Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

Google’s AI-powered ‘inclusive warnings’ feature is very broken

vice.com

Post reply on HN
Print view View original

691–700 of 1001 posts

Re: Google’s AI-powered ‘inclusive warnings’ feature is very broken

#691

Post by politician »

31098347 wrote: post #659

There's a few different things to unpack with this feature. (1) It would be useful in the context of a general style guide, like a white-label Grammarly. Corporations could set their own prompts for words, phrases, and structures. This would make documentation more consistent. (2) This is dystopian as fuck. Google has the ability to see, aggregate, and now influence what you write in Google docs and Gmail. Who is mak…

Grammarly is also an OS-wide keylogger.

https://www.kolide.com/blog/is-grammarly-a-keylogger-what-ca...

They claim not to be a keylogger, but, you know, Amazon claims that Alexa isn't always listening too and/but/yet also that it'll wake up immediately when you say the right wake word. So, I'd take their claims with a lot of salt.

Re: Google’s AI-powered ‘inclusive warnings’ feature is very broken

#692

Post by UncleMeat »

Tr3nton wrote: post #620
UncleMeat wrote: post #593

Earlier quoted context omitted.

MLK explicitly supported affirmative action and other programs designed explicitly to benefit black people in order to make up for past discrimination. I suspect he would be tired of being used as a justification for absolute colorblindness.

So Affirmative Action is revenge discrimination?

No. MLK considered it to be a step on the path to equality.

Re: Google’s AI-powered ‘inclusive warnings’ feature is very broken

#693

Post by UncleMeat »

troupe wrote: post #641
UncleMeat wrote: post #593

Earlier quoted context omitted.

MLK explicitly supported affirmative action and other programs designed explicitly to benefit black people in order to make up for past discrimination. I suspect he would be tired of being used as a justification for absolute colorblindness.

What he stood for changed over time, but at one point he said he hoped the decisions people made about his children would be made based on his children's character and not the color of their skin.

Yes he did say that one line in one speech. But if you review the totality of his writing, it is very clear that total colorblindness in the face of inequality was not his goal. The "I Have A Dream" does not represent some shift away from an earlier opinion. It simply doesn't represent the opinion that a lot of reactionaries think it does.

Re: Google’s AI-powered ‘inclusive warnings’ feature is very broken

#694

Post by makeitdouble »

31098347 wrote: post #659

There's a few different things to unpack with this feature. (1) It would be useful in the context of a general style guide, like a white-label Grammarly. Corporations could set their own prompts for words, phrases, and structures. This would make documentation more consistent. (2) This is dystopian as fuck. Google has the ability to see, aggregate, and now influence what you write in Google docs and Gmail. Who is mak…

> Google has the ability to see, aggregate, and now influence what you write in Google docs and Gmail.

Interestingly Microsoft has that power with the Office suite for nearly 40 years now (only a decade if we want to focus on the cloud connected area), do you see specific dystopian influences on society stemming from that ?

Re: Google’s AI-powered ‘inclusive warnings’ feature is very broken

#695

Post by redox99 »

nullc wrote: post #556
gdulli wrote: post #29

I spent many years in the bluest corner of a very blue state. But still I can't imagine any objection to the word "landlord" outside of (1) a right wing attempt to satirize the left, (2) a Google product brainstorming meeting, or (3) a rare sincere outlier or concern troll. In any of those cases I don't think this is necessary. That said, if the word "landlord" organically fell out of favor and out of usage to be rep…

> (1) a right wing attempt to satirize the left It's hard not to secretly suspect that some of these things arise that way-- satire that is so spot on that it gets adopted as the truth. > and out of usage to be replaced by something else I wouldn't care Fundamentally that's why language bullying works-- it doesn't matter what words are used so long as the communicating parties understand each other. Not only does it…

> It's hard not to secretly suspect that some of these things arise that way-- satire that is so spot on that it gets adopted as the truth.

There have been many PR/issues on github that were either very likely, or certainly submitted by trolls, that have been accepted.

[1] Although nowadays it's common to see master/slave being replaced, back in 2015 it was not common, this one one of the first things that dropped that terminology. The fact that the user is called pcbro and has a picture of PC Principal from South Park, makes it somewhat likely that it was a troll.

[2] I think based on the language used is almost 100% certainly a troll

[1] https://github.com/apple/swift/pull/165 [2] https://github.com/microsoft/vscode/issues/87268

Re: Google’s AI-powered ‘inclusive warnings’ feature is very broken

#696

Post by arrakis2021 »

This is sure to dismantle systemic racism.

But in all seriousness I think this stuff is great. It pushes more people away from insane woke culture. I’m starting to see even the most progressive friends and colleagues pause and think twice about how cringeworthy all of this is.

Let it happen.

Re: Google’s AI-powered ‘inclusive warnings’ feature is very broken

#697

Post by Shaanie »

MikeDelta wrote: post #522
tedsanders wrote: post #272

Earlier quoted context omitted.

Putting myself in the shoes of a kid learning about computers, I find some of these phrases to be more clear, not less clear. Two examples: Allowlist and denylist are much clearer to a kid learning about computers than whitelist or blacklist. If you've never heard of whitelist before, it's sounds like a list of things that are white. If you've never of allowlist, it's pretty obvious what it means - a list of things t…

It appreciate the idea to make language more precise where possible. However, terms like motherboard and whitelisting are just a few in the universe of complicated (yet gender neutral) terms like GPU, CPU, DDR, parity, firewall, IP6, etc. Of course, that is not a reason to not improve the language where one can, but I don't know if it will help that much in CS.

In reality these types of changes often just make learning more difficult, because now you have to learn that it's called both "motherboard" and "mainboard". At least for a generation or two.

Re: Google’s AI-powered ‘inclusive warnings’ feature is very broken

#698

Post by ridaj »

aendruk wrote: post #627
ridaj wrote: post #603

Earlier quoted context omitted.

What about when the landlord is a company? Or an actual man in fact? We're going from unnecessarily gendered to unnecessarily neutered word.

Like most things, it depends on the situation. The point is to find something that does work. Surely there’s room for improvement?

A nudging system like this better have a much higher signal to noise ratio otherwise it really amps up the clippy-like annoyance. Keep in mind this isn't something you have to go turn on in a menu for "help me write inclusively", it's enabled by default for all; and while I might personally appreciate the reminders when they do work, I'm not sure as many users of the software think playing word police should be a priority.

Re: Google’s AI-powered ‘inclusive warnings’ feature is very broken

#699

Post by hnburnsy »

murderfs wrote: post #521
itsyaboi wrote: post #41

Earlier quoted context omitted.

Since you work at Google, maybe you can answer this: does internal tooling exist to flag occurrences of "bad" words in say, code reviews or shared documents?

Yes. There's even a bot that calls you out if you use "bad" words in internal chats.

So is there a committee or person responsible for maintaining the bot bad words list?

Re: Google’s AI-powered ‘inclusive warnings’ feature is very broken

#700

Post by sodality2 »

politician wrote: post #691
31098347 wrote: post #659

There's a few different things to unpack with this feature. (1) It would be useful in the context of a general style guide, like a white-label Grammarly. Corporations could set their own prompts for words, phrases, and structures. This would make documentation more consistent. (2) This is dystopian as fuck. Google has the ability to see, aggregate, and now influence what you write in Google docs and Gmail. Who is mak…

Grammarly is also an OS-wide keylogger. https://www.kolide.com/blog/is-grammarly-a-keylogger-what-ca... They claim not to be a keylogger, but, you know, Amazon claims that Alexa isn't always listening too and/but/yet also that it'll wake up immediately when you say the right wake word. So, I'd take their claims with a lot of salt.

> Amazon claims that Alexa isn't always listening too and/but/yet also that it'll wake up immediately when you say the right wake word

Strictly Alexa, the device, is always listening in order to trigger on the wake word. It's just that audio doesn't get sent to Amazon until it hears a wake word (then it just streams the audio to Amazon to process it. locally it's only smart enough to listen for a wake word). This is verifiable from sniffing network traffic, it's not sending enough to be a live audio stream at all times unless the wake word is said.

Post reply on HN