruft.
For reference, I optimized my website a while ago and ended up with the average page being < 15K.
Try it, on any connection in normal use today it should load faster than you can blink.
Add a few pictures and it would still work quite comfortably.
Not all of them. For a trivial example, an in-depth article about a topic would be penalized more than a more epidermic one, because it would contain more text, graphs, etc.
> For reference, I optimized my website a while ago and ended up with the average page being < 15K.
I feel you, I absolutely hate that my website uses a bunch of stuff for almost no reason. 80 KB of CSS or 90 KB of font isn't awesome, and I started a project to provide information on how to create lighter websites (https://www.lightentheweb.com/, stalled a bit).
All I'm saying here, though, is that the solution is more nuanced than "heavier = lower". AMP is definitely not a step in the right direction, though. I wonder if Google could give sites that used a specific CSS file and no JS a bump, rather than loading things on their own domain.
But such trivial examples should be trivial to get right no? It's all about the relative weight of the various factors that determine the ranking and a 'lightweight' article could be recognized as such.
How do you recognize which is which?
Sounds like a publisher-assisted "readability view" would solve both problems, though. It would allow the user to view a "lighter" version of the page and search engines could just measure how large the "non-light" version is and penalize accordingly. I'm afraid that would incentivize the publisher to put ads and other crap in the "light" view, though.
You as the publisher/author have to implement an AMP page yourself and then alert Google of its existence.
A lot of the useless pages are 3 stories of scrolling through large images with very little text.
Also, "data" as in tables and small/vector charts isn't all that big.
Tell that to a photographer, designer, artist, illustrator, etc
I don't think that would be a good metric for relevance though and in most case, the total size isn't an issue. Which is why AMP is good, low file size when it's needed, full size when it's not.
For me, I'd given up on trying to browse the web on my smartphone until Firefox Focus came along. I realize Firefox for Android now allows extensions, but I still used it as a typical browser. I signed into my accounts, I had expectations for a personalized experience, etc. With Focus, it's just a quick visit to a site and then it disappears. I can find the information I want, quickly, yet again.
User here. I like AMP. At the end of the day, I want to read an article. Most of the time, Reader view works. But not always. AMP reliably and quickly renders legible web pages. This wouldn't be an issue if 90% of newspaper websites didn't look like they were designed by turnips.
If Google really wanted to improve the quality of the web they had excellent tools to achieve that: the weighing of the various elements that go into the ranking algo.
But instead they chose to do an end-run around all standards processes and to attempt to capture mobile content (which is an extremely valuable slice of all web traffic) at the expense of competition and the openness of the web.
And so you are now a very willing pawn in the endgame for the domination of the world wide web, which will eventually result in Google being the sole provider of your content by virtue of the web being consumed more and more through mobile devices.
If you're not concerned I totally understand, but maybe you should be.
Doesn't Google already do this ? There are unintended consequences to increasing speed's weight in search rankings. Principally, it disadvantages smaller players. (Also, the New York Times would consistently rank under Buzzfeed.)
> you are now a very willing pawn in the endgame for the domination of the world wide web
Most people get news from the walled gardens of Facebook and Twitter . The minority of us sharing links to newspaper articles probably also have direct relationships with publishers. For example, I read the New York Times, Washington Post, Economist, Bloomberg and other papers directly on their sites because I subscribe to their newsletters.
> maybe you should be [concerned]
Maybe, but I don't think so. The things I read about in the newspapers AMP gives me faster access to yield better (i.e. more serious and more actionable upon) avenues for my attention. In every context where AMP is a worry journalism's decline seems more important.
Well, assuming equal relevance yes. But the NYT is not exactly a small player and could easily up their game in the page speed department. Besides that you could slowly ramp up the weighing to give parties the time to adjust.
> Most people get news from the walled gardens of Facebook and Twitter .
And this is what worries Google. Their long term adversary is Facebook (not Twitter, at least, not nearly so much).
> The minority of us sharing links to newspaper articles probably have direct relationships with publishers.
Not necessarily. I use a variety of aggregators.
> For example, I read the New York Times, Washington Post, Economist, Bloomberg and other papers directly on their sites because I subscribe to their daily newsletters.
> The things I read about in the newspapers AMP gives me faster access to hand me better (i.e. more serious and more actionable upon) uses for my attention.
Yes, it's a very seductive proposition, and one that given Google's power in the marketplace they just might get away with. See also: Microsoft bundling their browser and tying it in at every level possible in to the OS and other abuses of monopoly power.
> In every case where AMP is a worry, journalism's general decline is more important.
But that's the same kind of argument that people use when India has news about their space program: Why don't they fix hunger first. It's possible to both worry about journalisms decline and monopoly power abuse on the web, especially when it concerns the distribution of news.
When I'm looking for news, I want the highest-quality source first. AMP gives everyone the option of going fast now while kinks are worked out back home.
(I should mention that my default search engine is Duck Duck Go. I usually !news (Google News), though, because I like Google's news results better.)
> it's a very seductive proposition, and one that given Google's power in the marketplace they just might get away with
They might. But a lot of things might happen. AMP makes good journalism more accessible. If Google abuses that privilege down the road, it's not like they're legally unassailable.
> It's possible to both worry about journalisms decline and monopoly power abuse on the web, especially when it concerns the distribution of news
Lecturing newspapers about why they should have had speedier websites went nowhere. Seeing that the same content produced more views through AMP got the message across.
I don't think the analogy to India's space program is accurate. Technological development lifts standards of living. There's a solid argument for doing both because each helps the other. Killing AMP and forcing to choose users between crappy websites and efficient walled gardens is counterproductive.
Well, no, you want the highest quality source first that loads fast. Otherwise you wouldn't care about AMP in the first place!
> AMP gives everyone the option of going fast now while kinks are worked out back home.
So, do you expect that Google will ever give up their stranglehold on the sites that use AMP once they are in?
In my experience such measures tend to become permanent fixtures with their own hooks embedded in the various fabrics making removing them hard to impossible once they have achieve critical mass. We may have already passed that point, in which case you can safely ignore me.
> They might. But a lot of things might happen. If AMP makes good journalism more accessible I'm all for it.
That's your privilege. I have yet to have a day where I could not read more news than I had time for so I'm not familiar with good journalism not being accessible but I concede that this may be your situation.
> If Google abuses that privilege down the road, it's not like they're legally unassailable.
Who will take that up? You? The EU? The publishers?
My guess is that once entrenched it will be impossible to get rid of, a parasitic element in the pipe between content producer and consumer.
> Lecturing newspapers about why they should have speedier websites went nowhere.
> Seeing that the same content produces more views through AMP, however, seems to have gotten the message across.
Well no, I still don't see news websites getting any faster. But they're already ridiculously fast, consider that 30 years ago you had to wait a full day to read the news and now you can read it in 10 seconds. It's great if that can be cut down to 0.01 second and I'm all for doing that and incentivizing those companies to do this but in the end it is their web property and the market should be allowed to sort this out rather than that a search engine monopolist that is already the gateway to almost all pages read on the web becomes the gatekeeper to our news.
> In the meantime, Google's cash flows get to subsidize this IT expenditure for the news industry while consumers get faster news sooner.
No, Google's cash flow will not subsidize this because they now have the eyeballs. Sooner or later there will be a push for monetization.
> Killing AMP and forcing to choose users between crappy websites and efficient walled gardens, however, is counterproductive.
There are other choices besides.
AMP isn't forcing me to read fast newspaper A over slow newspaper B. It's letting me read slow newspaper B fast. I've never clicked on an AMP link for a source I didn't know.
> In my experience such measures tend to become permanent fixtures...Who will take that up? You? The EU?
You made an analogy to Microsoft bundling browsers earlier.
> I still don't see news websites getting any faster
They are .
> [websites are] already ridiculously fast
That's your preference. Mine is for them to be faster.
> Google's cash flow will not subsidize this because they now have the eyeballs. Sooner or later there will be a push for monetization
AMP is open source. If we'd gone the penalise-slow-pages route, every newspaper would have had to "roll their own" AMP in house. Even if Google throws a tax on later, readers got fast pages sooner and publishers got free code.
> There are other choices besides [crappy websites and efficient walled gardens]
In theory. In reality, that's the battleground.
That's called HTML. Nothing special.
Maybe if you're using the Trump definition of fast...
True. But I do care what tech makes the magic happen and the consolidation of the web into a very small number of silos is a dangerous development for many reasons and AMP is one plank in that consolidation game.
> Google is actually in a position to fix the problem and why not let them benefit while at it?
Because their stated reasons for AMP do not align with the implementation details, AMP grabs way too much control for a simple improvement in speed.
And to give credit where it's due, I'm glad you do. While we disagree on this point, I have a great deal of respect for your comments and found our discussion enjoyable.
Your analogy may be reasonably apt, but if so, it shouldn't make anyone who makes their living thinking about systems sanguine.
It does turn out many consumers will trade ownership for convenience. Particularly if convenience comes at a fraction of a cost. Meanwhile, there are side effects: the people actually creating the product (recorded music) get paid a vanishing fraction of what they used to. That changes the economics of actually producing recorded music, shifting the ability to do it to people who can get their money elsewhere. Some people like to try and obfuscate that reality with talk of "new business models" and "innovation," but it all boils down to the assertion that people who make recorded music should have to do another job in order to provide cloud record collections like Spotify to a consumer for the cheapest price possible.
The consumer, as you've pointed out, doesn't want to think about this. Whether they think about it or not, it will shape what kind of music gets produced and by who, so perhaps they should think about it, even if the incentives are more long term and non-obvious.
Buuut even assuming they don't care to do that... it doesn't mean that people who do care to think about how that affects the industry don't have every right to "moan about the philosophy."
Same goes for AMP. Solutions like AMP will absolutely have second order effects. Many users don't care to think about them. Doesn't mean they won't be subject to those effects, that those effects are just about philosophy, and it surely doesn't mean that people who are interested in them have any obligation to stay quiet.
And of all the things to actually have this really be something people waste time on. Working out the economics of production and distribution of music in a time when making copies is just short of free is at least actually a tricky problem. The idea that making mobile pages that load fast is a tricky problem that requires a Google engineering solution is utterly ludicrous.
This isn't about "philosophy." This is about systems and consequences. The consequences are real to the user whether they're aware of the chain that produced them or not.
> Yeah, artists may get paid less now, but Spotify and its ilk pretty much removed piracy from the vocabulary of regular consumers. Back in 2004, music piracy was just something you did
2004 is a poor point of comparison if what you want is to get the big picture or even just answer the narrow question of whether the choice really is between cut-rate cloud record collection services like Spotify or free-for-all piracy. 2003 is when the iTunes Music Store launches and 2004/2005 is when you really start to see the rise of digital music retail. Between then and the early 2010s you see those services become widely accepted (even embraced by non-technical people) as an alternative to piracy, and you see revenue from those retail services rise just fine without streaming. In fact, I've seen some reports that suggest that by 2012/2013 the profits from digital retail was on its way to the profits from physical formats.
So the choice wasn't necessarily between fiddly inconvenient piracy and butter smooth streaming experience.
And the battle doesn't even necessarily play out on the field of consumer choice, really. I mean, in a perfect world I might well expect enough consumers to recognize how they vote with the wallet will shape the world, that when we value McDonalds the economy produces McJobs, that when your airfare dollars are ultimately decided by the lowest price you'll get a shitty airline experience.... but yeah, people don't. That's actually why it's more important for people who can see the consequences to discuss them, publicly, loudly, maybe even forcefully enough that product owners who can't see past their A/B testing (and may not even have arrived at the points where they understand the limits of that along with the value) might pay attention.
> My problem is that a lot of developer-type solutions exist in the Silicon Valley echo chamber and will never see adoption because they don't take into account the big picture....
We're concerned precisely because we see the big picture. Understanding that Google has the power to shape the landscape regardless of how that effects the value of the landscape as a whole doesn't mitigate the responsibility to talk about it, or imply that there's a missing larger picture.
> Google will do what they want to do and they're big enough to where if they lose a few million hardcore advocates in the process it's no skin off their nose.
If you're right, and we've reached the point where Google essentially no longer has management and engineering talent that cares to be a good steward of the web or no longer has the incentives to understand how AMP isn't, then that's a much heavier indictment of Google than most of AMP's opponents have leveled so far.
And that's even if you can get everyone to agree that weighing the different aspects of a change says it's an overall benefit.
Even if AMP pages would render 1000 times as fast it wouldn't matter, it would still be an abuse of power. The relationship between a reader and the source of their content should not be mediated by Google. What point is there in HTTPS everywhere campaigns and such if Google MITMs every news article you read? What point is there for a content provider to even have a website if Google will end up serving the content?
This is simply bad, the only upside is the speed gain and that is one that could be sorted out by the market with a gentle nudge from Google rather than by some kind of monopolistic end-game on their part.
What I don't understand is why they cache the page themselves, and make that the requested version.
It would seem like any cached page from a CDN would be just as quick.
If Google preloaded the page from a non-google server, there would be a few problems. The most important one is that it would violate the user's privacy. That server would have in it's logs a request from a user that didn't click on the article, and could intuit information about the user's searching behavior from that request.
Less importantly, without the page being served from Google's cache, there is no way to guarantee it's actually a valid AMP page. The only thing Google knows is that it was valid AMP the last time it was crawled, which could be days ago.
And a waste of bandwidth if the user never clicks.
That's one problem.
> without the page being served from Google's cache, there is no way to guarantee it's actually a valid AMP page
Yes there is. You can validate pages as AMP on any CDN. You just can't have any benefit unless they're hosted on Google CDN.
TL;dr AMP is 100% build on web tech. Pre-rendering achieves the performance but relies on history.pushState and iframes, which doesn't allow presenting a URL that is not on the same origin as the search page.
We think that the UX trade off in the URL is OK given the performance benefits. Given that it is a trade off we now have a bunch of projects to mitigate those trade offs. More coming soon, including lots of improvements to Safari. My team works directly on WebKit to fix bugs that affect AMP (but also the web as a whole, since AMP is just web tech).
Also, is someone working on a successor to AMP that doesn't break or try to replace the browser's own UI? Maybe something done at a lower level, maybe the HTTPS level? E.g., I could imagine something where the browser is informed that www.google.com is loading/proxying a web page on the user's behalf, and through some kind of verification system, it determines the website has given www.google.com permission to do so, and the browser UI updates itself to show the right address in the URL, etc. Maybe there's a better way to do all that, just one idea off the top of my head that seems infinitely better than AMP from a UX perspective.
Though it'd also be nice if you could get websites to make their actual main sites faster for everyone, maybe using a method like jacquesm suggested. That coupled with the low level proxying might be a nice alternative to what we have now.
This is not about technology. And whatever goal the AMP project states it has could have been achieved - and better - in other ways, such as the suggestion elsewhere in this thread to simply penalize page weight. If people wanted a consistent user interface across all websites they would have stuck with Videotext.
> TL;dr AMP is 100% build on web tech.
Yes, so was the search engine that put each result page in an iframe. Only difference was they didn't have a monopoly on search. And it does not make it right.
> Pre-rendering achieves the performance but relies on history.pushState and iframes, which doesn't allow presenting a URL that is not on the same origin as the search page.
Minor technical details, not relevant. If you feel the AMP discussion is going to be swayed by technical bits you're simply out of touch.
> We think that the UX trade off in the URL is OK given the performance benefits.
Who made you the deciders of what the UX of the web should look like? Stick to generating the best search results rather than trying to co-opt the entire web one little bit at the time and leave the UX to the browsers, it would seem you have enough input there already.
> Given that it is a trade off we now have a bunch of projects to mitigate those trade offs.
The only trade-off that will satisfy me is AMP dying off because websites will stop to support it. But as long as Google is strong-arming content providers to use AMP that won't happen. It is no longer fair play as far as Google is concerned. If it ever was.
> My team works directly on WebKit to fix bugs that affects AMP (but also the web as a whole, since AMP is just web tech).
Consider doing something more useful with your talents. For instance, fix the long standing issue with the SERPs that makes it impossible to cut-and-paste URLs pointing to PDFs.
* The post it's replying to admitted motives behind those critical of AMP to the discussion.
* The idea that page weight is a problem best solved by a solution like AMP is hostile enough to the principles the web is based on that it's hard to bar either motive or outright thoughtlessness from the discussion. Defenses on either front are also admissible.
As for the job insult: there are plenty of UI/UX issues that Google could fix tomorrow if they wanted but instead they chose to work on this abomination and on top of that caused a bunch of work for companies well outside the Google ecosystem just to make things work that weren't broken before. That's a very inefficient way to allocate one's resources even at the scale of Google.
Finally, DanG has written a lot about the 'principle of charity', it appears to me that you have done everything that you could in order to take the very least charitable view of my comment and then you attack me on that interpretation.
I care about the web, in a way that I don't care about much else (HN maybe), and any and all attempts to subvert it should - in my opinion, which you are free to ignore - be fought tooth and nail lest we lose what came at a pretty great price. If you're comfortable with walled gardens and with large corporations gobbling up more and more of the open web to put it behind their store fronts then that too is fine with me. But you're not going to shut me up on this subject, nor am I going to moderate my language because you feel I should. Take the tone as a measure of my feelings on the subject.
Yeah, this is really within the bounds of "civil discourse":
> Consider doing something more useful with your talents.
It's not like they're forcing you to click on the AMP results. When Chrome started experimenting with SPDY, one could have asked them "who made you the deciders of what the protocol for HTTP transfer should be?". Yet from the lessons learned from SPDY was born HTTP/2.
If I were only interested in results from Reuters (the top result in the previous search, which was AMP), <<trump on afghanistan reuters>> returns 3 "condensed" AMP results and a non-AMP result above the fold, followed by 8 non-AMP results.
I don't think you're going to get far though as you're obviously just "opinionated" and probably not "really interested" in the "technical discussion".
As long as "We know that most users love the experience" why deign to discuss it with the ones that don't?
I'm really curious to see where using AMP as a proxy speeds up the page, versus applying similar optimizations on origin server?
If it's a cached page (served from any normal CDN), and crafted with similar optimizations  wouldn't it be just as fast, without needing to do the URL redirection?
Even if it were difficult to convince developers to implement the changes, couldn't many of the optimizations could be rolled into a module like PageSpeed ?
 - https://medium.com/@cramforce/why-amp-is-fast-7d2ff1f48597
 - https://developers.google.com/speed/pagespeed/module/
Edit - Reformatted links; I forgot HN doesn't use Markdown.
Getting similar performance out of on-origin serving is a priority for us, but we aren't there yet. One challenge is to give it the same scale as the AMP cache.
Criticism of AMP certainly includes technical aspects, but most objections I encounter are ethical rather than technical.
Did Google explicitly consider ethical considerations when creating and launching AMP? If so, what did this process look like, and where could we find out more?
If not, isn't it time we expected ethical review to be part of significant changes in web infrastructure, the same way we would with significant changes in physical infrastructure?
Now that someone got rid of that it's suddenly an "ethical issue"?!
I cannot opt out of AMP if I use Google as my primary search engine. It's about market power.
The risk critics are concerned about here is a future where the internet no longer "belongs" to the people on it anymore, and despite the malicious nature of the abuses you named, they hardly represented that class of threat.
People here are opinionated about it, but users love it.
So, instead of actually using AMP, you could just follow the AMP guidelines and get most of the speed benefits with almost no downsides. But Google started prioritizing AMP pages in search results, so now there's a downside either way.
You can beat AMP's performance with hand-tuned optimizations. Its goal is to uplift performance across a significant percentage of web content (instead of just that developed by the few experts).
Is Google going to penalize my sites if we don't use it? No, AMP is not a ranking factor.
We don't include advertising in our pages, so should we care about AMP?
Advertising is to AMP like Advertising is to the Web. E-commerce is a big use case where advertising is uncommon.
I understand that AMP is not a ranking factor (except for its inherent speed, perhaps). However, is there not a carousal at the top of SERPs for pages in your AMP Cache?
So while not an organic ranking factor, it would still affect SEO. Is that a fair comment? Or am I missing another perspective?
As far as I see it, Google sees this as valuable to their interests, and have made clear what the talking points are and what should and should not be responded to. That's the only reason I can think of for this counter argument to be continually ignored by him.
In fact all the people here are users. Funny how that works.
I’d prefer hosting all JS locally. (Or at least within of EU jurisdiction, as Privacy Shield is likely going to fall in the courts, too, as Safe Harbor did before, and I don't want to end up liable for that (and I don't want to sell out my users to some foreign tracking company))
So I'm not sure whether AMP is just something controversial as a lot of things from Google tend to be it seems, or if there is some real technical merit as to why it's so bad apparently.
So basically the argument boils down to "because I have shit internet, everyone else should suffer too".
I'm astounded at how much better it is than the last time I tried it.
Switched to it in Chrome a few weeks ago and don't miss Google. Which is shocking to me.
It's been successfully bringing me the correct stackoverflow results I would expect when I need to search.
Plus, I'm really starting to like the bangs, which allow you to search language-specific docs directly without any setup.
For example, if you use a U2F key, you need an extension for that to work - and that extension doesn't work with FF 57. I honestly don't remember the last time something broke with Chrome.
Or smaller things, like the fact that I can't use the keyboard to manage bookmarks (in FF on macOS, backspace doesn't delete bookmarks).
But the worst is in FF, Ctrl+Shift+N isn't new incognito window, it's restore the previous window. Years of muscle memory don't go away overnight.
Firefox is an interesting choice but the Chrome plugin ecosystem seems much more complete for me than Firefox has today.
It'll be interesting to see what happens in the space. I was a Mozilla user, then Firefox and now Chrome user. I have no issue going to better platforms.
I've been trying to use DDG more, unfortunately I'm not so impressed. Quite often end up researching for Google.
Hello. Sorry, but I have to express my deep disgust for what you do. "Give us your content and we'll rank you higher' is the opposite of what made the web and Google awesome.
If you care about user experience, downrank pages based on CPU usage, Memory usage, number of tcp/ip connections, number of domains connected, number of ips connected, number of bytes transferred, percentage of time that the UI is non-responsive and so on.
Trying to force the web into your little AMP cage is a terrible idea. It's the Facebook way of doing things. Are you really so afraid that you have to join the crooks? Would you have managed to take such a big part of the pie from Microsoft if you had been afraid of them and played their game?
Don't fight oppression with oppression. Fight it with freedom. Get rid of that AMP crap.
The only problem I have with AMP is that (google search) users can't turn it off. Me - I love AMP. You would love it too if you had to browse the internet on shitty third-world 3G connections(if you're lucky! Otherwise you're stuck on EDGE or GPRS). Not all CDNs are equal when browsing from less developed countries.
AMP is great for the next billion internet users, not so much for the HN crowd who don't mind multi-MB page loads without worrying about timeouts. Google should make AMP opt-out so that those who can afford to turn it off have the choice.
Purely from the reach it can potentially have if everyone does start using is great. Everybody is already optimizing so that they get crawled by Google, better faster and more efficiently. This is the logical next step.
From a ranking perspective, I think it's important that the HN community stops thinking just in terms of geographical metrics. Yes, in the US, UK etc where our offices are it gets blotched. But our next users are not going to be there from these countries. If at all, being a PM for a global product here, tomorrow's users, paying or free will be from probably from a country where AMP users would be the most impressionable today.
You would love it too if you had to browse
the internet on shitty third-world 3G connections
There's more reason Google wants you to use AMP than "it makes the web faster!". That's just the sell. They REALLY want you to use it, enough that they don't want you turn it off. Maybe something to do with protecting ad revenue?
I heard that AMP was developed by people connected to the search team. And search team is probably another name for "AdWords cash cow team"
It's worth noting that since your comment was made, the idea "AMP use leads to better search rankings" is now directly contradicted by the TL of AMP (cramforce) in a comment this thread .
Also, FWIW, I don't have any issue with AMP as a technology, but I have a big issue with how AMP cache is used by Google.
So AMP at the very least indirectly impacts rankings (and thus impacts rankings), and it may directly impact the ability to appear in the carousel (which impacts rankings).
Would love for him to refute this, but I suspect that won't happen.
Note how over time more and more of the things that Google used to simply point to are served by Google directly. AMP is just another element in that strategy.
We did add the link away from AMP, so that users who cared have a per-link opt out. We know that most users love the experience.
But haven't you seen the level of vitriol that comes up every time AMP is discussed? This isn't a minor thing, a lot of people REALLY dislike it. I have SERIOUSLY considered switching to a different search engine entirely because of this.
The fact that there is now a link to get to the normal page after I've already gotten to the AMP page is not a solution to the issue. As an end-user that feels like I'm complaining that her website is too hard to read because of the noisy background and you're telling me I could just use print preview where the background doesn't show and read it that way.
This feature has significantly change the way I use Google for the worse, and in the time since it was first launched it really hasn't gotten any better for me. It's so bad my OS vendor is starting to incorporate workarounds directly into their code.
If I Google something how do I open five tabs to go look at? I have to ignore every AMP result because I can't open them in a tab. So would Google deems to be the "best" content is now unavailable to me.
I don't have a choice to simply turn it off.
Not everyone will. If you read the comments here there are people who do like it.
But there are clearly a lot of us who HATE it with a passion because it's made your website so much harder to use.
PLEASE give me an option to turn it off.
> But there are clearly a lot of us who HATE it with a passion because it's made your website so much harder to use.
You really seem to overestimate the size of HN readership compared to the rest of the internet.
It is a minor thing. We are a niche audience living in a bubble. We're not - by far - the typical or target users. Our usage behavior is - mostly - irrelevant to Google.
The world has over 3 billion internet users. Let's say that HN has 3M unique monthly users, and that only 10% of those comment, and they are split between liking and hating AMP.
That gives you 150k people who hate AMP compared to 3 billion users on the internet.
Why would you build and maintain a feature for 0.005% of your users?
Even taking at face value the assumptions that yield your numbers (not obvious to me that's safe):
150K is still a pretty large audience.
It's an influential audience, to the extent that tech and therefore thinking about tech is influential.
It's an audience that understands the implications of technology choices like this and can articulate them.
Also, I don't understand the assumption that this audience is completely complementary when it comes to understanding how the rest of the population thinks. I expect it's at least as likely a bet that there are people in the rest of the population who sense something is off or some aspect of the experience is degraded but can't understand and articulate the issue at as high a resolution as it is likely that everyone in the rest of the population simple does not care.
And you really don't have to be some kind of programmer or other ubernerd to see some of the relevant issues in play with regards to URLs and walled gardens.
I know HN loves to think they're influential but I'd suggest than less than two handfuls of HN users actually have measurable influence outside of HN.
No, it's tiny. And considering how HN readers usually user adblock, and rarely click on ads, it is a completely irrelevant audience for Google.
But let's say you make $10 out of each user per year. That's $1.5M in yearly revenue. It barely pays for 3 employees at a large tech firm.
> It's an influential audience, to the extent that tech and therefore thinking about tech is influential.
>It's an audience that understands the implications of technology choices like this and can articulate them.
Considering how even HN readers are divided about AMP, I'd say that the net impact of HN audience on AMP is zero. We're not as influential outside the tech bubble as you think.
What if it was 10% of all users? Or maybe 20%?
Where would you draw the line… what percentage of users would have to dislike amp for you to think it was reasonable to add an option?
I'm not sure what number I would choose (ignoring my bias that I don't like it and therefore my number would be smaller than probably reasonable).
One easy piece of math is (cost of adding and maintaining that feature) vs. [(users who dislike) x (% of users who would quit you) x (profit per user)]
For example, if 10% of all users disliked it (300M users), but only 1% of those would leave your services (3M users) and you made, on average, $10 per user in profit, you'd be putting $30M/year of profit at risk, which is probably enough to hire a few people to build and maintain that feature.
Mostly on HN. And it's clear from this thread that a lot of vitriol comes from deeply ideological positions rather than real technical reasons. That won't be fixed by adding an opt-out.
My assumption is that the idea logical issue is much stronger here on hacker news than among the public at large. Much like views on free software purity.
My guess, for lack of better reasoning, is that the percentage of people on hacker news who hate the experience is roughly the same as those in the "real world". I wouldn't expect that being more technically inclined would make people more likely to like AMP or dislike it (ignoring idea logical issues).
I wrote a comment earlier about my experience with AMP from my perspective as purely a user. Please excuse some of the rantiness of it. Like i said, I can get a little frustrated with AMP at times.
You know that most users love the experience, or you know that most users take the path of least resistance and don't click away using a UI whose purpose isn't particularly clear yet?
How exactly do you know that?
Also, given that the indicator for an AMP page is a greyscale, hard to see icon and words, people who don't want to use AMP might hit it because they are in a hurry and the text they are looking at is black and noticeable.
Which means even less reason to trust you.
Google search basically is a function where you input a search string and it returns a list of URL's matching the search criteria. That's all I need and want it to do. AMP breaks the core behaviour of Google Search, it no longer provides a list of URL's that match my criteria. Basically most search results aren't actual search results anymore.
It's equally possible the majority of people just don't care. That doesn't mean they like AMP or would op-in to it given the choice.
While this is most certainly true, it doesn't mean that data retention and web ownership aren't values worth fighting for.
This whole discussion reminds me of the people ranting against RMS who many years later come to the conclusion that he probably was right all along only by then the horse has bolted.
The long term consequences of AMP are to consolidate Google's power and to make Google even more of an intermediary deciding who gets to see what. It's an end run around the kind of openness that is a cornerstone of the web and I care deeply about that openness and will oppose vocally any attempt to hijack the web for fig-leaf reasons.
Note how Google is slowly boiling the frog by doing all those things that they said they'd 'never do' and how their 'don't be evil' motto gets trotted out time and again to show us how they are good little web citizens when in fact they ram their unilateral solutions down our throats in the interest of their bottom line while pretending it is all about the 'user experience'. Accidentally gaining control over the content delivery of the pages containing advertising tags of competing ad networks sure must look like a nice and juicy bonus.
Borders and Toys'R'Us both decided to partner with Amazon for a while. I'm not sure why they thought that was a good idea, but that was their choice.
What I care about is the user experience. That's where I'd be happy with an option to simply disable it on my devices.
I think it's kind of unfortunate that it's not possible for people to discuss AMP without both issues coming up and confusing things. But that's life.
But forget all that, on a simpler level, if you do nothing most things stay the same. If you're happy with the status-quo, you're welcome to be silent. Others prefer vocal protest.
If the majority of users merely tolerate AMP... I wouldn't exactly call that a win for Google.
A Google engineer saying their users love their products is probably expected. They were probably fed this by some marketing/sales droid.
How do you achieve this with a non-AMP website? It sounds great.
Congratulations, you re-invented WAP and turned back the clock on the mobile web 20 years.
In terms of the AMP open source project governance rules I am THE tech lead.
I also manage a team at Google that works full time on the AMP open source projects and act as an overarching tech lead, but there are several sub teams that have specialized tech leads (e.g. for UI, ads, analytics, caching, etc.).
Is that what you wanted to know? :)
This is identical to the ranking boost given to Google Shopping (for which Google was just fined 3 billion EUR), so you can expect the same result.
Isn't it different because (presumably) the AMP links aren't owned by Google?
The fine was specifically for anti-competitive behavior, i.e. consistently surfacing their own links over those of shopping competitors, which at least with AMP links isn't necessarily the behavior, though there is likely some overlap there.
Google only gives the carousel position benefit to sites using the AMP js from Google's CDN.
If you use any other CDN, you don't get the benefit.
From a legal standpoint, this behavior is clearly different from their last slap on the wrist.
Google doesn't want the web to be fast or lean, it incentivizes you to create a fast version to be shown from google while the slow version is shown everywhere else.
Google: fast, everyone else: slow.
And amp positioning and ui display is absolutely a ranking boost.
No, everywhere else it's the normal speed your servers can deliver it to clients. Nothing about AMP is slowing down your normal web page.
The positioning and display are a ranking boost for whoever chooses to implement AMP pages - they are not explicitly targeted at Google services, which is the behavior the EU fined them for.
Anyone who thinks the EU is going to fine Google for AMP is actively avoiding reality at this point.
Faster and slower are relative terms.
> Nothing about AMP is slowing down your normal web page
Nothing about AMP is making the web faster, which it says it's doing on the project homepage. It's making a fork of the web to be shown on Google that is 'faster'.
> The AMP Project is an open-source initiative aiming to make the web better for all. The project enables the creation of websites and ads that are consistently fast, beautiful and high-performing across devices and distribution platforms.
I understand how it works and Your arguments so far aren't changing my mind.
Thanks for posting the homepage text, for posterity.
I believe there is research showing that little adornments like that convey trust and increase the CTR vs other listings that don't have them.
The easiest way to do the "more generic solution" is to remove that turd burger from search results. Interwebz fixed.
So, like regular webpage visits?
Now Google has forever(?) broken links in search results, in order to count click-throughs - so I suppose all browsers should special case that too...
I can't for the life of me figure out why they married an ok html/css/js subset (amp pages) with an absurdly invasive proxy technology.
I hear they say "speed", and "75%" - but if that's going from 130ms to 50ms complete page render does it really matter?
"we fucked up again web to wap level and now everyone needs to jump along because sharing is royally screwed"
I really wish some interdimensional telepathic Fart would Goodbye Moonmen this TL into abandoning AMP.