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

Google fires another AI researcher who reportedly challenged findings

engadget.com

Post reply on HN
Print view View original

221–230 of 238 posts

Re: Google fires another AI researcher who reportedly challenged findings

#221

Post by faeriechangling »

bitshiftfaced wrote: post #218
faeriechangling wrote: post #188

Earlier quoted context omitted.

It’s not disputed by anything other than googles PR. By well established California law legal resignations include an end date, she did not ever specify an end date when threatening resignation, and thus she never legally resigned. She was thus fired. It would be bias in googles favor to act as though she were not fired which is what any court would find. Sort of a “discuss the controversy” sort of thing over a very…

> By well established California law legal resignations include an end date, she did not ever specify an end date when threatening resignation, and thus she never legally resigned. She was thus fired. What law is that? Either way, I'm confused about what the issue is here. California is an at-will state, so she could quit at any time, and they could fire her at any time. When you make an ultimatum, you're saying you'…

It's case law, I can't find the link to the case I saw the first time this story ran that made it clear that Gebru was fired and did not resign. I did find this however:

https://edd.ca.gov/en/uibdg/Voluntary_Quit_VQ_135

>In P-B-39, the claimant gave notice on October 24 that she was quitting effective November 15. The employer permitted her to work only until October 31. The Board held that the claimant was discharged and said:

>. . . the claimant was not permitted to work to the effective date of her resignation and the employer did not pay the claimant her wages through that date. The claimant did suffer a wage loss by the action of the employer in accelerating the last day of work.

Gebru was going to arrange a meeting with Google to discuss when she would resign, and Google opted to claim she had already resigned and simply fired her. This constitutes a firing as Gebru clearly did not intend an immediate resignation.

>Either way, I'm confused about what the issue is here.

I bring this up merely because the neutrality of the journalists reporting on the case has been brought into question, and I want to defend them describing the situation as a "firing" as being the neutral thing to do. Additionally, I think Google's twisting of labour law for Google's PR/Financial benefit speaks to Google's character in the entire affair, and we must fairly consider Google's actions when judging Gebru's, whose actions were controversial.

Re: Google fires another AI researcher who reportedly challenged findings

#222

Post by chaps »

yodon wrote: post #176
chaps wrote: post #172

Earlier quoted context omitted.

The point I'm hoping to show isn't a "smoking gun" (frankly, I doubt any single document, or set of documents, would meet your criteria), but to show that research is deeply motivated by funding rather than for "pure science". Bad results that don't fit the narrative of the funders or political body's interest either don't get published, or are downplayed to the extent that they're effectively unpublished. Anything r…

Reading your second set of links, it sounds like much more a case of "the media does a bad job of reporting on science" than "the research is corrupt." If you want to improve the quality of science reporting, that's probably better handled by criticizing the reporters than by criticizing the researchers (most researchers already work hard to keep their work out of the popular press, precisely because the popular pres…

You're reading things at a very surface level, friend. Good luck to you and yours.

Re: Google fires another AI researcher who reportedly challenged findings

#223

Post by boppo1 »

BeetleB wrote: post #123
pharmakom wrote: post #114

Earlier quoted context omitted.

If you cannot give an ultimatum then you can always be pushed around.

> If you cannot give an ultimatum then you can always be pushed around. No finer example of false dichotomy than this. Hers is a textbook example of how not to negotiate. By textbook, I mean literally every textbook on negotiations that I've read. Ultimatums are a tool of last resort, and they all caution their high failure rate. It also is typically a signal of a weak person.

>literally every textbook on negotiations

Aside: can you recommend the best ones? Getting to Yes is all I've read, but it seems a little 'fluffy'. I want something more rigorous, perhaps with game theory and models.

Re: Google fires another AI researcher who reportedly challenged findings

#224

Post by boppo1 »

throwawaygh wrote: post #209
brian_cloutier wrote: post #115

Earlier quoted context omitted.

I wonder if you've had an in-person conversation with someone who would identify as being a part of the AGI safety community. I've talked with several people and their arguments are a lot more sophisticated than an intuition that ML models would have human motivations. This post isn't written by somebody inside the community but he presumably has access to them and has had conversations with them which shape his beli…

> I wonder if you've had an in-person conversation with someone who would identify as being a part of the AGI safety community. Yes. Many. > I've talked with several people and their arguments are a lot more sophisticated than an intuition that ML models would have human motivations. I don't have to think much about refuting this. Sure, okay, sophistication. Or not. Whatever. The sophistication is still mostly philos…

>Even in the limit, where ML sysetems do some non-trivial set of human-like tasks (which we aren't even remotely close to yet, btw)

I'm a layman in AI/ML and I've seen this stated a lot by people actually programming ML stuff (as opposed to "working in the space" as a blogger/manager/marketer etc.)

How can I concretely convey this concept to friends & family panicing about AI from crap they read in NYT/Economist/WSJ etc.? 'Some guy on HN who sounded like he knew what he was talking about says that article you read is sensationalist' doesn't pass their 'expertise' test.

Re: Google fires another AI researcher who reportedly challenged findings

#225

Post by BeetleB »

boppo1 wrote: post #223
BeetleB wrote: post #123

Earlier quoted context omitted.

> If you cannot give an ultimatum then you can always be pushed around. No finer example of false dichotomy than this. Hers is a textbook example of how not to negotiate. By textbook, I mean literally every textbook on negotiations that I've read. Ultimatums are a tool of last resort, and they all caution their high failure rate. It also is typically a signal of a weak person.

>literally every textbook on negotiations Aside: can you recommend the best ones? Getting to Yes is all I've read, but it seems a little 'fluffy'. I want something more rigorous, perhaps with game theory and models.

> I want something more rigorous, perhaps with game theory and models.

I haven't read any of those. The thing is, those are great for longer term negotiations, but not for short term day to day ones - you don't have the luxury of evaluating things from a game theory/model perspective. Even Getting To Yes is a bit poor in that regard.

Re: Google fires another AI researcher who reportedly challenged findings

#226

Post by istinetz »

godelski wrote: post #197
nostrademons wrote: post #194

Earlier quoted context omitted.

> it was our human bias of knowing Obama is black As a side note - Obama is multiracial (white mother). As a multiracial person who makes no effort to fit into American racial categories and whose general attitude toward them is "Your social construction of race is fucked up, man", it astounds me that people would expect an AI to see what's not actually there . Biologically, Barack Obama is half-black and half-white,…

Well the goal is that the reconstruction would be "him". Obviously the reconstruction isn't Obama. Limiting the discussion to race isn't a productive means to resolve the errors. There are more than race and they are likely coupled.

>Obviously the reconstruction isn't Obama.

okay, so the model is not a perfect oracle? What's your point, then, if it's not about race?

Re: Google fires another AI researcher who reportedly challenged findings

#227

Post by sijnapq »

yablak wrote: post #118

The TF-Agents team replicated the RL training (with the corresponding teams' very deep collaboration) and open-sourced it here: https://github.com/google-research/circuit_training It pretty much gets the same results as found in the Nature paper. The original codebase was heavily research-focused, used TF1, was impossible to run distributed training outside of Google's infra, and made it hard to try algorithms other…

If you seek truth, search on moma or ask for experiments for nature rebuttal document. Read the document and come back for comments :)

Re: Google fires another AI researcher who reportedly challenged findings

#228

Post by Y-bar »

tmp_anon_22 wrote: post #70
Y-bar wrote: post #60

Earlier quoted context omitted.

Maybe they thought Google was a mostly okay company with good potential and were under the impression that what they were hired for was to help them improve? There was after all a time when many of us really believed Google was a good kind of company and that it generally adhered to the "don't be evil" ethos. I can see myself falling into a similar trap of wishful thinking, especially if corporate communication reinf…

> Maybe they thought Pretty sure the thoughts that existed before they viewed their offer letter were different from the thoughts they had afterwards. And there's nothing wrong with it unless you for some reason want to ignore the biggest bias in the room.

What did the offer letter contain?

Re: Google fires another AI researcher who reportedly challenged findings

#229

Post by dekhn »

godelski wrote: post #197
nostrademons wrote: post #194

Earlier quoted context omitted.

> it was our human bias of knowing Obama is black As a side note - Obama is multiracial (white mother). As a multiracial person who makes no effort to fit into American racial categories and whose general attitude toward them is "Your social construction of race is fucked up, man", it astounds me that people would expect an AI to see what's not actually there . Biologically, Barack Obama is half-black and half-white,…

Well the goal is that the reconstruction would be "him". Obviously the reconstruction isn't Obama. Limiting the discussion to race isn't a productive means to resolve the errors. There are more than race and they are likely coupled.

actually, if you ahve a blurred face of obama, there are many reconstructions that are consistent with that data. His skin is pale, his features are selected from several genetic histories... the whole complaint about that example was overblown.
Post reply on HN