[dividuum]

Personally I don't think the idea behind AMP is bad. But the implementation is dangerous as it artificially fragments the web. I guess fewer would oppose AMP if google made some machine verifiable guidelines for "light" webpages that would earn them this "icon". Linked from the "fat" page by some "link" meta tag.

[godot]

>if google made some machine verifiable guidelines for "light" webpages that would earn them this "icon". Linked from the "fat" page by some "link" meta tag.

That is exactly what AMP is; with the exception (a huge one, I know) that Google also then caches the page on their server and serves it from a google.com host.

[avaer]

For Google, the hijack is the prime feature; AMP is the PR vehicle that makes it swallow.

AMP could have been done without the huge exception, but then Google couldn't profit from it.

[claudiulodro]

The fast, non-blocking content loading is the main feature. I took a radical path when developing my site and made it AMP-first. Instead of having an AMP version of each page, every page is its own AMP version because it's an AMP page.

Even served from my cheapo shared web host and not Google's AMP cache, I have pretty-much instant loading of all pages: http://multithreaded.link/2017/08/lyft-customer-acquisition-...

It's a good framework for building super fast pages. I will admit it's a little riskier to build a site on top of technology a large company owns, but this is a risk that you also have when you use React or other frameworks.

[smt88]

> this is a risk that you also have when you use React or other frameworks

React isn't a good analogy. When you use React, you can get 100% of the value of the library, even if Facebook disappears overnight.

By contrast, much of the benefit of AMP is the caching aspect, which relies on Google.

[stephenr]

> The fast, non-blocking content loading is the main feature.

You don't need requires-js-to-render markup or a cache operated by a privacy whoring ad company to make a page load quickly.

Edit: also, comparing to react or whatever js framework is like saying "look at my new cast iron shoes, they're so much lighter than those old lead ones".

[kmoe]

Your "About", "Contact", and "Privacy Policy" footer links don't work (they go to "#").

[claudiulodro]

Yeah sorry about that! I just launched it recently so I'm still getting everything going. :)

[an27]

Nice, on FF I get a white page because Google domains lurk in my hosts file. I suppose I'm not your target audience.

[leeoniya]

> For Google, the hijack is the prime feature; AMP is the PR vehicle that makes it swallow.

this is 100% true. yet somehow there are still those who swear this isnt the case. i've ranted about it previously [1].

[1] https://news.ycombinator.com/item?id=14695268

[cobookman]

I think it's also about prefetching amp pages from Google search. You can't do that without the cdn.

Sure google could modify chrome to allow this and not block it. but you'd be asking for even more technical blockers. And you'd have something that only works for a small portion of the web.

[avaer]

Sounds like the perfect use case for a nice standard CORSy web prefetch API that all browsers could implement.

And this would work for at least as much of the web as is currently putting effort into AMP.

[cramforce]

We are thinking about this. The problem is privacy. AMP can pre-render a page from publisher X without publisher X learning that this happened.

This is important for avoiding that you e.g. would suddenly see re-marketing ads for pages that you have never clicked.

[baddox]

Regardless of profit, if Google just made the AMP spec and didn't incentivize publishers to use it, well, no one would use it.

[avaer]

Not saying that's true, but if it is, wouldn't that mean AMP is not a desired product?

I fully support a "faster page spec" that search engines incentivize, just like I support a ranking boost for using HTTPS.

But AMP isn't that; AMP is incentivizing the payment of tribute data to Google under the veneer of a faster page spec, to steer the rhetoric more favorably for the company.

[nine_k]

It's desired by the (non-paying) users, but not necessarily by web publishers. Here the desires of said users and Google align: they both want to see a search result quickly.

[adrianratnapala]

How much of this is actual desire by web publishers, and how much is just collective stupidity. Do they really want their websites to be a hassle to use? To have readers turn away because it doesn't load faster enough, or at all?

I think it is more likely that they just higher web-developers, according to some cultural norm about how to select devs and about what features to specify. And those devs choose frameworks according to other cultural norms about what makes a good web framework. They are partly, but not completely constrained in this by the requested features set.

The result is bloated JS applications where mere pages are required, but the complexity of the system, and the social inertia is so great that nobody can fix, or even perceive the problem.

[baddox]

> wouldn't that mean AMP is not a desired product?

Well, kind of, in the same way that common-sense lightweight HTML/CSS/JS pages aren't a desired product for the people who make those decisions on publisher websites.

[brianwawok]

I want it as a consumer.

If AMP went away tomorrow I would just browsing on my phone. Ad blockers are not good enough, and I refuse to fight popups and scroll blockers and blah blah blah in 2017.

[Tom4hawk]

Maybe it's time to change websites you visit? If website is actively trying to prevent me from reading it I just don't do that.

If you are a publisher and think that you need AMP than it only means that your website sucks big time.

[notatoad]

>For Google, the hijack is the prime feature

what does google get out of it? The experience is better for users who like fast, lightweight pages, but as far as i can tell there's no tangible direct benefit to google when you load an AMP page. The only benefit to google is if their users like the experience and continue to use Google search so they can access AMP-cached pages. If users don't like AMP, google gets no benefit.

or am i missing something?

[avaer]

Google's entire business model is tracking users as they use the internet in order to target ads.

Every AMPed page is contributing work towards improving the effectiveness and appeal of Google's product (selling ads), by adding data to the targeting algorithms.

If every piece of content on the internet is tracked -- through full UX hijack on Google's infrastructure -- that might be a trillion dollar holy grail.

[tyingq]

The ux hijack will progress over time as well. Google knows how to boil users slowly so they never notice the heat.

AMP carousel pages have some subtle capture features that regular AMP pages don't. They will add more, then slowly move them to the regular AMP pages over time.

[notatoad]

you think google doesn't already track the links you click on the SERP? they aren't learning anything new by putting a tracking code on the page that link takes you to.

[stephenr]

People don't share google result links.

People absolutely share amp links.

Also, amp being google.com hosted means no more pesky same-origin privacy protections like limiting cookies.

[fastball]

I don't understand why this matters though?

Google is tracking you. If you don't want to be data-mined, you shouldn't be using Google. If you are using Google, it means you don't mind being tracked. It means you are placing the convenience that Google provides above your concerns about what a faceless corporation can do with your data. If you don't mind being tracked, AMP is not a problem.

I understand how you could dislike Google. I don't understand how you could dislike AMP specifically, but not necessarily google.

[stephenr]

seriously? If you want to use any of a companies services, you have to agree to like any and all of their future changes to said service, no matter how monopoly-like they are or how many privacy protections they hack around?

[kuschku]

I want to prevent my users from being datamined, even if they found my site via Google.

But I also don’t want to end up basically deranked on Google by sites that have nothing to do with the result.

Just look at this example to see how AMP distorts the actual search ranking, by moving pages to #1 that have nothing to do with the topic, but happen to be on AMP: http://i.imgur.com/84FvZmA.png

[shimon_e]

Fact is, Google would already have a complete profile on you for advertising purposes even without AMP. They offer more than enough services that track you that they don't really need AMP for tracking purposes.

[notyourwork]

How does google profit from serving these pages opposed to original content creator?

[the8472]

> That is exactly what AMP is

Except that there is mandatory 3rd-party javascript that needs to be loaded just to get it to render. That doesn't make sense if it really just were about lean pages

[ninkendo]

That's an extremely disingenuous statement. It's not just "a [huge] exception", it's the entire basis for google's strategy. AMP isn't just a standard for how to construct a website, it's a way for external entities like Google to host your site directly inside their site, so that they have ever-increasing access to your browsing behavior. Now Google doesn't just know what results you clicked on, but what links you're following inside those results, how long you stay on the pages, etc.

Facebook, twitter, reddit, etc already utilize similar techniques on their mobile apps to make it so "web links" you click just keep you in the app.

But the web (as in, not bespoke mobile apps, but what you get in a general-purpose web browser) is supposed to be decentralized and resistant to any one entity owning the end-to-end experience.

I'm aware that any other entity (like Bing for instance) could do the same thing that Google does, and that the AMP standard isn't what's at fault here, but there should be no doubt as to Google's motives.

[grey-area]

So, that's not what AMP is. There's a reason google don't drop the incredibly unpopular hijacking of content, and it's not to do with edge caching, but control.

[kuschku]

And modifies the pages UI.

And forces you to load content from Google’s CDN.

And so on.

[stephenr]

You don't need new js-required-to-render markup to make a page lightweight

[rhizome]

With Google I can never relax to think that something new will be a good thing for me, a user, but their "make mobile viewable" (or whatever the message is) reformatter for sites that still have 10 year old CSS could certainly result in some sort of signal displayed in search results. A mobile-viewability indicator, perhaps. I dunno, it sounds kinda hokey when I write it out like that, but I have to think they're going to productize it at some point.

[stephenr]

It's pretty easy to detect the number of external scripts, media objects, fonts and images a page attempts to load.

But that wouldn't give google control and importantly, a sure-fire run around same-origin privacy protections.

[mikeytown2]

>if google made some machine verifiable guidelines for "light" webpages that would earn them this "icon".

If you get a score above 95 or (X) on google page speed insights that might be a metric that everyone can agree on.

[BoorishBears]

I can't for the life of me understand why the icon couldn't just be a link to the AMP version (or be located near the current cached link is), and the link would link to the normal page