Hacker News Reader

I'm curious about this as well. How easy would it be to game performance measurements? For example, when you can lazily load content it becomes much more difficult to measure. The advantage of the AMP restrictions for them is that they know for sure certain performance metrics will be met but this is much more difficult for arbitrary webpages.

View on HN

Run a WebPageTest on an AMP link and tell me it fits within the 1 second promise AMP keeps making.

Without prefetch, on a lot of the pages I've cross-checked manually, it doesn't. So there's already gaming from the other side of the aisle in a sense.

That's a good point. I think google is able to measure this to some extent, but yeah I guess anything could be gamed.

Reply on news.ycombinator.com