Hacker News Reader

Comment by snek | Hacker News Reader

You're asking the wrong question. AMP isn't bad, Google ranking AMP pages over equally-fast non-AMP pages is bad. But Google's main goal here is to preload/render content without the content knowing it is being preloaded, so that the page appears to load instantly when the user clicks on it.

What I think actually needs to happen is a standard for deferred navigation where the UA can be told to load a bunch of resources and then choose one of them to actually navigate to (basically what AMP does). The problem here is that Google is (as we type here) actively coming up with horrible standards like signed exchanges so that they never have to send users away from their own domain, so I don't think they would be fans of a standardized system that killed AMP.

View on HN

AMP is also bad, because the technical steering committee is 3/7ths Google employees[1], with the rest made up of similar platforms that use others' content, like Twitter and Pinterest.

So content-producing organisations are increasingly strong-armed into building websites the way the platforms want.

Also, having large tracts of the web built using the same small set of severely limited components means we get dull, samey websites susceptible to the same hacks or bugs where things like interactive features are much more difficult or impossible.

I wrote about this and more here: https://unlikekinds.com/article/google-amp-page-speed

[1] https://github.com/ampproject/meta-tsc/blob/master/README.md

I think a better solution would be for Google to heavily demote pages with slow load times, popup banners, notifications, etc... It would incentivize authors to deliver clean/fast pages while still leaving them freedom to design pages any way they want.

> AMP isn't bad,

No, it is clearly bad. You seem to even agree. The stated reason for creating AMP may be good, but the technology and standard itself is bad.

Like all things this is not as black and white as you'd think. AMP as a web-components toolkit is actually fairly nice and I totally get the flip side of why people think it's dangerous.

> AMP as a web-components toolkit is actually fairly nice

It is? What do you like about it? What does it help you do more easily?

AMP is in no way necessary to build "user-first" fast loading pages.

AMP was created to allow aggregators to ensure they could cache and serve content without the user leaving the portal.

As publishers are desperate for the traffic, they have adopted AMP to make the aggregators happy.

The user is never given a choice, they just end up stuck with a cached AMP page and confused about how to reach the publisher's real website.

This "user-first" framework is really more "user-last"

Reply on news.ycombinator.com