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

AI researcher Timnit Gebru resigns from Google

platformer.news

Post reply on HN
Print view View original

371–380 of 926 posts

Re: AI researcher Timnit Gebru resigns from Google

#371

Post by dash2 »

GVIrish wrote: post #270
jeffbee wrote: post #51

Earlier quoted context omitted.

I don't think modern HR policy really works like that. If one person complains to their HRBP that the speech or actions of another person at the company makes them feel personally attacked, derogated, or establishes a hostile work environment then the company will be forced to take that matter to the second person. HR isn't looking to setup an Oxford debate between persons A and B. They are required to resolve the ho…

Sure but HR wouldn't share the feedback on her paper and give her an opportunity to update it. Instead they demanded that she retract the paper. How is it fair or standard to make someone retract a paper because of anonymous and secret criticisms that the researcher isn't even given a chance to address?

Actually, to be fair, anonymous referee reports are absolutely standard in academia, and they may be standard part of Google practice in this field - and if so that might be good practice. Waiving anonymity could compromise the integrity of the intellectual process.

Re: AI researcher Timnit Gebru resigns from Google

#372

Post by malcolmgreaves »

A woman of color that spoke truth to power. And surprise -- the old white guy running the show didn't like it. Sad -- Jeff Dean's lost an incredible amount of respect and trustworthiness. Looks like he's sold out and is now a Google-guy (see reports of his email to Brain staffers in which he has multiple, easily refutable lies).

Re: AI researcher Timnit Gebru resigns from Google

#373

Post by bitcharmer »

contemporary343 wrote: post #233
Bukhmanizer wrote: post #219

Earlier quoted context omitted.

I work in research, and citation count, especially of a single paper, is such a dumb metric for evaluating performance. There are about a million reasons that one paper gets cited over another that don’t have to do with quality. Does the same paper get 1000+ citations if it were published in 2010? Or even 2015? Probably not.

It's an original research paper (not review) and it was published in 2018. So yes, 1000 citations at the very least conveys that it was noticed and had an impact. Also, anyone who works in the field of AI fairness and ethics research knows about it. So yeah, it's an important study.

Andrew Wakefield's paper which started the anti vax movement got much more citations than that.

What does that tell you?

Re: AI researcher Timnit Gebru resigns from Google

#374

Post by dash2 »

black_puppydog wrote: post #220
southstreet wrote: post #132

Earlier quoted context omitted.

"feedback" here seems to refer to feedback on the scientific paper that is being retracted, not personal performance feedback or anything like that. When I red this first, wasn't immediately clear what this was about.

Good point, that also took a bit of disentangling for me. As such, I find it really quite baffling that someone would be unwilling to share that scientific feedback with her, or who raised it.

Some refereeing processes are anonymous, and there can be good reasons for that, like preventing academic logrolling. It's standard in academia.

Re: AI researcher Timnit Gebru resigns from Google

#376

Post by malcolmgreaves »

VHRanger wrote: post #281
curiousllama wrote: post #60

Earlier quoted context omitted.

Even if it were the case, responding to "retract this paper" with "No; I prefer to resign" is/should be par for the course for a researcher.

They submitted thee paper for internal review one day before their external deadline. They set themselves up for failure regardless of the paper's content.

Their paper was accepted by the conference's peer review system. Google, not researchers in the community, wanted it retracted. The only logical explanation is that it was good science that did not put Google is a positive light.

Re: AI researcher Timnit Gebru resigns from Google

#377

Post by qppo »

inglor_cz wrote: post #327
qppo wrote: post #296

Earlier quoted context omitted.

Diversity initiatives are like a graduated social tax. As a passably straight white man I can afford to pay more than an LGBTQ person of color. Just like I can pay more in income taxes than people in different financial situations. It doesn't bother me if an applicant pool is made a bit bigger for marginalized people than myself, because I do not struggle to fall into any applicant pool whatsoever and have not for mo…

"As a passably straight white man I can afford to pay more than an LGBTQ person of color." How is this true? What if someone is a broke straight white man with three kids and someone else LGBTQ person of color is a young childless heir of fortune? How is the first person possibly privileged in comparison with the second one? Looking from Europe, America seems to be very determined to recast its very obvious class pro…

I mean if you compare a rich person to a poor person yes the rich person is going to come out ahead in most social metrics but we're not speaking of wealth inequality, but social divides outside of it.

That said, in the United States, non-white men were systemically marginalized and prevented from participation in society to the same degree. There are numerous examples of this over the last four centuries, but they did not decrease in fervor until about a generation ago. For many reasons, our class divide is correlated to our racial divide. That doesn't mean we don't need to fix both, just that the there are two problems to fix there.

Looking from America, it's hard not to be frustrated at Europeans on their high horses talking about class versus racial divide. Is it so hard to recognize that we have different social problems, and a different history? And that things we talk about might be colored one way, because race, gender, and sexual orientation impact real people on a daily basis? Do you reject the idea that bigotry exists independent of class and it's something we need to work on?

Re: AI researcher Timnit Gebru resigns from Google

#378

Post by joshuamorton »

inglor_cz wrote: post #327
qppo wrote: post #296

Earlier quoted context omitted.

Diversity initiatives are like a graduated social tax. As a passably straight white man I can afford to pay more than an LGBTQ person of color. Just like I can pay more in income taxes than people in different financial situations. It doesn't bother me if an applicant pool is made a bit bigger for marginalized people than myself, because I do not struggle to fall into any applicant pool whatsoever and have not for mo…

"As a passably straight white man I can afford to pay more than an LGBTQ person of color." How is this true? What if someone is a broke straight white man with three kids and someone else LGBTQ person of color is a young childless heir of fortune? How is the first person possibly privileged in comparison with the second one? Looking from Europe, America seems to be very determined to recast its very obvious class pro…

There are still absolutely situations where I'd prefer to be the first person over the second. For example, when interacting with a police officer, or when visiting certain towns. Or when attempting to get medical care (there's strong evidence to suggest that physicians are worse at treating black patients, and are prone to ignoring complaints from women, independent of class).

While class problems do, absolutely, exist in the US, race problems also exist.

Re: AI researcher Timnit Gebru resigns from Google

#379

Post by jerome-jh »

black_puppydog wrote: post #2

I think this is actually the better link to be discussing, rather than the twitter threads shared before. The most relevant parts (in terms of back story) should be this: > A week before you go out on vacation, you see a meeting pop up at 4:30pm PST on your calendar (this popped up at around 2pm). No one would tell you what the meeting was about in advance. Then in that meeting your manager’s manager tells you “it ha…

> behavior that is inconsistent with the expectations of a Google manager.

Maybe my remark is down to earth but, she literally told her group to stop working. I do not know of any company where this is an acceptable behavior for a manager. Besides the disagreements, it sounds to me that was the action that caused Google to pull the trigger.

Also from Dean's email it is pretty clear there are a number of new products down the pipe based on AI. At one point top management decided the company should stop denigrating its own products.

Re: AI researcher Timnit Gebru resigns from Google

#380

Post by dash2 »

adamsea wrote: post #369
dash2 wrote: post #355

Earlier quoted context omitted.

I'm torn on this. Actually, if you think you're a researcher and then feedback comes in this weird way, that seems pretty weird. At the same time... "people like me who are constantly dehumanized" - really? I mean maybe, but to me this sounds like hyperbole, and actually, weaponized hyperbole. This is someone who's a respected and presumably very highly paid researcher. Dehumanized ?

“People like me,” not referring to her privilege, but to women and people of color, I would guess. Since there are serious culture and systemic biases / prejudices against women and PoC.

Sure, but either she's constantly dehumanized or she isn't, and if she's actually rather powerful and important, and is then claiming that by being a black woman, she automatically gets a share in the victimhood, then that doesn't seem very convincing.
Post reply on HN