Comment by crazygringo | Hacker News Reader

Comment by crazygringo | Hacker News Reader

Parent Full thread

I don't think any of that is accurate though.

Manifest v3 blocks user tracking -- if the request is blocked, any tracking attached to it is blocked. I'm sure it's not 100% perfect, but it's certainly working well enough in practice.

And what malware are you talking about? If a request is blocked, it's blocked. It doesn't matter if it's an ad or malware.

Manifest v3 is better at #4, because the junk isn't loaded, and the blocking is more efficient in terms of CPU.

And then #5 I don't know what you're talking about. I use Stylus and Tampermonkey to customize webpages and they continue to work great.

So I just don't see the evidence that "Google is 100% boiling the frog here". That's what everyone was saying, but now that Manifest v3 has come out, I just see adblocking that continues to work and uses less CPU to do it.

I see a lot of fearmongering around Google, but now that the results are in with Manifest v3... they just don't seem true. You're making all these claims, but I just don't see the evidence now that we're seeing how it works in practice.

Explain to me how uBlock Origin can realistically go from 100,000 to 500,000 dynamic rules down to 30k rules(only 5k of those can be dynamic) in the Lite version without losing the ability to actually block everything?

These limits are easy targets for ad networks to overwhelm or outmaneuver.

    That's what everyone was saying

Everyone was saying that the new API is less capable than the old API at blocking things. DeclarativeNetRequest IS less capable; that's just a fact.

No one was saying that adblockers would literally stop working, so it's beyond disingenuous to dismiss people's issues with these changes by just saying 'works for me'.

What evidence would you actually accept anyway? Do you need a leaked internal document from Google saying literally 'devs, go neuter adblockers' before you believe Google might have bad intentions surrounding people's ability to block ads and tracking?

If security and performance were the actual driving forces of DeclarativeNetRequest, then they would have simply added it in addition to the existing webRequest block functionality. uBlock Origin and most extensions would have happily moved the majority of their rules to the static list if it meant better performance and privacy while keeping around the webRequest blocks for the things that actually need it.

Google has gone from having only one nuclear-level option for influencing adblockers (aka delisting) to now having its boot softly pressed against their necks and plenty of levers to pull. And you want me to look at that and go, 'There's no direct evidence of malicious intention there... so perfectly normal and/or acceptable behavior by the world's biggest ad company'?

> Explain to me how uBlock Origin can realistically go from 100,000 to 500,000 dynamic rules down to 30k rules(only 5k of those can be dynamic) in the Lite version without losing the ability to actually block everything?

I don't know and I don't have to. All I know is uBlock Origin Lite is still blocking everything. So it seems like 30K rules is plenty? Like it's not a meaningful difference for end users if it's blocking 99.99% vs 99.9999% of ads?

> No one was saying that adblockers would literally stop working

That's sure what it sounded like. That it would literally be so bad you'd have to switch browsers because of how degraded the experience would be.

> What evidence would you actually accept anyway?

The fact that the adblocking experience was significantly degraded for the average user -- e.g. that now 10% or 25% of ads were getting through.

> And you want me to look at that and go, 'There's no direct evidence of malicious intention there... so perfectly normal and/or acceptable behavior...

Yeah, pretty much. As far as I can tell, security and performance seem to justify the Manifest v3 changes. Occam's Razor says you don't need anything else. If you think there's malicious intention, then the onus of proof is on you.

I was told, time and time again, than Manifest v3 would result in an adblocking experience so bad that people would start switching browsers because of it, that Google was cracking down on adblockers to neuter them. Now that it's here and my adblocking works just as well, maybe even better (if it's sped up page loading times) -- then sorry, as far as I can tell the malicious intention was made-up.

> That's sure what it sounded like. That it would literally be so bad you'd have to switch browsers because of how degraded the experience would be.

> I was told, time and time again, than Manifest v3 would result in an adblocking experience so bad that people would start switching browsers because of it

Once enough ads catch up with the new limitations. Right or wrong, we're still too early for that.

> Explain to me how uBlock Origin can realistically go from 100,000 to 500,000 dynamic rules down to 30k rules(only 5k of those can be dynamic) in the Lite version without losing the ability to actually block everything?

I will take this one.

First, your limits are out of date. The static minimum is 30k, but can now escalate to an order of magnitude higher depending on how many extensions are installed. The dynamic limit is now 30k, of which at most 5k can be "unsafe". Source: https://developer.chrome.com/docs/extensions/reference/api/d...

Second, even if the limits were correct: consider the possibility that 99% of those rules are irrelevant, out of date garbage that blocks nothing anymore but haven't been removed because there is neither process nor incentive on the extension dev's part to do so.

Ad uses pattern. UBO adds matching pattern. Ad switches to new pattern. Cat and mouse.

This happens widely, rapidly, and on an ongoing basis. The result is that the rule set is large and grows rapidly, but very little of it is actually useful day to day. From the user's perspective, the only cost is that the browser very slowly gets continually less performant, which they will not attribute to the extension.

This isn't hypothetical. I'm on the Chrome team. We analyzed the rule set contents. This is why we proposed the initial limits we did: they were plenty large enough to allow all the extensions we analyzed to do everything they actually wanted to do, if only you stripped the cruft.

The rule size increases since then primarily come out of a dialog process with ad blocking devs about their process and needs and what they see in the wild coupled with what we think we can manage to keep performant. There are compromises. I'm not on that team so I can't speak to details. But it's part of an honest attempt to have a dialog.

There are usually simple explanations for things, if people were truly willing to consider them without bias.

Thank you for providing this valuable explanation -- I haven't heard this perspective expressed elsewhere. The fact that old rules would never get deleted but continue to drain resources makes some design decisions make a lot more sense.

One particularly pernicious outcome is that some ad blockers tout their rule set sizes as a feature, and users choose among blockers based on it, when if anything it is probably negatively correlated with blocker quality -- it's not necessarily a sign you're comprehensive so much as that you don't care about efficiency or cleaning up after yourself.

That's of course an oversimplification. But people who believe they're technically knowledgeable and adept are just as likely as other folks to fall for bullshit and be convinced to do things contrary to their own self interests. It's just a different type of bullshit.

No one wants to hear that, because we all want to tell ourselves that maybe everyone else is gullible, but WE'RE smart and rational. To a close approximation, though, none of us are.

Now consider that the extra drain can be practically zero, and you get back to those decisions making less sense

Google lifting the arbitrary rule count limit would go a long way towards building trust.

Reply on news.ycombinator.com