Hacker News Reader

Comment by crazygringo | Hacker News Reader

This should not be the #1 story on HN right now. :(

It's a single anecdote that deleting Chrome on two computers sped them up. It provides zero evidence (even anecdotal) that it has anything to do with "Keystone" specifically. It provides zero evidence for the idea that "Keystone" is able to "nefariously hide itself from Activity Monitor". And it also completely contradicts the normal user experience of Chrome, which is that most people's computers don't slow down after a Chrome install. (WindowServer on my MBP usually uses <10% CPU and I've used Chrome for many, many years.)

Why is this nonsense being upvoted? I get people dislike Google and Chrome, but wouldn't it be better to stick to things backed by real evidence?

View on HN

Indignation attracts upvotes, then flags. When an indignant post is relatively empty or misleading, flags will eventually win, but it's like the immune system—it takes some time to marshall enoguh white blood cells.

This is a nice case because moderators only saw it after the process was done.

Hundreds of people commented and how many tens of thousands read it first though? Reminds me of the moderation problem big tech companies face. Humans are prone to confirmation bias which means that when they read inaccurate information that confirms a bias it even further twist people's brains. For example, this article likely twisted thousands of people just a little further against Google even with though there was no good reason for it.

My grandpa twenty years ago had some legitimate gripes with the other side of the political aisle. However, he has seen so many biased and false news stories that his confirmation bias has accepted as truth without critical thinking that he now believes the other side is literally made up of evil criminals intent on destroying the country.

The only solution I can think of is something like a libel law that makes anyone who publishes false information liable to anyone who reads that information... Basically, cut out the enormous amount of garbage produced by media and people chasing pennies. People should only publish things they firmly believe based on evidence to be true.

That's a different and much harder problem. I don't think there's a way around it. Community mechanisms need time for people to evaluate what they're looking at.

It's also an old problem - the adage "A lie flies around the world while the truth is still getting its trousers on" traces back to Jonathan Swift:

Falsehood flies, and the Truth comes limping after it; so that when Men come to be undeceiv’d, it is too late; the Jest is over, and the Tale has had its Effect.

A perfect 18th century description of internet dynamics!

https://quoteinvestigator.com/2014/07/13/truth/

Dang, I can't tell you how much I enjoyed you applying that old quote to the modern day. The more things change...

The interesting thing here is I follow Loren on twitter - and he has dozens of credible people repeating the experiment and getting the same result, including people that I would consider unimpeachable sources such as:

https://twitter.com/benthompson/status/1338412989473382400

Well, I agree—that's interesting. It implies that something real is happening, though not that the OP's interpretation was accurate. It often happens in such cases that the observation is right but the interpretation is wrong [1]. It's when the interpretation rushes into angry denunciation, and especially a campaign using baity tricks to rile people up, that fair-minded readers start to have adverse reactions.

A discussion oriented more around a question (what explains this odd phenomenon?) could be a good thing. A more neutral submission, with a less baity title, might get a different reception from the community. But the combination of overstated claims and baity denunciation is usually going to trigger an immune response.

[1] I have no idea if it's wrong or not, just that a lot of informed commenters reacted that way.

Indeed. It’s almost as though Loren was reading this thread an took your advice. His post as of a few hours ago (he continues to troubleshoot this with google and tons of people experiencing the same thing)

https://twitter.com/lorenb/status/1339372013232779264

“ Also, sincere apologies for any misleading language in the original post, coming down off my fury that this was possible (helps not wanting to hurl your mac through a wall).

Thanks to everyone for coming through now with samples/evidence.

Something's up, but I'll shut up now.“

Thanks, as always for the calming and wise guidance Dang.

> The only solution I can think of is something like a libel law

Let's bring things a little bit back into perspective, because on the whole nobody's getting hurt here.

It's just Google Chrome. It's not a person. A thread like this isn't going to hurt its dominance of the browser ecosystem. And if it did, it would even be good for the diversity.

I don't think it would make a really big dent in people's bias against Google, because slowing your computer down is way way WAY less bad than the things that people biased against Google are worried about (being a gigantic faceless corporation with power equal to nation states that is controlled by something that is not, does not care for democracy).

It's like badmouthing Thanos because he has bad breath.

If true then Chrome has some bad code, and what else is new. If false then ... then what. At worst somebody deinstalls Chrome, and then?

Let's not bring libel law into this, or in fact even compare it to biased and false political news stories.

My WindowServer usage was over 90% before but it went back down to under 10% after I did what he said. I also only needed it to test one site I was developing. I had the exact same things he did: MBP slow even when nothing is running, WindowServer always running high. So I think I can vouch for it being true. It seems to have worked for me also.

EDIT: Maybe it's just placebo. I checked the CPU usage on WindowServer before and after deleting & restarting.

After reading your post I did the same thing. Similar results 90% to 20% in my case. Anecdotes aren't data but I am glad I tried.

Yeah, I had my friend try it as well because he said his was suggish as well. He said he had a similar experience

If you’re having visible performance problems, file a report with Feedback Assistant. It’s ok if it seems to be a 3rd party app causing the issue.

(High CPU use isn’t necessarily a real performance issue and might be caused by just having Activity Monitor open.)

Apple does not support or fix third party apps.

This is 100% chrome, with their annoying updating agent, which you can't remove

You're awfully confident for someone who has provided no additional evidence.

The OP claims the high CPU usage went after uninstalling chrome and rebooting their machine.

Rebooting their machine.

The thing we tell everyone to as the absolute first step in trying to solve any problem with a computer, "have you tried turning it off and on again?".

This isn't proof that chrome or keystone or anything at all caused the problem. We have nothing to determine the root cause.

I've had numerous problems with keyagent. It was horrible. I spend way too much time on trying to get that monster to work normally or get it removed.

But it's an "essential part" of chrome

Apple likes it when the products they sell work properly and would like to hear about it when common third party products are causing that to not happen.

Yeah, next time I'll try that as well! Completely forgot Feedback Assistant existed, but I'm glad it is back to normal speeds now.

Maybe you are right. A lot of comments here are complaining about the post in terms of scant evidence. Someone could pretty easily myth-bust this, but I have not seen that in the comments here yet.

This is a fair comment.

At the same time even "hard" evidence would likely get dismissed as anecdotal, and there's certainly enough of it now (and even plenty in the past) to point a clear finger at Chrome/Keystone.

This certainly beat filing it in the black hole that is Chromium's bug reporter where it would have been ignored / works-on-my-machine'd / or dismissed as anecdotal there.

Whatever it is doing is sketchy and causing WindowServer to thrash. And this is not the first sketchy thing it has done.

I think it's kinda sus that Google doesn't provide any sort of instructions or script to uninstall Keystone. Why not?

It's literally Chome Bad = upvote

I mean Chrome is bad on Macs in terms of resource usage and speed.

But this post alleges it is really bad when not running, which seems like huge news if true.

I don't see enough evidence here, but I'd be open to the idea that this is true if someone looked into it deeper.

It would have been a lot better as "Having CPU Load Issues you can't identify? This might be worth a try"

'how should communities amplify or suppress anecdotes' is a good question

'facebook advertises based on IRL conversation' is another example of something that feels true to people but nobody has made stick as far as I know

upvotes don't always mean 'yes, guilty, firing squad'. I think they can mean 'yes tell me more, let's bookmark and start investigating'.

Agreed, about 25% of the content on HN seems to be Google hate.

Google, Amazon and Facebook. Microsoft gets it in the comment section, mostly. Only Apple is truly safe on HN.

>Only Apple is truly safe on HN.

Not even close to being true...

Yea, I would've agreed ten years ago but Apple gets plenty of criticism here these days.

None of the top 30 stories right now mention Google

I will say that it seemed like this was a pretty big story to be not be covered in the tech press. I do wonder if there really is much evidence.

On the other hand, I generally feel that Chrome is garbage that slows computers down, so I wouldn't be surprised. But I don't see enough evidence here.

I think the thing is that the title is ‘chrome is bad’, and that’s something most of the crowd here can sort of agree with for a variety of reasons.

Why don’t you let the vote buttons do the voting? If you’re not interested in content you can ignore it and not upvote it

Rust = Good

Google = Bad

This is peak hacker news. If this were a comment it would be deleted, but for headlines, the less information the better.

Comment. (comment.com)

Reply on news.ycombinator.com