Given Google's domination in the browser market, there really needs to be some accountability. Web browsers are an incredibly important tool for most(all?) businesses and if changes are not documented (including stealth patches) this is very concerning.
I wish this gave Chrome the corporate "death penalty", i.e. kicking it off of corporate infrastructure.
Google being able to silently update the browser might also have GDPR implications as no-one can be sure that the next version (you didn't notice) exfiltrates sensitive data.
Sounds like a pretty ordinary bug, combined with people not knowing what "experiment" means in the context of Chrome.
When new functionality is added to Chrome (or Firefox, or most other large software projects with automatic updates), you don't add it for everyone at once; if you did that, and it turned out to be buggy, everyone would be exposed to the buggy version. Instead you roll it out for a few users, and gradually expand the subset of the userbase which has the feature, while watching for both automatic and manual bug reports. If the bug reports don't come in, you enable the feature for everyone. While the feature is turned on for some people but not for everyone, it's called an "experiment".
You don't think there's any difference between what Google is doing and the protocol for signing people up to a drug trial? Is it obvious they are equally consensual?
10 comments
[ 2.8 ms ] story [ 34.2 ms ] threadOr would a higher up have had to have signed up on this? Yikes!
Generally, de-centralized decision making power is good within parameters/guard rails.
Google being able to silently update the browser might also have GDPR implications as no-one can be sure that the next version (you didn't notice) exfiltrates sensitive data.
When new functionality is added to Chrome (or Firefox, or most other large software projects with automatic updates), you don't add it for everyone at once; if you did that, and it turned out to be buggy, everyone would be exposed to the buggy version. Instead you roll it out for a few users, and gradually expand the subset of the userbase which has the feature, while watching for both automatic and manual bug reports. If the bug reports don't come in, you enable the feature for everyone. While the feature is turned on for some people but not for everyone, it's called an "experiment".
Yes, because it's not communicated at all to non-technical users that this stuff is happening at all.
https://news.ycombinator.com/item?id=21542716
https://bugs.chromium.org/p/chromium/issues/detail?id=102483...