5 ms·
Took a performance benchmark in both Firefox and Brave. Both of them are unusually slow (addons/extensions disabled, fresh profile), but Firefox especially so.
by jeroenhd 2y ago
Took a performance benchmark in both Firefox and Brave. Both of them are unusually slow (addons/extensions disabled, fresh profile), but Firefox especially so. Source for the huge render times seems to be desktop_polymer.js, specifically the part that's registering and setting up custom web components. Once the website loads, performance becomes a lot better.
My guess is that a Polymer update made Youtube slower for everyone, but SpiderMonkey isn't particularly great at the kind of excess operations that have been added. Firefox in particular seems to suffer from complete UI freezes whereas Chromium browsers seem to just have slow tabs when the browser is overwhelmed.
While I certainly wouldn't be surprised if this is part of an anti-adblocker mechanism, not every slowdown on Google's websites is done out of malice. Some of it is just caused by bugs.
- IncreasePosts 2y agoAlso, it is Firefox on the desktop. 90% of YouTube views come from mobile, and Firefox only has about 5% market share on the desktop. So desktop Firefox is a half a percent of the overall users of youtube. Is it any surprise that maintaining support for a browser that delivers less than a percent of the total users is deprioritized or just forgotten about?
- miah_ 2y agoDespite those stats, if Google wants to avoid being called a monopoly they need to support it, and it needs to receive the same level of support as their own product.
- mozempthrowaway 2y agoIt’s not monopolistic to not support shitty browsers with little user base. This isn’t even the basis for the current antitrust ruling against Google, for what it’s worth. As the parent comment points out, if YouTube isn’t buggy for most of their users, why do they have to worry about it? We don’t expect either Microsoft or game devs to ensure their stuff works well with wine on Linux.
- miah_ 2y agoSo you work at Google on Youtube then right? The _WEB_ is a open standard. A Game running on Windows is a false equivalence.
- mozempthrowaway 2y agoWeb standards are open but that doesn’t mean you can’t build on top of them. YouTube is not an open standard. If Google wanted to end web support for YouTube and force everyone to use some sort of YouTube app even on desktop, they could do that. They’re not obligated to make sure YouTube works on EVERY browser especially when an insignificant amount of their users come from there. A game running on windows isn’t a false equivalence either. A lot of games are built on open standards and open source (e.g. OpenGL, various game engines, etc) and still won’t work on wine/linux. If a bunch of people are using windows in order to play their games, that’s not too different from a lot of people using chromium browsers to use YouTube. Also, I work at Mozilla.
- positr0n 2y agoI'd argue the opposite. You need to be a monopoly and have basically every eyeball in the country on your website for it to be worth spending the engineering hours optimizing a browser that only delivers a fraction of a percent of your viewers.
- pessimizer 2y agoIf they're doing it to intentionally damage firefox because they feel it is competition, that's an argument for Chrome to be severed from google. If they're inadvertently doing it because firefox is so insignificant as not to be worth thinking about, that's an argument for Chrome to be severed from google.
- IncreasePosts 2y agoHow is Firefox having abysmal market share an argument that chrome should be severed from Google?
- seanw444 2y agoIn a vacuum, it's not. The market is completely dominated by Google though, and they exercise that authority and mindshare brazenly, to the point that it's easy to see how it affects Firefox. Firefox functions great. It doesn't function great with Google products/services such as YouTube in this case, which has no real competition. What's someone to do? Switch to Chrome. One might even make the assumption that the repeat offenses over time, across various services, across various fields, demonstrates... malice.
- benterix 2y ago> demonstrates... malice Not demonstrates, hints at. In each of these cases, Google is very careful that you can easily come up with an alternative plausible explanation. It's hard to use the en masse argument if each of these problems can be accounted for. These guys are not stupid - they're happy with the current status quo where they have a de facto monopoly but they can pretend they don't.
- lenkite 2y agoSorry, it is absolutely hard-core malice. No hinting, after we found out last year that a deliberate sleep was introduced with a user agent check. This is bald-faced murder. Its like shooting someone in public and stomping on the body. No need to dance around it. https://news.ycombinator.com/item?id=38345858 https://news.ycombinator.com/item?id=38345858
- ChocolateGod 2y agoI have a strong feeling the developers behind sites like YouTube and Reddit don't actually use the software they make, because both are slow and laggy to use on any computer I use. I just can't think of any other reason why they're both so infamously bad.
- ohthatsnotright 2y agoWorking on software rarely, especially at large companies, means you have any sort of agency over the features of the software. That's Not Your Job (TM).
- 01HNNWZ0MV43FF 2y agoAs they say, you aren't the customer, the advertiser is the customer. If the site works for the 95% of ad dollars who use Chrome on their phones, it works as intended.
- mrandish 2y ago> the advertiser is the customer. Excellent point. If any change on the site caused ad views or clicks to drop by >1%, automated tests probably shoot flares up to the exec level. Whereas the vast majority of users on FF are blocking ads anyway, so their lagging performance probably barely registers on those tests. While YT probably still has non-ad automated performance tests, in the case of a non-Chrome, non-Apple desktop browser, those tests probably run every odd Thursday and regressions send a toast message to an intern. :-)
- whstl 2y agoEven advertisers don't matter anymore, only metrics. Because they can be manipulated. The software engineering show isn't being run by engineers, middle-managers or even MBAs anymore. It's run by product people who treat it like a mini dictatorship. So releases like this have to go out, no matter the consequence.
- zagrebian 2y agoThey use it but in Chrome on a maxed out MacBook Pro
- neilv 2y agoSince YouTube has massive resources and famed tooling&processes, presumably either they knew about the Firefox problems before deploying, or (if there was a genuine QA mistake) they'd know about it very soon after deploy, when they'd be able to rollback if they wanted to. (Obviously, they know about Firefox, they've developed for it since it was available, and they've even been funding it.)
- PaulHoule 2y agoGoogle funding Firefox is part of the problem. It contributes to the perception of corruption around the Mozilla organization, just as it does when the bus has a supergraphic ad for a car dealership on it. If the EU was serious about privacy they'd fully fund Firefox.
- mozempthrowaway 2y agoThe perception of corruption is not without merit. Mozilla is pretty corrupt at this point. (Source: I work here). I really don’t get where this whole “the EU should fund it” idea came from and why it’s repeated so often. Why would the EU throw their tax payer money at another corrupt American corporation? Mozilla has been in bed with Google for several years, has horrible web compatibility, and is only barely still in the privacy lane. Besides, Europe isn’t the land of open source software and privacy. Look at the laws in the UK, France, and Germany; they’re not exactly privacy friendly. Look at the tech stacks at companies in the UK and Germany, they lean very heavily into the Microsoft/.NET world.
- PaulHoule 2y agoThen they should fork it and start a new organization. The EU and the world could have a privacy focused browser if they paid for it. If they don't they're going to always be waiting for the market to do it and it won't. Given that Mozilla wastes a lot of resources on things that are thoroughly pizzled (see https://en.wikipedia.org/wiki/Autofac https://en.wikipedia.org/wiki/Autofac) a browser could be maintained for much less than the current Mozilla budget. This is why "fully funded" as opposed to "funded" is key. So long as the organization feels the need to go around with its hat out it is going to be corrupt.
- genewitch 2y agoeven after load, youtube shorts (the doomscrolling section of youtube) has a 2-3 second lag between when the audio starts on a new video and when the video starts moving. seems monopolistic. what's goin on with peertube and rumble these days
- smegger001 2y agomy question is why isn't this being tested for before it is rolled out? shouldn't google be running testing on their sites for each of the major browsers before dropping it on everyone?
- aidenn0 2y agoAnother wild-ass guess is that a lot of garbage may be generated? I had to use an intranet site that was developed on an older version of vue, and that version apparently was a profligate allocator; difference between chromium and firefox was stark. On an only-slightly-related note, if anyone knows how to use the new firefox profiler when there is no internet access, please point me at it.
- rasz 2y ago>desktop_polymer.js its doing such fun things as patching tons of standard js methods (.append() .appendChild() etc) for no reason other than maybe fighting adblockers? When debugging own scripts/extensions running on YT every other call to standard function ends up calling some proxy in desktop_polymer.js