More than 1,200 Google workers condemn firing of AI scientist Timnit Gebru | Hacker News Reader

More than 1,200 Google workers condemn firing of AI scientist Timnit Gebru

theguardian.com

129 pointssoylentbeige237 comments

She gave her employer an ultimatum, threatening to quit if she didn't get her way. They called her bluff. I don't see how that is the same as being fired.

At that point she was on her way out, one way or another; most employers would rather not work with people like this.

I think the subtle, yet important, difference is that she said she would quit by a certain date and they terminated her immediately.

Given her role is likely just for PR it seems rather foolish not to at least give her two weeks notice or something.

This isn't at all unusual when someone is quitting on bad terms, is it?

In her group email she also asked other employees to stop working in any other related project.

I mean, I don't see how google could NOT terminate her after such an event. I am in a nation with extremly strong worker protection law but that action alone would be enough to fire you on the spot.

> but that action alone would be enough to fire you on the spot.

They are not as strong as you think then.

In countries where there are strong work protections law striking is a right, you don't get paid for the time you don't show up at work but you can obviously convince an entire department or an entire company workforce to go on strike, without consequences.

Without it the protection is not strong, it means an entire company against the individual worker, which is obviously unfair (regardless of what you think about work protections law, Google against a single human being it's an unfair battle, that's why "strong protection" means that the worker deson't have to fight alone and can call other workers to join the fight).

As I said in another sibling post, you cannot strike that way. You have to follow procedures to organize a legit strike.

If I wrote an email like that not only my union would not protect me, they would most probably tell me the employer took the right decision because I put at risk other employees positions (if they followed my reccomandations to stop working just because I said it)

Reading the comments here I feel like HN lives in a completely parallel universe. The number of employees who have the ability to threaten to quit and not immediately get fired are very very few. Even at the executive level you can't just make ultimatums for your employment and expect that they won't consider just firing you. Even if you're the representative of a union speaking on behalf of thousands of employees you need make sure you have enough leverage before making demands like that.

This not some crazy conspiracy that she got fired. It would be noteworthy if someone who did what she did didn't get fired.

The reality is that the research organizations at big tech companies have always been a very different kind of labor relationship than typical rank-and-file business units. Many of the senior or otherwise high-profile researchers at FAANG companies would otherwise be professors or academics, in positions where there is essentially unbridled academic freedom and no expectation of separating personal convictions and principles from the academic environment.

Google lured many of these researchers away from academic roles, with the tacit promise that they would be able to continue their research - and at times political - endeavors relatively unencumbered. Google has built a research empire on this promise, which has been responsible in part for their continued dominance in various markets.

My point here is that these researchers have a lot of leverage. Advocating for action within an organization, counter to the current immediate goals of management, is only taboo (and thus “fireable”) in most companies because workers have almost no leverage, generally speaking. Not because it’s some inherently “bad” thing. If Google becomes seen as a place that fires researchers for advocating for their beliefs, especially in a capacity that they’ve been explicitly hired to do, then I think a lot of researchers would - and rightly should - reconsider whether such an environment is consistent with their values as academics and, often times, activists. Google’s research group is enormously valuable to the company, and there’s only so much reputational damage of this nature the organization can sustain before academics decide to take their talents elsewhere.

> with extremly strong worker protection law

Serious question: what makes them strong if employers can't organize a simple strike?

You can organize a strike BUT this is absolutely not the way you do it.

You have to follow strict procedures, work with your union and go into one or more formal negotiating tables with all the representatives of all the interested parts before even thinking about giving an ultimatum like that. And for sure you cannot communicate in that way using directly your work email

But can you or can't you be fired for that?

Because in my country you can't be fired for using the work email for that.

Communicating with other workers using their work email is absolutely allowed, courts ruled about it several times [1].

I usually receive trade unions (there are several of them) communications on my work email from the unions' work email, because they have been authorized by the company to send them.

In this case she wrote directly to workers using work email, but having no union or labor protection laws it doesn't make any difference whether she could or couldn't, she could be fired anyway without having to provide any reason.

In countries where there are strong laws protecting workers she would have written to the union members and they would have done the same thing: write directly to the workers.

Of course she did it at Google so it is different, but here the strict procedures to call a strike are only necessary if a public service that requires continuity risk to be interrupted, otherwise unions are only required to alert the company that the strike is going to happen, but have no requirement whatsoever on how to organise it.

Which sound logical to me,strong protection means IMO freedom to collectively counter the actions of the company, if that's not allowed the protection is not strong.

[1] Court of Catania, Labor Section, February 2, 2009 "The RSU employee can send trade union communications by e-mail to the employees of the company during their working hours and to their company e-mail address using his personal e-mail address"

In many countries you do need a vote for a strike to be organised. You can't just spam people with a request to stop working. Strikes are a powerful tactic that have evolved a lot of formality around them to try and ensure the outcomes aren't totally destructive. Italy is a rather unusual exception to this. Perhaps it's a contributing factor to the long stagnation of the Italian economy.

> Perhaps it's a contributing factor to the long stagnation of the Italian economy

Perhaps.

But doesn't explain stagnation in Japanese economy where there are no Italian unions.

Or why France did much better despite having even stronger work protections laws than Italy and wilder strikes (the gilet jaunes for example) or those happened last year against the pension reform where the workers of public transportation went on strike - for weeks - without even announcing it.

Even countries like Germany, Singapore, Switzerland and Finland are doing worse than Pakistan in the GDP growth race.

My point was than if there are stronger work protection laws somewhere else and the laws of your country are weaker, they are not very strong, they are moderately strong.

Re: last point. Fair enough. That point is sound.

Japan seems to have evolved a work culture very similar to strongly unionised societies but without unions. Japanese salaryman culture is famously a culture of employment for life with unusually strong loyalty between employee and employer, hence weird things like "banishment rooms" that you don't find elsewhere. If it's the end results that matter and not the means, Japan might not be a good counter example.

I think French strike law sounds tighter than Italian strike law? The French are famous for striking but strikes must still be a collective decision and related to a specific set of issues, whereas Italian strike law really does sound incredibly broad and vague.

With respect to Pakistan, that's doesn't mean anything, poor countries always have very high GDP growth. It's easy to grow something small and backwards by a lot because you can get a lot of relative growth just by copying what other countries do, and less absolute improvement is needed to get a percentage point of growth to begin with. You can only compare GDP growth rates between countries of a similar level of wealth.

What's the point of the two week notice in this case?

as I understand it, she was on vacation at the time

In California you get paid for accrued vacation when you are terminated. She's not out any money, so why keep her on as an employee while she's out?

I think that would have been the civilized thing to do. But that is the exception, not the rule in the bay; I.e once you have made your intentions known, and are perceived as someone who could cause damage in the interim, you are out straight away. Feels like a natural consequence of at-will employment.

You can't hand in your resignation effective at a future date. That's not how it works. If you resign, you resign. It's then up to the employer whether or not to hold you to whatever notice period is set out in your contract. Many places will not do so and release you effective immediately.

Of course, if you leave on good terms, your superiors might look at you kindly and work with you to meet not just their needs, but also your needs. That's why it pays to be humble and generous in your interactions with other people.

This is par for the course at most companies with IP to protect. If you quit they just walk you out.

In the US, in countries with more worker protections it is not. (unless they pay you for that 2-week, month, 3-month period when you aren't working for them, but they can't remove you from worker list).

Yes this is correct. The primary difference in at will employment is that I also am not required to give my employer notice, whereas in the countries you mentioned I am.

That's not true.

You can as an employer walk away whenever you want, you just don't get the salary for the difference between the notice and when you left.

Example: if you have to give 30 days notice (very common in Italy) and leave on the spot, after your resignation have been accepted, you have to give the company the equivalent of 30 days of pay.

Which is exactly the money they should give you aniway if you stayed for the entire notice time.

But usually the employer and company find an agreement before coming to that.

If employee is found using company resources inappropriately (in this case, mailing list to rant) then you can't give that employee 2 weeks notice. Google will then have to consider possibility that she will abuse their mailing lists.

Google has internal mailing lists dedicated solely to rants (and related discussions) and others on which rants are certainly well accepted.

Google exercises a subjective standard to determine whether or not an employee is using mailing lists inappropriately (e.g. James Damore).

Source: ex-google

So if you don’t do what your boss says at a job, even if you disagree, your boss can fire you, and this is good. But if you tell your boss that they should do something or you’ll quit, this is bad and makes you unemployable.

You realize that this is just an argument that whoever is in a position of authority over you is right by virtue of having that authority?

It's called "insubordination".

Is insubordination to have a different opinion, and then offer paths to agreement between professionals?

Is insubordination to represent your responsibility area, as Ethical AI Researcher?

Is insubordination to send an email in order to establish dialogue around matters of disagreement between professionals?

Good point. I think its important to leave a link to the paper here for HN. Please read it or at least its intro & conclusion.

https://gofile.io/d/WfcxoF

After reading, it seems very much in line with her role as an ethicist. She merely states that there are various risks associated with processing large datasets.

Its a bit sad that they attempted to censure such an innocuous paper that without malice states “there are risks to consider.”

> She merely states that there are various risks associated with processing large datasets.

What's the strongest argument against her "merely" doing this?

We haven't heard any arguments, only that management can do as they please when they don't like results of AI Ethics Research. We'll likely not get anything substantial, ever, either.

As I've stated in other post, this is very, very troublesome for AI and Google.

OTOH, Google has much emphasis on ethics in AI/ML otherwise. But the whole company, especially management, need to put actions in alignment with their words on this matter.

What is the strongest argument that you can come up with? Not that's been presented.

My strongest argument is AGAINST managers at Google getting to micromanage, censor and to revenge-fire a researcher, just because they personally disagree with their published AI Ethics Research. A job said researcher was tasked with and leading.

Such behaviour, unless corrected, undermines Google's entire AI/ML and cloud operations.

That's not what I asked. If you can't argue the counter of your own position in the most honest way possible then either you're not prepared or not honest.

It's called steel manning.

The best rebuttal so far has been to decline to comment.

However, without revealing names the paper outlines internal company matters, presenting opinions and tautology that do not fit proper format of professionally published research. Instead of reworking a flawed paper, the researcher escalated and exposed individual opinion, publicly. Such company representation is out of scope for research positions. As Google could not agree to terms of email, it was taken as resignation notice.

Let's try applying this, shall we:

I believe we should have no curruption in Russia. By your logic, unless I can give a strong argument in favour of corruption, I am not prepared or not honest?

Are you prepared to give a rigorous argument in favour of: terrorist attacks, slavery, eugenics, facism, child labour, sexism

Bribes functions as oil in the machinery of society, and empowers strong men against rampant decadence. So is working for good when it is sanctioned and goes to the right people and their agendas, comprising of: terrorist attacks, slavery, eugenics, facism, child labour, and of course sexism.

This is what I said.

> If you can't argue the counter of your own position in the most honest way possible then either you're not prepared or not honest.

I didn't say that if your position is on the opposite side of child abuse and you don't want to change your position that you aren't honest.

Yes, I'm prepared to try and make the strongest argument possible for things I may not fully understand, until I do. Whatever those things make me feel.

Sometimes the strongest arguments have large flaws (mostly moral with you examples) but sometimes we just don't understand the opposing arguments and then just create strawmen to rage against.

I'm opposed to this type of irrational emotional behaviour. It causes things like mob violence which I'm very familiar with.

So when someone can't articulate the opposing views then they need to be warned or directed to do so. If they refuse and run back to a strawmen then yes they are not honest and obviously not prepared.

uMkhonto we Sizwe was a terrorist organization. But they had a very strong argument for why terrorism was an appropriate action.

I don't say that I agree with their use of violence, but they had an argument that makes blanket statements like "terrorism is evil" harder to make.

"No artificial neural network is near a point where we can talk about it having moral responsibility separate from its trainers’ and deployers’ – but we can make it sound like it does, and exonerate them, if we call it AI." https://twitter.com/shashashasha/status/1335067153402355714

Modern ethical AI / ML, is both that and beyond that. Issues range vastly in how complex algorithms can be used, misused and misinterpreted. There are also issues of uncovering hidden biases, unfairness, manipulative effects, privacy issues, risk and impact analysis, explanatory AI, who is accountable for accidents and malicious acts, etc. The list goes on covering both the legal and the political. This may not be unique to ML / AI, but is harder to quantify and prove, with more prevalent usage of complex algorithms.

https://analyticsindiamag.com/how-is-ethical-ai-different-fr...

https://www.kdnuggets.com/2016/03/ethics-machine-learning-ta...

The strongest argument for Google is it not wanting people to know that there are risks, no?

The strongest argument for her not doing it is keeping her job vs maintaining her professional responsibility as an ethicist, no?

That's not a strong argument. Can you come up with anything better?

Do Google employers work in barracks?

Without commenting on the politics of this particular case, I think what you describe is built into the definition of "boss".

Not universally though. In countries with stronger labor protections and unions, you can’t just fire people for philosophical disagreements (not without those disputes being mediated anyway).

I don’t understand how we arrived at some conception of workplaces as little dictatorships, where bosses can treat people however they want and that’s somehow good or legitimate.

A boss controls the work that you, via a contract, agreed to do.

If you want out of the contract the boss cannot make you do the work. You have agency, they cannot threaten you with violence as a dictator could.

Managers are supposed to exercise power on behalf of the company. In this case, it’s unclear whether the manager was exercising power in the best interests of the company or of himself.

In my opinion, might doesn’t make right. Managers aren’t de facto right purely by virtue of being a manager.

Really depends on your definition of right. Do you mean legally? Morally? Logically?

When it comes to who has the ability to steer subordinates, managers are always right. Except if their managers disagree...

So just blanket appeals to authoritarianism then. Completely rational.

A company is not a government... Yes, someone who owns a company has the legal right to do whatever they want, as long as it's not criminal, with their property.

They have complete authority... They don't have to have someone else endow it to them or give it to them...

So yes, they are by definition authoritarian. That's one of the cornerstones of being able to own private property.

Other people’s labor is not a company’s private property. That was known as slavery and we have mostly abolished it in the developed world.

But my point is that all of your arguments rest on an unquestioned assumption about how ownership translates to authoritarian control of others without explaining how or by what justification that power is arrived at.

The answer of course is that it’s a social construct enforced by states to varying degrees, present and historically, and thus no way inherent to the thing itself. Appeals to status quo are therefore not justifications for the validity of the relationship, as it has changed over time, proving malleable to all kinds of political and social forces (e.g. the abolition of slavery, the formation of unions, worker collectives). Defending an authoritarian relation to property as it relates to other people’s labor thus require an actual and explicit defense, which you have so far not provided.

Citizen of country with some of the strongest worker protections here: not over philosophical differences.

Very much so for refusing to do work assigned to you.

And it doesn't mean "bosses can treat people however they want". It means "you are employed to do a job, if you refuse to do that job, you are no longer employed"

As much as I support employment rights, you are paid by a company to do their bidding. If you don't like it start a company, but then you're just doing the bidding of the customer.

This is a uniquely American notion of employment. “Do my bidding or be done away with” is not the norm across countries with strong unions and worker protections.

No, this is not uniquely American. Which country are you from that doesn't require employees to act in a way as directed by the company that employs them?

Literally any country with strong laws protecting strikes and work stoppages? In many parts of Europe, I can convince my coworkers that our bosses are treating us unfairly, unsafely, etc, we can walk off the job and not get fired. I doubt any employer would want us to do that, so it would be entirely against their “direction.”

"Do my bidding" is different from "do my bidding being paid way below market rate" or "do my bidding and I don't care if that involves a high likelihood of losing a limb".

Also, "do my bidding" should be written as "do your job."

I'm from a country with strong labor protections (striking, 3 strike dismissal etc.), and you're moving the goalposts.

Just to clarify, in many parts of Europe you cannot do that, you would need to be a recognised union and in many cases at least hold a vote of some sort, there would probably also be requirements around negotiations. You would also be restricted in what you do to people who refuse to join your strike.

The UK is an example of a European country with relatively tough union laws. The list of issues that it's legitimate to strike over is restricted, votes are required, the general secretary of the union must authorise the strike and so on. Strikes do happen of course but not because some random middle manager threw a hissy fit about feedback on a paper she wrote.

Giving ultimatums and threatening to quit generally isn't a good way to remain employed with any company.

> So if you don’t do what your boss says at a job, even if you disagree, your boss can fire you, and this is good.

Gonna take a quick guess you're still in undergrad.

Yes, this is literally how employment works. Your future boss's job will be to use your labor efficiently for the company. You may disagree with your boss on an assignment, in which case it would be wise to schedule some time and professionally discuss your concerns. Your boss may be persuaded, and good bosses will try their best, within reason, to keep their subordinates content with the work they're doing.

However, at the end of the day, the boss may disagree with your points and still require you to do the job as assigned. If you refuse, you will likely be fired. Note that you also have agency here: If you don't like the work the boss makes you do, and you feel like your concerns are consistently falling on deaf ears, you can quit.

Just because that is how the world works, doesn’t mean it has to be a that way (or should be that way).

It doesn't have to be no, but y s it should...

What is your proposal?

Do we have the email with the "ultimatum" in it? I think there's a world of difference between

"Do these things or I'm gone and taking my team with me"

and

"These things are so fundamental to my role as an ethical AI researcher that it's difficult to see myself doing this job without them"

IANAL, but I suspect a court would also interpret them differently.

And even so, Google likes to position itself as a nice employer, which listens to their employees' concerns. De-escalating the situation would have been nice - immediately terminating the concerned employee makes Google seem like another giant faceless emotionless multinational corporation (which it may well be).

Initially I took Google's side given Gebru has history of bullying so aggressively and has no qualms to play race/gender cards at every opportunity she gets. But now that more details (including the paper in question) has been revealed, it's hard to stay on Google's side. It is clear that the peer review process that Jeff Dean had described has been optional for all intent and purpose. It says along similar lines that right there on their internal website. Too many Googlers have attested that they have bypassed the process with no consequences. Gebru's paper itself has no real technical novelty. There was no reason to fire her. It doesn't seem out of line that Gebru was befuddled, frustrated and made a threat on the basis that she had right for the information. Jeff Dean could have easily offered her feedback and transparency in decision process. He could have agreed to let her make corrections later in the submission process. He did none of that and doesn't want to come out clean.

Having said that I still feel Gebru is extra-ordinarily self-entitled person. Few months ago she insisted that Jeff Dean take her side in Twitter feud she initiated. She has the mindset of either you are with me all the time or you are racist+misogynist. She complains about "micro agressions" towards her from everyone but she herself is super aggressive. When disagreement arise, she marks you down as enemy and launches attacks to destroy you completely with sole purpose of making you homeless, jobless and making sure your kids die of hunger. I wouldn't wish upon her as colleague to anyone. You might have right to disagree with me but this is what I have felt. I still take her side in this current episode nonetheless because I think she is on the right side.

Ultimately we are all playing opinion game here and only few people knows what really happened.

I think that's a balanced view. The fact is, Google might have constructively dismissed her because of her history and attitude and they knew if they outright fired her, it would be a PR nightmare. Letting her resign like this makes it much more complex and doesn't open them up to any serious threat of litigation.

She can try to sue them, but now that she made it public it's quite unlikely any court would look at it.

All in all, it's obvious Google didn't follow their own supposed research protocols, but this whole thing is much more complex and I think it has very little to do with her paper.

> All in all, it's obvious Google didn't follow their own supposed research protocols, but this whole thing is much more complex and I think it has very little to do with her paper.

She still had a job after she published the paper, she still had a job after she told people to stop working, she stopped having a job after she demanded confidental HR records of people who feared giving her rather simple feedback.

Citation: https://old.reddit.com/r/MachineLearning/comments/k77sxz/d_t...

Sample comment:

Another ex-colleague here. I was not going to participate in the discussions but your post made me realize objective truth should come out. I do believe she actually thinks she is making the world a better place but in reality any interaction with her has been incredibly stressful having to carefully weigh every move made in her presence. When this blows over her departure will be a net positive for the morale of the company.

To give a concrete example of what it is like to work with her I will describe something that has not come to light until now. When GPT-3 came out a discussion thread was started in the brain papers group. Timnit was one of the first to respond with some of her thoughts. Almost immediately a very high profile figure has also also responded with his thoughts. He is not Lecun or Dean but he is close. What followed for the rest of the thread was Timnit blasting privileged white men for ignoring the voice of a black woman. Nevermind that it was painfully clear they were writing their responses at the same time. Message after message she would blast both the high profile figure and anyone who so much as implied it could have been a misunderstanding. In the end everyone just bent over backwards apologizing to her and the thread was abandoned along with the whole brain papers group which was relatively active up to that point. She has effectively robbed thousands of colleagues of insights into their seniors thought process just because she didn't immediately get attention.

The thread is still up there so any googler can see it for themselves and verify I am telling the truth.

What I do know is that you don't hire a PR team to make you look bad. You hire them to make you look good.

You don't hire an ethicist to make you look bad. You hire them to make you look good at what you do: diversity in AI. Gebru failed at her job.

I think you aren't necessarily wrong that internal committees like this are about PR. It is simple industry self-regulation. Expectations should be adjusted accordingly.

But that would also mean Google exploited her for that.

I haven't seen much by myself, but I witnessed that flame with Y.LeCunn and that alone gave a lot of information.

I can't understand WHY so many people obviously think that it is totally fine to be behave utterly unethical, especially when a person whose behaviour is in question is "professional ethics researcher". Looks like they think that a big goal justifies the means. Welcome to Russia circa 1917, or Germany of 193x.

Insanity((

It's really sad watching people bend over backwards trying to give this person a break.

You admit she's a bully, super aggressive, constantly makes spurious accusations of racism, attempts to destroy her enemies totally and is in general a nightmare employee in every way. Then you say you take her side because Dean could have given her "transparency"? Are you aware she was demanding the identities of people who had criticised her work, presumably so she could then go and bully them too?

This person is a 100% A+ straightforward case of someone who should be fired and in fact should have been fired long ago. All those attributes are the sort of thing that absolutely qualify someone for dismissal. If you were feeling trollish, you might even say it's black and white.

Agree 100%. We should not be protecting bullies regardless of race, gender, position in the company whatever. If you don't treat people well you are out.

I didn't even know that she had a history of bullying before I concluded she was in the wrong. My incomplete understanding is she submitted a paper, google said it was not submitted in time and they made her retract it, and in response she demanded the names of those who reviewed the paper or else she would resign. That is completely unacceptable behavior and they were right to accept her resignation. The fact that she has a history of bullying makes it even worse for her. Of course they aren't going to give her the names of those people so she can bully them too.

> "would rather not work with people like this."

Her clout/skills/PR and her attitude are a package deal.

Google wants to use her for PR, but expects her to be a lapdog.

I conflicted about many things, but respect her for having backbone.

She said "at a later date" leaving it up to her discretion. Google accepted a resignation that didn't exist, and will probably be liable for it in some way.

A draft copy of the disputed paper was posted over in Reddit, if this is of interest to anyone:

https://old.reddit.com/r/MachineLearning/comments/k77sxz/d_t...

https://gofile.io/d/WfcxoF

Whether or not I agree with the conclusions, that’s exactly the sort of paper one might expect to be written by someone whose job title includes the words “ethics” and “AI”. I’m not sure what Google expected.

The other curious thing is, behind the pink-boxed redactions are the names of four other Google employees. I'm wondering why they haven't been reprimanded in the same way. At least, not that we know of.

Well, she wasn't reprimanded for writing the paper nor for publishing it. In fact, she wasn't even reprimanded as far as I know for telling other employees to stop working on DEI. She told them she wanted confidental information or she will quit on a date she would like. They said they would not give her that information and that she is no longer welcome at the company. This isn't really reprimanding or even a normal firing. If I tell my boss do X or I quit and they say "Thanks for your work, we accept your resigination and the legal notice period applies." Fair play.

> She told them she wanted confidental information

I understand this is only because upper management had chosen to anonymize their review of the paper, which is apparently quite unusual in Google.

What stood out to you?

The whole paper doesn't go beyond surface-level analysis. The main point is that language models reflect the same biases as its training data. Very little of it is technical. Most of it is about examples of how language models aren't as politically correct as the authors want it to be (e.g. calling female doctors "female doctors" when it's not relevant to the context). It's a legitimate concern, but anyone who remembered Microsoft Tay would know that.

the paper was not the problem, her attitude & reaction to peer review was the problem - giving ultimatums, demanding to disclose names of the reviewers, and broadcasting emails to entire team calling for sabotage.

Her behavior fits the definition of 'toxic employee', it was unprofessional by any standard.

Pink squares used to hide the authors, but the names are still written behind in the PDF...

I'd been looking for that. Thanks for posting.

She wasn't fired over the paper. She was fired because she demanded Google name the employees who gave (anonymous) negative feedback about the paper, and that anyone going forward who gave feedback on her work also be named, or she'd quit.

> As shown in the amount of compute used to train deep learning models has increased 300,000x in 6 years, increasing at a far higher pace than Moore’s Law which posits that the amount of computation that can be done per unit area would roughly double every two years. This means that power consumption per unit area is not staying constant as implied by Moore’s Law [108].

IIRC Moore's law doesn't make any predictions about power consumption at all, let alone ones meriting being cited in a scientific paper.

yeah, I've debunked this particular line as being technically nonsense. If I were a reviewer of the paper, reaching a sentence like that, I'd return to the editor with "revise and resubmit- this doesn't meet the basic technical level of expertise we require for publication."

if you build 2X the compute capacity with current nodes, you didn't violate moore's law.

According to a seemingly credible account from a former employee, Timnit terrorized a Google Brain manuscript discussion forum with unfounded accusations of racist disregard for her contributions https://www.reddit.com/r/MachineLearning/comments/k77sxz/com...

The thread is filled with agreeing comments from many people who seem to be ex-Google and who had access to the relevant thread.

This is crazy, I'm not sure why people aren't taking this more into account. By all accounts Timnit was an extremely toxic person to work with. Sure, perhaps she shouldn't have been fired for this paper, Google probably saw an opportunity to remove her that wouldn't be a PR nightmare for them. But regardless, it's hard for me personally to see them as being in the wrong given Timnit's history.

Gebru's brand of Twitter woke-scoldery is taken for granted these days. Par for the course. We all know the tired rhetoric that will be trotted out to defend it, so much so that we could write it ourselves.

A member of the corporate elite criticizing, let alone firing, a black woman for race baiting is now beyond the pale. So they got rid of her indirectly, by making her working conditions unappealing enough to induce her to quit. Too bad some Director Karen of Whatever lost her nerve and ended it on her terms rather than Gebru's. No finesse!

I find the semi-regular condemnations from employees of big tech companies hollow. If you work for one of these companies and find their behavior unsavory, go work somewhere else - this is actually what might cause executives at these companies to reflect on how they might change course.

The fact that most of these employees could easily get well paying (just not as _ridiculously well paying_) jobs elsewhere to me feels like they are not staying out of necessity, rather they are looking to assuage their conscious through statements like these, and continue collecting their exorbitant pay.

The Facebook “walkout” was a particularly hilarious example of this when you consider the second step of the walkout: walking back in.

If you look at trade unions, you see that walkouts, work slow downs and temporary strikes are commonly used.

The goal is never to destroy the company but to simply force them to the bargaining table by reminding them of the economic power workers provide.

A total strike or worse a union quitting a company all together is a nuclear option.

It may work with physical work, but at FANG companies software engineers will just make up for it on the weekends at nights. Stack ranking works wonderfully at making engineers to compete in working as much as they can (which is not a bad thing, just how the system works).

All that stuff sounds like systematic abuses that a union would reduce.

The problem of the effectiveness of Facebook employee walkouts stems from the fact that they aren’t a real union, so the employer can just force them to make it up without providing for any additional compensation.

Any protest is a null option until the whole thing catalyzes and they form a union to address the repeated disregard the company shows them.

The only way for the company to avoid this is to self regulate or at least pay lip service to buy more time.

,,All that stuff sounds like systematic abuses that a union would reduce.''

It doesn't always feel like it when you get an above average salary. Also it really depends on the team, how much the tech lead and the manager can push back to external requests...I don't see how adding even more politics can solve political issues. The fact that it's super easy to move between teams is much more important than anything the union could achieve.

Stack ranking causes political backstabbers to rise to the top and filters out the conscientious. It probably functions well enough in companies that are fundamentally extractive, but it will put a company that is supposed to innovate on a self destructive spiral.

Microsoft improved a lot ever since it dumped stack ranking.

Microsoft was using per team stack ranking, Google global stack ranking. That's a huge difference I think. The problem is that the managers go to the promotion commitees and started exchanging promotions with eachother, so the system that was working before doesn't work anymore.

This wasn’t a labor dispute about being treated fairly but a position employees were taking on principles. The truly principled walked out and didn’t come back.

It’s just another example of the “participation culture” that modern parenting and social media have made commonplace. Who needs to do anything real when you can just upvote or retweet? You get the same sort of participation trophy that you’ve been taught to aim for since childhood.

>"that modern parenting"

Back in my days we would chop off heads and assisinate politicans! These kids can't do anything right!

> Back in my days we would chop off heads and assisinate politicans!

back in my days, we would unionize and block the factories until the company died or we got what we wanted

The peak of this behavior led to a small war between coal miners and a sheriff and a small airforce that bombed the strikers that only ended when Warren Harding threatened to send in the military.

They ultimately lost but what happened there led to the sympathetic growth of the AFL and CIO and ultimately to the New Deal and the rise of the middle class.

They say history doesn't repeat but it rhymes. I wouldn't be surprised if the next series of pitched struggles between American owners and labor result in extreme violence also, before giving way to a system that is slightly more egalitarian.

> Back in my days we would chop off heads

Ironic that guillotine memes are so popular among the kids these days, no?

The 70s are back in fashion.

This is such a bad-faith, shallow take. Ascribing the act of bringing awareness to “participation trophies” is nonsense.

I also disagree with the parent. Bringing change is almost always easiest from within. “Don’t like it? Leave.” stances aim to do more to attack the person involved than it does to take a critical look at whatever problem it is that’s being discussed.

It's a moral problem. If the organization you work for is immoral, by knowingly continuing to work for them you are materially cooperating in the evil they are doing and are morally culpable yourself. Sure, you can refuse to materially cooperate, but then they will fire you anyhow, as we see here.

You don't need to leave, you can strike for example. "Condemning" does nothing.

It is not "don't like it? leave.", it is "stop working for immoral organizations". Would you consider a guard of a concentration camp ethical if they "condemned" the actions of the government but continued to support them via their work? I wouldn't.

The fact that they're working at Google in the first place shows they're not that ethical to begin with.

But nearly everyone has limits, and the petition signing limit was reached for these 1200 Google employees.

We'll see if the quitting limit is ever reached for any of them.

What’s the opposite of “participation culture”? Not participating?

I feel like signing a letter of protest is doing something “real”. It’s communicating to leadership of the company that you have concerns about what has transpired. It’s not nothing.

Ayn Rand's following, probably.

Is there an equivalent to Godwin's Law for Ayn Rand?

Parent post is probably referring to the relatively new phenomena of issuing "participation prizes" instead of prizes for winning, this insulating the participants who "lost" from the stigma associated with being a "loser".

Or they could just actually unionize instead of this reactive outrage. It’s unlikely Gebru would have been fired under these circumstances if she had a union contract.

As a manager, Gebru wouldn't have been allowed into a union.

Is she really a manager? In factories the foreman was considered part of the Union rank and file. PMs and other developer leadership could be integrated into a union.

Basically if you can be promoted up into a position it can be a union job. If you need a business degree to do it though then it’s non union.

I don't know exactly what her job responsibilities were, but it's not the case as a rule that any position you can be promoted to is a union job. Anyone who has independent authority to "hire, transfer, suspend, lay off, recall, promote, discharge, assign, reward, or discipline other employees, or responsibly to direct them, or to adjust their grievances, or effectively to recommend such action" is by law a supervisor, and they can't join an NLRA-protected union because supervisors aren't protected by the NLRA. As far as I know, all Google managers have the power to assign, reward, and discipline their reports.

If they can only do it for specific purposes and their actions are reviewed then it’s possible they could gain union memberships.

For example foremen hire, train, supervise and evaluate workers. Yet they have a union.

I can’t say for certain if it’s a protected union but it is a large and existent one in the US.

> Is she really a manager?

Yes.

> In factories the foreman was considered part of the Union rank and file.

She's not a foreman.

> Basically if you can be promoted up into a position it can be a union job.

That's not how the NLRA works; in plenty of workplaces people can, even if they rarely do, promote without limit and most management positions don't require business degrees, but that doesn't make all those positions NLRA-protected.

I hope this does encourage more workers in this industry to unionize.

It's unfortunate that people aren't sufficiently concerned that their employer holds a significant amount of power over them - or at least, not enough that they'll do anything much about it - even if they do happen to have a job that pays higher than average.

Unions can be valuable resources for all workers.

> Or they could just actually unionize instead of this reactive outrage.

Some of the employees involved are in management (as was Gebru), so they don't have that right, nor would that protect people in positions similar to Gebru's.

If their goal is to promote savory behavior, isn't it beneficial to have an internal voice trying to guide the company's direction?

There is no shortage of engineers who would happily take a job at Google. If the folks raising concerns decided to all leave, isn't the net result that Google would operate much as it does today (maybe a little bit slower to get products to market) but without any internal check on its behavior?

And since the issue under dispute here is AI ethics, shouldn't we the non-Googlers want to make sure people willing to voice concern about unethical behavior remain at Google?

If they were serious about the strength of collective action, they could unionize. That due process would have protected a fired employee more than any petition.

"shouldn't we the non-Googlers want to make sure people willing to voice concern about unethical behavior remain at Google?"

That depends on whether they're actually effective. If their concerns are ignored and they're not able to effect change, then their presence at the company is useless as far as moving Google in an ethical direction is concerned.

> they are looking to assuage their conscious through statements like these

Statements like these still put pressure on Google. It's possible to care enough to say something, without caring enough to go beyond talk, but that doesn't necessarily mean it's all performative.

How does it put pressure on Google? Where's the leverage?

Because it threatens Google's PR.

They're often meant to be somewhat hollow. The referenced letter (https://googlewalkout.medium.com/standing-with-dr-timnit-geb...) uses dramatic language to issue pretty anodyne demands: Brain leadership should meet with her former team and explain what went wrong, Google should publicly explain what it was they found objectionable in her paper, and Google Research should confirm they're committed to research integrity and academic freedom. Someone could easily sign this letter without thinking Google's a bad company at all.

The public sphere is very heavily slanted. Few people dare come out under their own name and condemn the bullying behavior of Timnit, for fear of being subjected to that very bullying, which has drastic real life consequences. The letter demands to push the conversation into the public sphere, where one sided Twitter and media pressure can be fully applied. One has to be very naive and sheltered to sign such a letter.

The protests against dragonfly and maven seemed to have an effect, which is the opposite of hollow?

Big Corps have always been taken that stance. Google managers must be saying the same thing up and down the food chain.

But the fact that it hasn't stopped walkouts is interesting. It means there are influential people within who back the walkouts.

Or more likely, the corps realize that the walkouts are an ineffective form of virtue signaling and it would be a PR nightmare to do anything about them, so might as well let them happen since they really mean very little.

Walkouts are close enough to strikes that it would be pretty dangerous for a Google executive to try and stop them. (I think some of them may actually be strikes by law - I don't know the exact legal definition.)

If your job is an ethicist, your company hires you to demonstrate that they are ethical.

If they truly aren't then you either have to sacrifice your integrity or kick up a fuss. Additionally once her integrity is lost she loses market value.

Her market (whitewashing/signaling of ethical behavior of a large company) is an odd one but she behaved it the most rational manner given the circumstances.

Moreover by kicking up a fuss she probably has sacrificed an non disparagement termination bonus. Real $$$. So she's probably not talking shit.

I agree with all of this, except you miss that she started playing those ultimatum games with her boss, this particular act, let’s be real, is something that gets you fired from most places. There would be a lot of merit to what you’re saying and a lot more sympathy for this woman if she hadn’t made these missteps.

It’s true that ultimatums usually get you fired.

However it’s also true that they can simply be a reflection of an otherwise unworkable situation.

In a regular employment situation, I think that if you get to the point of wanting to issue an ultimatums, it is just an indication that the company is not a good fit for you any longer. On an individual level it’s better to recognize the signal and leave without burning bridges.

However if you role is ‘ethicist’, it is fundamentally your responsibility not to turn a blind eye to or be complicit in unethical processes, so giving an ultimatum may have been the only option available to retain integrity.

I don’t think we can declare these to be ‘missteps’ given her role. Perhaps Google is simply not capable of being ethical, and is going to fire people who challenge that. What would not be a ‘misstep’ under these circumstances?

"she started playing those ultimatum games with her boss"

Is it a game or is it just a negotiation?

My comment was about the 1200 workers who condemned the firing.

There have been similar recent public statements such as MS workers about use of AI for the pentagon, FB employees on content, etc..

Simply I mean the best way to support a wrongfully ousted ethicist would be to leave yourself, rather than keep getting paid by the execs who wronged her.

The work of AI ethics - especially in the vein of Timnit Gebru - is essential. My statement, was one of support for her.

>Simply I mean the best way to support a wrongfully ousted ethicist would be to leave yourself

What are you basing this on? I can think of at least one example where this pressure actually worked (caused google management to reverse course). ICE contracts, I think?

Your strategy would seem to be more in Google management's favor - clearing out all the people who might kick up a fuss when they try to engage in unethical behavior.

A typical recruiting fee is 30 percent of an employees annual salary.

A typical onboarding process takes weeks or months of paid employee time.

Employees value to a company increases substantially over time due to knowledge of culture, process, etc.

Executives care a lot when company actions cause employees to leave. Human capital is one of the essential ingredients to a successful company.

These things are real money and execs care about the impact on bottom line.

This is typically the most important aspect of an exit interview. In my experience, the employee's reason for leaving is thoroughly examined.

Does "condemning an action" have the same impact?

It’s easy to recommend others to leave their jobs. Do you like absolutely all the aspects of the business you work for?

No, but I'm not at the point where I'm signing an open letter to condemn my employer for something. However if you are at that point, you may at least consider that you should be working elsewhere.

Why? If the workers can apply pressure and get their employer to change whatever objectionable behavior it's doing, why shouldn't they?

Why shouldn't workplaces be more democratic? We got rid of kings everywhere else, why should we keep them around just because they're signing our paychecks?

I'm in total agreement with you. My response was a criticism of those who sit idly by and collect their massive paycheck, and perhaps signing an open letter from time to time to sleep better at night. It's just paying lip service to the cause.

I absolutely hate my workplace. But I've hated most of the other workplaces I've been at. My ideal situation would be to work at NASA, but I've heard some questionable things about that workplace too. Point is, unless you run the show, you have little control over the culture and something will always suck because humans are individuals and not all of us like conforming. So, I just suck it up and work so I can live. Maybe one day I can do my own thing and establish the culture, but I have to be willing to sacrifice and do what it takes to build a successful business. My guess is that if I do that I might just end up becoming the very thing I hate. It's easy to be morally superior when I have nothing on the line.

>> go work somewhere else

That may play into the company's long term goals to winnow out the people not willing to tow the party line.

Not sure what the best option is, maybe for those that have accrued weeks of PTO they can all take it at the same time.

The problem with this view, though, is imagine if you're in that situation, and you leave, only to be replaced with someone with more amoral views. It's the opposite of a solution imo.

At the risk of invoking Godwin's law....this is exactly Rodolf Hoss' defense at the Nuremberg trails:

"Don’t you see, we SS men were not supposed to think about these things; ... somebody else would have done just as well if I hadn’t"

https://www.facinghistory.org/holocaust-and-human-behavior/c...

I would be extremely cauticous about using the argument of "well If I don't do it, someone else will".

But that implies that people protesting are doing bad things themselves, as opposed to just working at the same company.

I really hope someone at Google's PR department is paying attention to comparisons like these, and that they realize there are consequences to allowing Google's brand to get dragged through the mud by their relentless antisocial behavior.

Google went from being a tech darling where their "Don't be evil" motto was taken seriously to "Google employees are like Nazis".

Hope someone in authority at Google is doing some serious soul searching as to why this might be.

But notice there is a difference between following orders and staying in a role but using what power you have to enact change. There is a third route.

So if the company does something unsavory, just walk away? Not even trying to change it first?

If you don’t like a policy, go live somewhere else?

If you find mold in a fruit, go shopping in a different place?

I mean, it’s a valid strategy. But obviously not the only valid one.

It’s not wrong to like the other aspects of the job and try to change the quality that they disagree with.

Why not try this first though? You complain, and if complaints aren't met you then quit. It's kind of childish, in a way, to not voice problems you have until you just up and quit one day. It's much more professional to at least try and communicate.

If you just saw someone get terminated at your employer, and you considered it unfair, wouldn't you fear that speaking up might get you disfavor, perhaps due to whatever element resulted in the earlier unfairness?

Personally, I've been a scaredy-cat about this. I actually didn't go to a different FAANG a couple years ago, partly because I anticipated a dilemma: the company had been in the news for some alleged unfairness, and I decided that I could easily imagine being there, and trapped between wanting to speak up on behalf of a fellow employee facing a similar unfairness in the future, and suspecting that I would be terminated if I did. Had I already been there (rather than merely anticipating the situation with a hypothetical) I don't know whether I'd speak up. (Personally, given past battle scars, I'd probably make quiet and delicate internal inquiries, trying to find a sympathetic person in a position to look into and correct or mitigate what seemed to be unfair, while being clear that I'm still a team player.)

So, I do think speaking up in a big-corporate environment, in the public manner these people have done, would tend to come with some personal risk in most companies. When people do it (perhaps) despite a perception of personal risk to themselves in doing so, we should pause and consider what they are saying.

Also, I don't blame people for not anticipating this situation to which they're objecting (based on the information available I've seen at this point). If the concern had instead been "I'm shocked -- shocked! -- to find the amassing of intimate personal information of Internet users going on here!", then I'd wonder how they couldn't have realized that before going to that company, but I don't think that's the concern.

I'm not taking a position on the concern itself, since the information I have so far is incomplete and complicated -- only suggesting that, when people are saying something at some personal risk, those people and what they're saying shouldn't be dismissed offhand.

To say the least those employees are misguided. Gebru was hired to research AI, yes she spent all her time being an activist. A pretty lousy one too. Case in point: 1. Her paper's claim on CO2 emission was not scientific in any way (projecting 100% of utilization of hundreds of desktop GPUs, really? Looking at emission without looking at how much would be saved or what other benefits are, really?). 2. Her paper claims that lacking support of all languages in BERT marginalizes minorities, which was similar to how Trofim Lysenko politicizing and moralizing biological science in Soviet. 3. See the quote blow. If this is not activism, I don't know what is. The only remaining questions are, why would Google AI hire her in the first place, and why should she work in a research org? Note it's perfectly for her to work in Google or any other company so choose the companies. I just thought people do real research in a research org.

"Citation: https://old.reddit.com/r/MachineLearning/comments/k77sxz/d_t... Sample comment:

The thread is still up there so any googler can see it for themselves and verify I am telling the truth."

Google is an company that demands extraordinarily high ideological conformance from its workers. Unfortunately for Google it new longer controls the exact specification of what is and is not acceptable thought at Google, and much like the French revolution it’s being pushed into directions that is management may or may not agree with.

What should be a fairly reasonable question – is this person a good employee or not is immediately contaminated by questions they have nothing to do with job performance.

"What should be a fairly reasonable question – is this person a good employee or not is immediately contaminated by questions they have nothing to do with job performance."

There are all sorts of non-performance considerations that are important, not just at Google, but at pretty much every other company.

One of those considerations is ethics. Another is legality. Another is public relations.

Which of these do you think companies should ignore?

I thought those are also job performance considerations?

If they are job performance considerations, then what is the person I replied to talking about?

Ideological conformance. Google has a particular world view that insists that privacy is not a human right, that certain political positions are verboten and unions are bad.

It's sort of meta but...

The article repeatedly mentions Gebru's race and sex and includes it for the only Google employee quoted. Why does that matter? Is she more reliable because of her skin colour? Is he less reliable because of his genitals?

The cynic in me feels like I'm being subtly trolled. Or maybe I'm just oversensitive? :)

I think they're trying to make out it's racial and she was slienced on pointing out a racial issue.

It just seems to be conflating so many things: personal racism against her (which I see no evidence of) and institutional racism (which is some of what she researches but then her and his race are irrelevant). That plus sexism. Plus academic freedom and employee-employer relations and PR.

In the end, it's a mess of nothings, rather and 1 solid something imho...

It's typical SJW narrative. People are tired of it, it's like the boy who cried wolf story. Hopefully at some point people will just learn to ignore these "woke" SJWs as they have their tantrums.

This article is from a couple of days ago. The story the article is about has been posted to HN several times and I think that there are may be some relevant updates to the story since this article was published if you wish to dive in:

https://hn.algolia.com/?dateRange=pastWeek&page=0&prefix=fal...

Per Google search, Google currently employs about 119,000 people.

So 1,200 condemn it, and 117,800 do not condemn it.

As a scientist, at the very least I would also need to know how many workers applaud the firing, and how many are actively uninterested, otherwise quoting that number is entirely without meaning.

Having read the paper (https://gofile.io/d/WfcxoF), I can't take the idea that Google wanted to suppress it seriously (as the article argues).

In the paper, another paper is summarized describing the environmental costs of training large models. It is then argued that global warming will disproportionately affect marginalized people. This constitutes "environmental racism" because the main beneficiaries of these models are rich white people whereas the people bearing the costs are poor people who live near the equator. In my mind, this is two-points-make-a-line thinking.

The rest of the paper is more interesting. It argues that machine learning is conservative in that it "reinforces hegemonic language". I think this is true. Models are trained using data, data is what exists rather than what we would like to exist. If you're unhappy with what exists (as Gebru is, terribly), this represents a problem. Two of her solutions are "curating" data (censoring data) and "working with panels of experiential experts through the Diverse Voices methodology" (constituting powerful groups made up of her ideological allies). I think these solutions are dead-ends. It seems to me that we have to make peace with the conservatism inherent in these models (and maybe all models).

The paper also argues (I think rightly) that problems will materialize "should humans encounter seemingly coherent language model output and take it for the words of some person or organization who have accountability for what is said". In other words, we grant other humans responsibility for what they say and do but what about models that can mimic humans very accurately? A similar problem exists today when a Google or Youtube account is erroneously locked by some algorithm. There's a sense that, really, no one is responsible for this outcome because an algorithm did it. Even if the owner manages to get his account restored, there's no clear way to assign responsibility for the mistake. Perhaps this problem will become even worse if we start to see algorithms powered by models that are near-indistinguishable from people.

Anyway, the paper doesn't seem controversial to me. It seems clear that Gebru's strident tone and personal style were the cause of her firing/resignation. Jeff Dean's claim that the paper was rejected for not citing other papers may be false. The truth seems much simpler: based on her own public communications, Gebru felt she was the subject of "harassment" and "dehumanization" but those grievances seem to be hyperbole, and the personal injury that she felt became an excuse for her to treat her coworkers poorly. In other words, she didn't drop her twitter persona at work.

Remember the last time Google tried to make the algorithm less biased only after some egregious cases of comparing Africans to gorillas have been found.

https://www.theverge.com/2018/1/12/16882408/google-racist-go...

At least Google did better last time ngl

But how many applaud it? (In secret because supporting this is something you might get fired for...)

You might get fired for... supporting your bosses' decision to fire someone?

More like, you support the decision, but your immediate boss doesn't, so they start looking for reasons to get rid of you that don't have any official connection to your stance. I remember a while back seeing leaked documents from Google where some managers shared an unofficial blacklist of employees to never promote and try to get rid of because of their political views.

Every non-union employee in every state other than Montana has this potential problem.

The difference is in smaller companies the unspoken rules are often spoken out loud and the employees don’t make Google money.

How does Montana deal with that problem differently?

Is that true?

Why do people call it "the land of the free" then?

Why are we conjuring hypothetical situations in which you are oppressed, instead of condemning the concrete case where Jeff Dean looked for flimsy reasons to get rid of Timnit Gebru?

CA is an at will state. He can fire her for no reason if he wishes, let alone a "flimsy" one.

There is a strange thing in modern discourse where there is an advantage in first establishing yourself as a victim before making your point.

Judging from the comments from some of her ex-colleagues, she was not exactly an easy person to deal with:

> The GPT-3 thread you describe was my first exposure to Timnit. Watching that thread unfold left me feeling upset, frustrated, and disappointed.

> I was so excited in anticipation of other Googler's reactions and insights about GPT-3, but that thread got immediately derailed by Timnit into claims of racism, not being listened to, dehumanization, that the whole forum became icy and dead after that.

https://old.reddit.com/r/MachineLearning/comments/k77sxz/d_t...

I'm still trying to gather more insight into what actually happened here. But you have to know upper management calculated there'd be blowback for firing someone of her stature, position, background, race, gender, etc. They did it anyway. That to me says that keeping her around was worse than the storm they knew they'd generate by firing her.

I get that Timnit Gebru is well respected in the field of AI ethics, and that the demand via HR that she retract the paper without discussion or appeal at the 11th hour sounds super shady

That email blast though. It made it so much worse for her position.

This article read like her paper was on bias. But isn't the paper on the environmental effects of large models? The article reads like Google tried to slience her on something racial.

I am more than happy to take one of their places.

Fire them, all of them!8-))

I thought she offered her resignation and they accepted?

She offered to resign on a particular schedule to work out things with her team if they insisted on retracting the paper. They terminated her employment immediately, while she was on vacation, because of an email she sent to the ERG about it, and they didn't even inform her direct manager.

Whatever they accepted, it wasn't the resignation she offered.

(And even if they had done so, I expect the folks here to remain unhappy about the circumstances that lead up to her offering her resignation. It's not a thing you do if things are going well.)

> She offered to resign on a particular schedule

It is not out of the ordinary at all for companies to walk people out on the day of resignation. The company is not required or even expected to abide by whatever schedule is presented to them.

If the employee says "I'm giving you 1 month formal notice of resignation", and the company responds with "effective immediately you are terminated without pay or benefits", at that point it's not a resignation any more, it's a termination.

In most countries (which are not "at will"), I believe if the employee had not breached their contract they could claim wrongful termination in this scenario and likely be awarded the pay they would have received for the missing notice period.

This doesn't mean the company is obliged to keep the employee on premises, on the network, active etc. It is not uncommon to ask the employee to leave the premises, have security walk them out, disconnect from the network etc out of security/IP/morale/whatever concerns. But they would still be either technically employed and paid while the notice worked out or paid in lieu of notice.

This is completely untrue, at least with 'at will' employments. Once you resign, you quit.

Working for any additional length of time after that often occurs if it is in both parties best interests. If it isn't though, the relationship is usually terminated immediately after resignation.

> This is completely untrue, at least with 'at will' employments.

The sentence you are replying to literally excludes 'at will' countries from the scope of their point...

>> In most countries (which are not "at will"), ...

Sorry, I was replying to the first sentence.

Then you are still mistaken.

Once you resign you do indeed quit.

But when you say you are going to resign at a future date, especially when you say you are going to resign at a future date only if certain conditions occur but will not resign if they don't... then you haven't resigned yet.

That's not special to employments. Similar applies to contracts in general, where a party tells the other party that at some future date, and/or under some condition, they are committing to exercise some clause in the contract. It's almost as if the message is a courtesy notice, an advance warning, of what the first party is going to do. They haven't done it.

In 'at will' locations the effect is much the same as quitting the same day because the company has the power to 'let you go' without giving a reason anyway - although the company must still comply with the contract, discrimination law, its normal operating procedure etc.

But even though the effect is much the same as quitting immediately - you are let go and employment ends - for anything where the difference between 'resigned' and 'terminated' matters, for example rights to government support, or loss-of-job insurance payouts, the company deciding to terminate early is a 'termination'.

> But when you say you are going to resign at a future date, especially when you say you are going to resign at a future date only if certain conditions occur but will not resign if they don't... then you haven't resigned yet

Maybe if you want to play semantics, sure. In the eyes of the employer though, you have effectively resigned.

I don't get why people keep mentioning how things are in other countries. We all know the US is bad for employee rights, why bring up how it is in other countries?

I think it's good to remind people how terrible the US is when it comes to worker rights. Many people don't realize just how bad they've got it over there.

Because people are debating whether it is a resignation (where the employee is in control of events), or a termination ("firing", where the company is in control of events).

Although US has 'at will' in most states, in many respects the way employment issues are judged in the US is similar to elsewhere. So the perspective elsewhere on what counts as termination seems relevant.

You're free to disagree, but that's why it's brought into the conversation.

> they could claim wrongful termination in this scenario and likely be awarded the pay they would have received for the missing notice period.

Correct.

That's what happens in Italy for example and I believe in large parts of Europe (I know for sure France, Germany, Sweden and Spain)

The missing pay is granted anyway, but with a wrongful termination the employer will most likely obtain a fair compensation.

A couple decades ago I was awarded a year of salary + the 3 months of notice we agreed upon when I was hired, for a wrongful termination after an acquisition.

Employers just need to pay you whatever period is written in your contract (which may be longer or shorter than 1 month). Which may very well be what happened here - nobody claimed she's no longer paid a dime, just that she no longer has access to Google resources.

You can't just say "i'm giving you 1 month of formal notice of resignation" and expect the company to be bound to that month... otherwise, to take it to extremes, what's stopping you from giving them a 20years formal notice of resignation?

In the US, in most states, 'at will' employment often means zero notice is required from either side. A contract can specify otherwise, as can other things (such as a company HR procedure).

> You can't just say "i'm giving you 1 month of formal notice of resignation" and expect the company to be bound to that month... otherwise, to take it to extremes, what's stopping you from giving them a 20years formal notice of resignation?

You actually can say that if you want.

What stops 20 years notice from having any effect is that the company can terminate you anyway in the normal way before that. I.e. they can still make you redundant, fire you, assess your performance etc in all of the usual ways, or in 'at will' locations just let you go without a reason. They will still be subject to discrimination law etc if they fail to comply with all the usual constraints on terminating employments, but nothing about your 20 year notice prevents them from following the normal rules at any time.

ps. I really meant to say "giving 1 month formal notice in accordance with my contract of employment", but I realised in the US people don't always have any notice in their contract (due to 'at will') so I removed the part about being in accordance with it.

Ignore employment as a special case for a moment. In general, for any contract (implied or not), parties can give each other advance notice of something that is going to take effect in future, as if they told the other party on the future date of the effect; the effect will not occur until that date. The other party in the contract can't just decide the effect has taken place earlier. You can also give conditions. In a sense, notice of this kind is just a courtesy, an advance warning if you like.

I believe in US 'at will' terms, Timnit effectively gave notice that she was going to resign at a specified future date under certain conditions. Timing had not actually resigned at the point at which Google is said to have terminated the employment. That makes it a termination.

I was mostly disagreeing with "wrongful termination" - a "conditional resignation" is a good reason for termination, and there's nothing wrongful to terminate the employment earlier than his proposed terms as long as you respect the contract (pay for the notice period, to the extent that such period exists).

Are you sure she isn't being paid for her notice period?

> It is not out of the ordinary at all for companies to walk people out on the day of resignation.

I've observed a few cases of this as a bystander. Generally, the employee escorted out had accepted a job offer from a competitor -- and that doesn't appear to be the case here.

This is just not how employment works, really anywhere in the world. The company is well within it's right to ask you to leave earlier to protect it's IP.

If you're a Google employee I have to say it seems like you guys a pampered beyond believe. Someone publically tells people not to work on something for the company and doesn't get fired for that but gets shown the door after demanding confidental information or else and you guys are unhappy? Stay at Google, because most companies would have had you out the door if you sent a public email telling people to stop working.

All those 1200 should threaten to go on strike and hit Google's profits.

1200 employees going on strike will literally have no negative impact on Google's bottom line, at least not unless they do it for a long time period. A lot of the money that comes in is basically automatic, it's being produced by ad sells and stuff like that. You'd need a massive subset of the company's staff to go on strike (not just 1200) before that would have any impact on whether the money-making machine continues to function. In a case like this walkouts or strikes by the tech workers have more of a symbolic purpose + impact on future deadlines/releases than a direct impact on revenue.

Maybe if the entire SRE department went on strike and stopped maintaining the datacenters that would put a dent in the numbers due to degraded service, but something like that would probably just doom the company entirely even if they went back to work after a bit.

> Maybe if the entire SRE department went on strike and stopped maintaining the datacenters that would put a dent in the numbers due to degraded service, but something like that would probably just doom the company entirely even if they went back to work after a bit.

It's the threat of being able to strike and potentially cause that much damage to the business that makes it powerful.

No-one really wants to strike, actually going through with it is the final step when the company fails to meet the workers' demands - no point in making a threat if you don't demonstrate it when the circumstances arise, it's toothless otherwise.

I respect that she sacrificed her prestigious role to force Google to at least admit where they really stand.

Unfortunately, it's clearly more complex than that. To be fair, it seems that she has made it about that to an extent and media ran with it. We haven't heard from google or her superiors. It's clear they didn't agree but to what extent? And why did she need to publicly quit? Again, she quit. She wasn't asked to leave or resign. She volunteered to do so because she had a moral disagreement with the company. And that's fair. We don't really know if she was treated fairly or unfairly just her claims of what happened.

"Old men's club" is real and need a shakeup. However, the elite clubs hurt both genders, not just one of them. It is not balanced by an equally forceful swing the other direction, but by emphasis on mutual respect for diversity of people. Sometimes measures need to be put in place to get the ball rolling, but they should be temporary.

I think we agree.

Yeah, but "The Middle Path" is a hard sell to the masses.

I think there's a weird seam of irrational rage around the mere suggestion that tech isn't some utopian meritocracy.

I find it jarring.

Yep. People mistake means for the ends.

Anybody may be susceptible to such ideology and short-circuited bias. However, technical people are often so focused on their details, we are probably even more vulnerable to this than most.

Threatening to leave is often times the ONLY leverage a worker has. What other leverage can a human utilize, in order to get their concerns addressed?

And how seriously does Google managers really take AI ethics, when they fail to resolve such an issue internally?

This is troublesome on so many levels and a huge red flag anything to do with AI and Google.

"Condemn firing" is a misleadingly narrow headline that focuses undue attention on the resign/fire question. The important part of the story is not the HR machinations of her departure, but what caused the situation to escalate to that point. The letter itself makes this clear -- its condemnations and demands go way beyond Gebru's employment:

https://googlewalkout.medium.com/standing-with-dr-timnit-geb...

> Dr. Gebru has faced defensiveness, racism, gaslighting, research censorship, and now a retaliatory firing

The letter goes on:

> In late November, five weeks after the piece had been internally reviewed and approved for publication through standard processes, Google leadership made the decision to censor it, without warning or cause.

The demands are also not about the firing itself:

> explain the process by which the paper was unilaterally rejected by leadership

Reply on news.ycombinator.com