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

We read the paper that forced Timnit Gebru out of Google

technologyreview.com

Post reply on HN
Print view View original

641–650 of 683 posts

Re: We read the paper that forced Timnit Gebru out of Google

#641

Post by akie »

whatthesmack wrote: post #262
akie wrote: post #230

Earlier quoted context omitted.

> a far-left “woke” interpretation of the world that I view as of immediate threat to our way of living, democracy and free speech. WTF dude. The core argument as to what you say is a "woke" interpretation of the world is to reshape the world in such a way that inherent privilege (for example being rich, being white, being male) does not completely predetermine anyone's outcomes. So that you can still "make it" even…

It seems like all of the actions to "rectify" the aforementioned issues clearly involve threats to democracy and free speech. For example, the far-left "woke" interpretation has historically made use of and encouraged "cancel culture" to punish people that disagree with that worldview, which deprives them of their right to free speech. This is not imagined -- many people's lives have been destroyed because of this. F…

I'm not sure how anyone can have this view of the world and keep a straight face.

However, here

> It seems like all of the actions to "rectify" the aforementioned issues clearly involve threats to democracy and free speech.

you at least acknowledge that there is a problem. So we just need to put our heads together and try to solve these problems... right?

Also, as a child of (at least some) privilege it takes a lot to see and understand the privileges you have. I hope you see yours, and I hope that you would want other people to have the same privileges that you have.

Re: We read the paper that forced Timnit Gebru out of Google

#642

Post by mustuhfa »

textgel wrote: post #521
mustuhfa wrote: post #412

Earlier quoted context omitted.

Why are you assuming anything about me? You realise I e used the word toxic in my whole comment history only related to this topic right? I, like many others, dont like the way she deals with people. It is toxic. You call a toxic person, a toxic person. Ample evidence for it. Its not outrage, its just facts.

I think you replied to the wrong person?/friendly fire? My message wan in reponse was to. > Toxic is an inflammatory and unnecessary word to use when no-one is privy to the actual facts.

Ah ok, apologies!

Re: We read the paper that forced Timnit Gebru out of Google

#643

Post by bisectable »

d_e_solomon wrote: post #533
thu2111 wrote: post #512

Earlier quoted context omitted.

We're talking about tech firms that have access to the entire internet and use it. Hypothetical examples involving imaginary datasets that nobody would use don't prove anything relevant. And note that my argument is not about whether you actually avoid bias, it's about whether you've done the most you can do to avoid it. If you used all the data you've got, then you've done the most you can, even if for some reason t…

> We're talking about tech firms that have access to the entire internet and use it. I think you're trying to say that a large enough dataset will be free of bias. I don't see how that follows. If I train a model on home mortgage decisions, I will replicate the bias on that currently exists on that dataset - https://news.northwestern.edu/stories/2020/01/racial-discrim... - unless there are conscientious choices to re…

Given the current public discourse climate in academia, where 99% of the researchers are politically left or extreme left, I recommend a grain of salt for any research that makes claims racial discrimination. The report you quote has at least a couple of strange holes. The paper itself is behind a paywall, thus beyond mere mortals reach.

> For example, in about 10% audits in which a white and an African-American auditor were sent to apply for the same unit after 2005, the white auditor was recommended more units than the African-American auditor. These trends hold in both the large HUD (Housing and Urban Development)-sponsored housing audits, which others have examined with similar findings to us, and in smaller correspondence studies

They fail to mention how large is the gap. Is the white auditor recommended 102 vs the black auditor 97, or 150 vs 50, or 200 vs 3? Without such critical information it is hard to form an opinion, unless one already has a large bias in accepting discrimination narratives uncritically.

> In the mortgage market the researchers found that racial gaps in loan denial have declined only slightly, and racial gaps in mortgage cost have not declined at all, suggesting persistent racial discrimination. Black and Hispanic borrowers are more likely to be rejected when they apply for a loan and are more likely to receive a high-cost mortgage.

They fail to mention the magic words 'when controlled for income'. America has a huge income disparity problem, which is conveniently forgotten behind the ongoing race (and gender) hucksterism. Assuming we'd wave a magic wand and fix all disparities across visible populations tomorrow, it will still not fix the fact that huge income disparities exist between individuals. Google engineers and researchers get paid 5 times the median national income or more, and (senior) Google management in the 10x to 1000x range. The vast majority of the population is stuck in dead end precarious jobs, with little social mobility, one medical emergency from bankruptcy.

Re: We read the paper that forced Timnit Gebru out of Google

#644

Post by int_19h »

kenjackson wrote: post #510
tinyhouse wrote: post #503

Earlier quoted context omitted.

Well, nothing in this story suggest they do. They might have an inconsistent review process, like pretty much most processes in these companies - from interviews to promotions. Let's see how many people follow her and leave Google Research due to their practices. My prediction is zero.

That’s the beauty of modern discrimination. You never come out and say you’re discriminating. Rather you just apply the rules selectively. An inconsistent process is the best best way to discriminate. My prediction is you will see some people leave Google due to this. Likely those who worked closest to her. Apple is already extending interest.

Modern? It's a trick that's probably as old as discrimination itself. Here's a rare case of an American judge speaking openly about this - the quote is from a 1941 Florida Supreme Court opinion (Watson v. Stone):

"I know something of the history of this legislation. The original Act of 1893 was passed when there was a great influx of negro laborers in this State drawn here for the purpose of working in turpentine and lumber camps. ... It is a safe guess to assume that more than 80% of the white men living in the rural sections of Florida have violated this statute. ... There has never been, within my knowledge, any effort to enforce the provisions of this statute as to white people, because it has been generally conceded to be in contravention of the Constitution and non-enforceable if contested."

Note, by the way, that this also references the other common trick - you can also make rules that are seemingly not arbitrary, but are easy enough to challenge for those with the right knowledge and/or means.

Re: We read the paper that forced Timnit Gebru out of Google

#645

Post by mam2 »

int_19h wrote: post #629
mam2 wrote: post #608

Earlier quoted context omitted.

I mean she did not HAVE to make an ultimatum. Real change takes hard work and time, and patience, and also playing politics with powerful people you dont like. Sounds like she basically made a one sided hit piece which was against the work some teams of brain, and took the "my way or the highway" approach. That wasn't so smart from her either...

Playing politics can take many different ways, and creating a scandal to draw attention can also be very efficient.

We'll see.. but minority activists have a tendency to use drama and twitter outrage a lot.

But if you take the minority part, google has a point: as a manager, she had a setback. Now instead of sucking it up and maybe wait for the next submission deadline and modify her article, she went and complain its discrimination on a internal mailing list of employees and then making ultimatums. :roll-eye: Any manager would have been fired.

Its only a special situation because she's black and work on ethics. Its privilege actually, more than discrimination here.

Re: We read the paper that forced Timnit Gebru out of Google

#646

Post by textgel »

mustuhfa wrote: post #642
textgel wrote: post #521

Earlier quoted context omitted.

I think you replied to the wrong person?/friendly fire? My message wan in reponse was to. > Toxic is an inflammatory and unnecessary word to use when no-one is privy to the actual facts.

Ah ok, apologies!

No problem :)

Re: We read the paper that forced Timnit Gebru out of Google

#647

Post by foota »

StavrosK wrote: post #174
foota wrote: post #139

Earlier quoted context omitted.

Her previous papers look pretty serious to me, along with her resume. I see nothing to suggest she's not qualified. https://ai.stanford.edu/~tgebru/

What do you mean by pretty serious? I looked at three, two are about better documenting ML models/data, and one is about inferring the political leanings of an area by what type of cars are parked in it. None seem particularly groundbreaking to me...

Honestly? I don't have much experience with ML and in particular issues of bias in ML (though I have a general understanding of a lot of the former), so I don't really feel qualified to evaluate her more recent work, but her older stuff definitely appeared like what I'd expect from a machine learning practitioner, so I don't think attacking her as someone that only understands social sciences (which I perceived the parent as doing) to be fair.

Re: We read the paper that forced Timnit Gebru out of Google

#648

Post by derbOac »

ghaff wrote: post #449
Wowfunhappy wrote: post #441

Earlier quoted context omitted.

I'm less cynical about this as a concept. Would you feel differently if, say, Apple hired a "Privacy Watchdog", with a long history of activism on the topic? Someone you could trust to speak out if something was amiss. If the person is later fired, that's a sign that something is wrong at the company. But if they stay , and have generally good things to say, that's a sign the company can be trusted. I do think this i…

It's a difficult type of role. See the NYT getting rid of its public editor/ombudsman.[1] I'd actually have trouble seeing Apple hiring someone like that given how tightly they control external messaging/speaking/etc. It's really hard for execs, including but not limited to PR, to have worked really hard on getting some message/narrative accepted by the media/analysts/etc. and then have a high profile employee on soc…

And yet here everyone is, discussing whether or not Google is secretly racist, with protest petitions involving 1000s of signatures. I had never heard of the name of Gebru before but now I have.

I'm not meaning to take sides here, and I think what you're saying needs to be kept in mind in understanding this, but it's hard for me personally at least to not see this in terms of some diversity-speech Streisand effect.

Re: We read the paper that forced Timnit Gebru out of Google

#649

Post by jart »

thu2111 wrote: post #376
hezag wrote: post #49

I did not read the paper (just like most people here), but by the title — “On the Dangers of Stochastic Parrots: Can Language Models Be Too Big?” — it does not look like the CO2 emissions thing is the main topic of this research. BTW, "Stochastic Parrots" is a very descriptive name for the problem > Moreover, because the training datasets are so large, it’s hard to audit them to check for these embedded biases. “A me…

Why? It sounds like ideologically driven circular reasoning. If you train an AI on the largest dataset it's possible to obtain then you have, almost by definition, done the most you can to avoid bias of any sort: the model will learn the most accurate representation of reality it can given the data available. Gebru is the type of person who defines "bias" as anything that isn't sufficiently positive towards people wh…

The greatest struggle facing the tech elite today is how to use the proletariat's data to train an artificial intelligence that doesn't inherit the proletariat's beliefs.

Re: We read the paper that forced Timnit Gebru out of Google

#650

Post by dragonwriter »

chishaku wrote: post #497
albntomat0 wrote: post #485

Earlier quoted context omitted.

I think we agree on the difficulty here. I believe the reverse is also true: if the person is fired for a legitimate, non-watchdog reason, they have an incentive to say the firing was due to them bring up privacy/ethics concerns.

believe Google vs believe a solitary worker (female, black, known political agenda)

> believe Google vs believe a solitary worker

But it's not a solitary worker, many Google AI workers have spoken up and said that even if the management narrative is true as to what happened in the specific case, it would represent a radically different review process — both as to what the review looked at and the request to retract rather than correct when there was a plenty of time for corrections for the conference — than is applied in Google AI research to other workers.

This strongly suggests that even if Google's incident-specific behavioral narrative is true as far as it goes that the entire incident was constructed as personally-targeted harassment, and that the narrative about qualitative motivations and regular process are false.

Post reply on HN