Great points - also, reading between the lines of his tweets, it sounds like the bug was pushed by a team he doesn't lead, so the best he can do is ask them "Please fix this, people are really pissed (at my team) because of this". But to the Image Search team, it's like, why should they do a costly rollback? It's easier for them to just sit back and say nothing, and let AMP soak up the blame on HN. Then they can fix the bug on their normal schedule, without their deadlines being affected.
This most recent change is just a bug in Image Search, at least based on the tweets I read. The extension seems to inspect the AMP article's URL and HTML, both of which are outside the scope of Image Search: https://github.com/da2x/amp2html
Seriously, I just don't get these hateful comments on Hacker News regarding AMP. It helps websites load fast on mobile, and users really appreciate when websites load fast.
What about the fact that it was originally called Page Speed Project and was a genuinely open standard; that that was subsequently shut down in favor of AMP? That the difference between the two is that Page Speed did not have attractive collusive lock-in effects while AMP does have collusive lock-in effects?
Did it work? Page Speed existed for a long time, and over that time, a huge portion of websites became bloated as hell and ridiculously slow. Hackernews is just about the only website I can think of that didn't.
"This webpage works best in Our Browser™️" was just a pointer to the user for a faster experience. It came at the small cost of crippling the web for a mere 15 years, probably costing a few billion hours user time wasted at horrible UX bred in a hostile corporate controlled environment.
AMP is just the same all over again, just with a singular player which might just be too big to fail. Oh, and also this time things like the information source of the western world is at stake, not only fan sites of specific locomotive models by some geeks.
But at least my websites load....slower actually if the page was properly built[1]!
> "...just with a singular player which might just be too big to fail"
while i agree with the overall sentiment that google is pulling a microsoft, i don't see how this line is true. if google went away tomorrow, how would it lead to an economic downturn?
it would suck if you're all in on google in some way, but most of their products have competitors and substitutes, so you could switch. maybe android losing some core services that aren't immediately replaceable might cripple phones for some folks. the engineers could certainly find new jobs, and a bunch would go on to create new companies from the things they're working on now while soaking up all the idle wealth around the world looking for any return at all.
i rarely use google for anything, so maybe i'm missing the obvious.
I hate that every fucking google search result on mobile web has its stupid little icon
I hate that there is no way for me to disable it as a user
I hate that it has muddied the waters in what the url bar means
I hate that it has trained users to not question fake url bars.
I hate that cloudflare so thoroughly jumped on its dick
I hate that we invented a way to fake the address in the url bar just for this stupid fucking feature.
I hate that we now have a system where somebody can share a page url with a friend, and that friend can view it on t he same device model using the same browser with the same settings, and will get a different page because one was viewing an amp page but shared it's real url.
I hate that every fucking amp page is lower featured in some way, and almost never works in desktop mode.
And most of all, I hate that it leads to everybody offloading shit onto google's servers.
AMP is not fast because it's served from google's CDN. That's a lie, It has always been a lie, and it will always be a lie. AMP is fast because it's incompatible with 99% of the bullshit client cpu heavy tracking and ad libraries, so they don't get included inside AMP pages.
We could have just had that, without all this stupid bullshit CDN redirection/misdirection bullshit.
And you know what, You want me to get off my hate train? Get google and all of the other search providers using it to solve complaint #2. That's really all it would take to get most of the hate to go away.
> AMP is not fast because it's served from google's CDN. That's a lie, It has always been a lie, and it will always be a lie. AMP is fast because it's incompatible with 99% of the bullshit client cpu heavy tracking and ad libraries, so they don't get included inside AMP pages.
AMP is fast because of prefetching. That's not a "lie." You can measure this. When you click an AMP link, the page is already downloaded. Prefetched pages load faster than https://motherfuckingwebsite.com/
But prefetching has a problem: when you visit a website, they can fingerprint you and track you. Prefetching from random websites in search results violates your privacy.
The only way to fix the privacy issues with prefetching is for the search engine (not just Google--Bing, too) to serve the prefetched page from their own CDN.
But wouldn’t that mean that your search results are going to contain much more data than necessary? Sending you’re results and prefetched versions of all the websites linked therein?
now that is pretty dangerous if i may say so... what happens when a search result for something totally innocuous turns up shady webpages that you never click on but get pre-loaded and cached by your browser?
"The only way to fix the privacy issues with prefetching is for the search engine (not just Google--Bing, too) to serve the prefetched page from their own CDN."
Except Google is probably the biggest threat to privacy on the planet. You're not "fixing" the privacy issues, you're making things worse.
It’s not making things any worse. Google already knows what your searched for and they already know what you clicked on, even without AMP. They probably know what you did on that webpage too because they’re probably using google analytics.
I don’t like google google either, especially because of their history on privacy issues, but they’re not getting much additional info from this. Just a stronger foothold on the web
> I hate that there is no way for me to disable it as a user
Yeah, seems like you should be able to. Just curious, where would you want that option to appear, and how would you want to see the search experience change as a result?
Even regardless of me being pro-AMP or anti-AMP, I'm not sure what you're trying to get at here.
- Google has settings for search and a page for these settings. For example, I have sometime in the past went to that page and set my preferences there. And now they're taken into account when I use Google's search. Like settings are wont to do, y'know.
I think you're approaching it from the wrong POV. First of all, the option needs to be on the site owner/developer side on whether to adopt AMP or not - and not have the decision of not adopting AMP adversely affect search ranking.
As a user, I should just see whatever the site lets me see. If they decide AMP is best for user experience, I would no doubt get it. If they decide to show a responsive mobile friendly site because it works best, then show me that.
My first guess is to only make it available in the first place to people using search while signed in. Could even sell it as a way to get the “faster” web by signing in. At least that way there’s a choice.
Making it default for everyone and then forcing people to sign in to opt out would be counterproductive from a privacy standpoint.
Could also just provide the original link right next to the AMP link in search results and let users decide.
Cloudflare seems to have provided a service that fixes a lot of the issues you mentioned, so I'm not sure why you're mad at Cloudflare for making the best of this bad situation.
> AMP is fast because it's incompatible with 99% of the bullshit client cpu heavy tracking and ad libraries, so they don't get included inside AMP pages.
> We could have just had that, without all this stupid bullshit CDN redirection/misdirection bullshit.
I agree the world could have faster pages without AMP. What would that take though? Would it take web developers across the world pushing back against including megabytes of extra JavaScript for ads/tracking, at all the companies who are currently using AMP?
Isn't that up to the developers themselves - whether they want to shoot themselves in the foot by slowing down their site by including lots of JS? Who is Google to determine how the Internet should be run? Shoving AMP down people's throat - because Google can as it owns majority browser and search market share - is akin to large governments in the world pushing their weight around and sticking their noses in other countries' issues.
I cringe everytime I visit a page that's AMP enabled, and usually bounce or just get on a desktop. Sure, promote a 'faster' web, but if it's at the expense of a horrible experience, why?
No, it's akin to large governments swooping in and regulating industries when they start hurting people en masse. Web developers weren't professional enough to make the web fast. AMP isn't the right solution, but the community itself was incapable of policing itself so it was just a question of time before it happened.
Before that, people were massively moving to the apps because the web got too slow and unusable on mobile.
Last time I checked, Google is part of the free market, no? This is exactly how free markets work. And if it turns out AMP is really bad for the web, then the reverse will happen too at some point.
> Web developers weren't professional enough to make the web fast.
Don't shit on (other?) developers to make your point. It's revisionism that only serves the purpose of taking more control from users and developers, which just conveniently gives even more power to the stakeholders who most benefit from the structurally unsound game businesses have been playing.
By the same logic, I could say the Google Chrome team (not related to AMP — just a metaphor) isn't professional enough to make a browser that uses less RAM. In reality, the team deals with a myriad of constraints and other priorities, both technical and commercial, that sometimes clash. Rendering speed and security are its major selling points; if both are met, management is satisfied and Google makes billions.
I'm loath to say this, but performance isn't where a lot of websites make money. It's the ads that people click, way too often by mistake. To feed this system, a bootload of tracking scripts, A/B testing scripts, overeager live support widgets...
Google is coyly selling preferred placement in exchange for performance, which may be the only way to get some businesses to care about performance. Developers are cranking out what the business owners are asking for.
Ask yourself why Google did not take aim at the ownership side, why it did not arrange discussions with major media companies etc. to establish performance as a company-wide goal, instead, put the responsibility on developers.
> Before that, people were massively moving to the apps because the web got too slow and unusable on mobile.
Yeah, the thought process to go from seeing the "use our mobile app instead! It's so much better!" banner, closing it and sending my eyes onto the content that I wanted to read takes at least half a second.
> No, it's akin to large governments swooping in and regulating industries when they start hurting people en masse
Okay, I can see that. The idea of Google being the gatekeeper and regulator of the Internet is scary though, especially given it's a for-profit corporation that has its own interests at hand. I think AMP would have been better if it started off as a proposal for it being a standard to HTML, voted on and governed by the W3C, instead of by Google.
> Web developers weren't professional enough to make the web fast
I would argue though, that web developers who are professional enough to adopt AMP, would be professional enough to ensure their site doesn't load bloat.
It is actually the opposite, it is a large corporation abusing its position to entrench itself even more. It has nothing to do with regulation. If google wanted to do the right thing all they had to do was penalize the bloat.
Except we've seen that it's not up to the developers, but some product manager or marketer who is willing to sacrifice the mobile experience for email popups, autoplaying ad videos with sound, loading 50MB of unoptimized content, and other value-sucking tragedy of the commons behavior.
Why is this Google’s problem to solve? And why, if it is somehow Google’s problem to solve (which would mean it’s any search provider’s problem to solve) does it require the site to adopt Google technology? Why not just penalize slow/bloated sites?
Considering the severe downvotes for your question: Nothing. It didn't make a difference when Google announced penalties for slow loading sites and nothing would make a difference now.
Hence why web developers of HN are so angry at AMP - it bans them from building sites filled with 20MB of JavaScript bloat. If you read through HN history, you'll see hundreds of comments belittling anytone that complained about webpage size and slow loading - usually defending it with a "noone has time to optimize things and build without these huge libraries" argument.
>If you read through HN history, you'll see hundreds of comments belittling anytone that complained about webpage size and slow loading - usually defending it with a "noone has time to optimize things and build without these huge libraries" argument.
Any examples? My limited reading experience here has been the opposite.
Nearly all such generalizations about HN are quite wrong; this one seems particularly off.
I think it's because comments we don't like make a much stronger expression than comments we do, so they weigh more heavily in forming our impression. This also explains why people arrive at such contradictory generalizations: it's because they like different things.
Technically speaking it does nothing of the sort. It helps websites be absorbed by Google so that Google can retain the user on their site. Which keeps the experience fast. But you don’t need Google to make your site fast, and in some cases using AMP can slow your site down. If it was just about making the web fast, there wouldn’t be AMP-only features. It would just be “fast only” and websites would have to meet certain criteria to pass that test. You should try using your brain a little more when you think about the motivation behind Google’s actions. Do you think the stakeholders at Google care how fast a user can exit Google?
Are you trying to point this out as being a good thing? I don’t know the answer because I don’t click AMP links, but I don’t have a problem with GDPR nonsense anyway because I’m not in Europe.
you still have to comply with the GDPR. if not, you can be detained if (e.g.) your plane makes an unwanted stop in the EU. (this assumes, EU brought a GDPR complaint against you or your company and you bothered to not respond)
Personal subjective experience does not show that AMP pages load faster. It has been the opposite, where Google servers have failed to deliver the page.
But mostly I hate that URL crap. Give me my URL bar functionality back with original host URL. And that there is no way to disable AMP from results.
I share the same distrust about Facebook and Google on HN, but only the latter seems to bring substantial downvotes.
To be fair, they have more employees, and I don't assume bad faith in the downvotes — I've been wrong more than a few times. Still, I sense that Google elicits a gratitude that is unjustified when it comes to relationships between companies, customers and employees.
If they cared about performance they’d simply weight it more in search rankings. I find AMP pages have worse user experience because they require 100KB of JavaScript to load and complete running before you see anything, which breaks with a couple decades of best-practice, including what other Google teams advise. I regularly see AMP load slowly or break because they didn’t follow those good practices whereas sites with good engineering practice load more reliably.
It also breaks URLs – instead of the real URL, you get a google.com link which doesn’t have a good desktop experience.
"Users appreciate" and "is it good for the world" are different metrics. I'm sure many users appreciate drive-thru daiquiri stores, and they get plenty of business, but it's perfectly understandable why people would dislike their existence.
HN's hatred towards AMP is something I've never understood. As a user, I'd much rather visit an AMP website than have to deal with the atrocities devs create when left to their own devices.
In general, HN has a pretty strong bias against Google and Facebook. I suspect it may be clouding the average user's judgement. AMP has done immense good for the majority of users of the mobile web.
Google has an internal system which allows engineers to transfer to other teams. If what you described happened to me, I would secure a transfer and then report the situation to higher ups.
IIRC that works only if you never got a bad performance rating, then nobody would touch you internally and your chance to transfer within Google is lower than to move to another company to a better level.
If you are a $race $gender, Google will hire you. The diverse candidate pool just means you need to work harder on your resume, interviewing skills, etc - since there's more competition.
Exactly, two engineers could have an entirely different experience based on their team. Engineers can transfer to different teams too, though that team will obviously look at the engineer's history of performance.
Two engineers on the same team can even have vastly different experience. Google is big enough that it attracts a huge spectrum of engineers and one person's hellish workplace is another person's bliss.
20% projects have been cut? I wish someone told me that before I started a 20% project which spanned multiple quarters, and still did well in my last performance review.
If an engineer isn't doing well on their 80% project, a 20% project might be frowned upon, that is true.
I guess the policy has changed. My impression was in the early days you could work 20% on your own stuff (no/few approvals necessary) and 80% on your corporate assignments. That's changed into something that, frankly, isn't very unique:
> In 2012 the firm began requiring engineers who wished to work on individual projects to run their proposals by their managers first. This was a significant change from the firm’s previous policy.
> In 2013 it was reported that managers had clamped down on staff taking ’20% time’ so as to avoid their teams falling behind in Google’s internal productivity rankings. Managers are judged on the productivity of their teams—Google has a highly developed internal analytics team that constantly measures all employees’ productivity—and so time spent on ’20% time’ projects would impact this.
2013 Google sounds like pretty much every other major company, in this area.
I think you can still do a 20% project. This was about the performance systems, and not about workload. Workload is easy peasy. Its almost a sinecure for most people because all you need to do is to optimize for promo.
So that refrain seems to imply the 20% project must be done as additional work after spending 40 hours a week on the 100% project. My point was you actually can spend roughly 32 hours a week on your main project, if you are performing well, and then roughly 8 hours a week on your 20% project, and still get high performance ratings.
