By "their CDN" I meant Google, Cloudflare and Microsoft. Can we set up our own CDN to serve our own content from our own servers and receive the AMP badge in search results?

Please disclose your affiliation to Google either in your bio or in comments, and don't post the same comment in multiple places.

> Can we set up our own CDN to serve our own content from our own servers and receive the AMP badge in search results

This...doesn't make sense. You lose the value of a CDN (both to you and to the consumer of your content, in this case Google and the end user) if you're rolling your own.

It no longer makes sense to be able serve our own content without it being pushed down in search results?

We were talking about CDNs because your collegue mentioned AMP CDNs, but the main point doesn't change: we cannot serve our own content from our own servers and get the same placement in search results as AMP content, even if our content loads verifiably fast and is as performant as an AMP page on the client.

> We were talking about CDNs because your collegue mentioned AMP CDNs

I have no clue who bdeurst is. They certainly aren't a colleague of mine.

> even if our content loads verifiably fast and is as performant as an AMP page on the client.

Can you explain to me how your page load time is 0ms? My understanding is that a correctly functioning AMP-cached page will load for the user in a whopping 0ms, because it can be preloaded.

The entire design of AMP starts from a fairly straightforward premise: "How do we reduce (user-visible) page load times to 0, safely, cross origin?" If your pages user-visible loading time is longer than 0, you're failing to keep up with AMP.

AMP pages are preloaded, that's how your get the 0ms load time. If Google would instruct the browser to preload other search results the same way, those would also be available in 0ms when the user accesses them.

See my statement about secure, cross origin preloads. You're asking a search engine to XSS attack its own users.

I think the correct term I was looking for is prefetching. That's a secure way to tell the browser to start loading search result links in the background.

> That's a secure way to tell the browser to start loading search result links in the background

prefetching isn't private cross origin:

> Along with the referral and URL-following implications already mentioned above, prefetching will generally cause the cookies of the prefetched site to be accessed.[1]

IDK about you, but I'd generally prefer that my cookies and IP not be exposed to all of the links that happen to be in the first page of search results.

[1]: https://developer.mozilla.org/en-US/docs/Web/HTTP/Link_prefe...

Browser specs can be improved, and new ones introduced. And even if safe prefetching is deemed technically impossible, the question remains: should we give Google and a handful of other companies disproportionate control over how we publish and consume content, for 50ms of load time?

> Browser specs can be improved, and new ones introduced.

Yes, for example Signed Exchanges, which on a technical level solves all of the problems of rel=prefetch (and a number of the problems with AMP, like link pasting and copying).

> Should we allow a handful of companies to be pinged every time we load a page on the web

I'm hopelessly confused here: you're only going to "ping" one of the handful of companies if you were referred by that company. (In a world with signed exchanges) You're not going to come across an AMP-cache link organically. You'll navigate to example.com directly, without anyone except example.com (and your DNS provider) knowing. The cache provider will only know if you navigate to the cached site via the cache provider. Concretely, you don't go to the Google amp-cache unless you're navigating there directly from Google's search results. Same for Microsoft/Bing.

So if your metric is

> Should we allow a handful of companies to be pinged every time we load a page on the web

Then yes, absolutely, because nothing changes!

Edit: To address your other question,

> should we give Google and a handful of other companies disproportionate control over how we publish and consume content, for 50ms of load time?

Alright: how is AMP materially different from <whatever other algorithmic choices rated search results before>?

You seem to be claiming that AMP is harmful to someone but who? It's not harmful to competitors or to end users, and its only harmful to developers if you make the most strained argument.

My premises here are that users actually prefer AMP results. You may not, but my understanding is that most users do. So from the perspective of an end user browsing the internet, AMP leads to an improved experience.

So it's good for users.

No one has yet been able to explain to me how its actually harmful to a web developer who now has an incentive to make AMP-compatible sites. Like sure, you now have to work with a framework you may not like, but that's not a compelling argument when people are claiming that AMP is a threat to the sanctity of the internet.

So it's not like bad for web developers, it's just sort of a lateral move.

That leaves competitors to the giants. But AMP is an open standard, and DDG could, if they wished, implement an AMP cache themselves today and it would just work. And they'd, if anything, benefit from the bigger players pushing that ecosystem. There's the potential for abuse via the caches.json registry, but the AMP project is aware of this and notes that the registry could be decentralized using Subresource Integrity or similar, if such a standard was adopted[1].

So again: I'm confused by how exactly it's bad, beyond the "I am forced to develop in a way I don't want to if I want to appear near the top of the search page", which isn't new.

[1]: https://github.com/ampproject/amphtml/pull/18495/commits/c5d...

I've actually removed that second question before I noticed your answer, because I knew you would then skip over the first one. Feel free to address the main point I was making in all my posts in this thread, whenever you feel ready.

> So if your metric is

Google being pinged obviously isn't my main metric, as you can see from all my posts in our discussion. My main concern is that publishers will be forced to use specific publishing mechanisms (AMP, Signed Exchanges) to appear at the top of Google Search results. That loss of control puts publishers in a vulnerable position, and hurts innovation across the web.

> Then yes, absolutely, because nothing changes!

Everything changes. Google's influence and control won't end at the moment the user navigates away to a top result on Google Search.

I updated the prior post, but to be a bit more concrete, what differentiates AMP signed exchanges from, say, HTTPS?

Google could serve an older version without the user knowing.

The signed exchanges protocol requires that the content be signed with a key with a short expiry date (< 1 week), sites are free to make it shorter. In extreme cases, the providing site could sign with an expiry of < 1 day or even something like 1 hour.

And iiuc, sites are still free to revoke their certs. So this is actually probably more secure compared to something like https in that regard.

[0]: https://github.com/WICG/webpackage/blob/6cc3237b36c2f9ce7534...

I see what you're saying but im not sure that's a problem - how many people are building CDNs in their garages that need to be certified?

Also, everything I say on HN reflects my own opinion and not any organization, which is what my profile states. I do not hide behind an anonymous username precisely for this reason. Poisoning the well by doxxing me doesn't change how the AMP standard works for CDNs either, and only serves to derail the conversation.

>Everything I say on HN reflects my own opinion and not any organizations, which is what my profile states

Sorry, but I agree with dessant. Your profile doesn't disclose your affiliation and nor do your comments. It's absolutely relevant to the discussion, because whether you want to admit it or not (or try your best to act neutral), your day job will have some influence over your opinions on these sort of projects.

You're right that it doesn't change objective facts about the specification, but I think it's misleading to suggest that, in general, external third-party CDNs are first class citizens in the AMP ecosystem when they're not treated the same within search.