Gitlab: don't discuss politics at work • Hacker News Board

Re: Gitlab: don't discuss politics at work

#921

Post by tudorw » Thu, Oct 17, 2019, 9:12 AM UTC

neonate wrote: post #280

anm89 wrote: post #129

I personally think this is awesome. I don't wan't some git hosting startup to be the arbiter of morality for society. The engineers, designers, and PMs shouldn't have an outsized voice in society because they have a specialized useful skillset and ended up on a successful product. If these users are breaking laws, then put them out of business via the courts and sieze the assets (the repos in this case) via legal mea…

Wow, that is such a clear analysis that someone should transcribe it. Here's the main part of what Naval says: "If Google, Facebook, and Twitter had been smart about this, they would not have picked sides. They would have said "We're publishers. Whatever goes through our pipes goes through our pipes. If it's illegal, we'll take it down. Give us a court order. Otherwise we don't touch it." It's like the phone company.…

If Google, Facebook, and Twitter had been smart about this, they... would have consulted deeply with lawmakers, regulators, copyright owners and other deeply entrenched parties before doing anything... move fast, and, er, what was it again ?

Re: Gitlab: don't discuss politics at work

#922

Post by repolfx » Thu, Oct 17, 2019, 9:14 AM UTC

izacus wrote: post #493

repolfx wrote: post #147

Earlier quoted context omitted.

I don't think even Google actually encouraged people to discuss politics at work. They never did when I was there. It was more a side-effect of decentralising the power to create forums. Anyone could create discussion lists there without asking permission, this was highly useful for coordination, making it easy to spin up new projects and avoiding siloing of information into people's private mail accounts. The flip s…

> This is a very far cry from the modern Google where arguing maybe girls find geeky stuff boring gets you fired, where they manipulate ML models to believe the world is some intersectional 'utopia' as part of social engineering programmes, where employees spend huge amounts of time erasing gendered pronouns from source code comments. Huh? Can you link some substantiations to these claims, it seems way bizarre?

Erasing gendered pronouns from comments: no, because that's a story I heard from former colleagues who are still there.

Arguing maybe girls find geeky stuff boring: this is what Damore did, but he wrote it much more formally and cited lots of studies. It boiled down to though, "most women find tech boring because they're women". Damore's essay is here: https://firedfortruth.com/2017/08/08/first-blog-post/

Manipulating ML models: https://developers.googleblog.com/2018/04/text-embedding-mod... with their stated example:

An example of bias in this context is if the incoming message is "Did the engineer finish the project?" and the model scores the response "Yes he did" higher than "Yes she did." These associations are learned from the data used to train the embeddings, and while they reflect the degree to which each gendered response is likely to be the actual response in the training data (and the degree to which there's a gender imbalance in these occupations in the real world), it can be a negative experience for users when the system simply assumes that the engineer is male.

This logic is broken and wouldn't have happened in the old Google. If you're predicting what response the user is most likely to type next, then "Yes he did" is a more useful prediction by any objective measure because most engineers are men. But here, Google AI Research concludes that the most likely prediction would be "a negative experience for users" and sets out to "debias" their models. Debias in this context means to bias the model away from learned reality and towards what liberal intersectionalists want the world to be, in the hope that by subtly manipulating people through AI predictions they can actually bring that world about.

Re: Gitlab: don't discuss politics at work

#923

Post by jka » Thu, Oct 17, 2019, 9:17 AM UTC

austincheney wrote: post #764

jka wrote: post #463

Earlier quoted context omitted.

The trouble with libertarianism is, ultimately, that libertarians only care about themselves. The kind of policy that GitLab is proposing here is fine - or even aspirational, under a certain lens - until one day you wake up and find that the same software is being used to destroy your community and compatriots. How could that happen? Isn't this free speech to max utility? Isn't that flawless? No, it's free utility to…

> The trouble with libertarianism is, ultimately, that libertarians only care about themselves. I am wondering on what you base this line of thinking? Do you have any examples where persons advocating for liberty are acting primarily out of self interest?

You're right, this was an off-the-cuff generalization and wasn't well explained; here's roughly where I was coming from:

- GitLab is claiming to do business with any customer, on the basis of free trade and contribution

- GitLab is encouraging a culture where employees do not discuss or talk about those relationships

- I believe that technology is not neutral and the behaviours we as software engineers define in code, which then may affect thousands or even millions, are a result of customer influence (product decisions, feature requests, pull/merge requests, etc)

The commit[0] I read describes this as inclusion and 'efficiency' (i.e. 'yours not to question why'), but the real outcome I foresee is: "we are going to do business with shady companies and/or countries, and we want to encourage a company culture where this is seen as fine and dissent is discouraged"

This is self-interest on the part of GitLab the corporation - and ironically it also reduces liberty and freedom for GitLab employees to speak openly.

Meanwhile I can understand the top-level comment's description of why this might in some sense be a good business strategy. There's not a whole lot of discussion around the negative externalities of being successful in a way that compromises values though.

[0] - https://gitlab.com/gitlab-com/www-gitlab-com/commit/b5a35716...

Re: Gitlab: don't discuss politics at work

#924

Post by ajscanlan » Thu, Oct 17, 2019, 9:17 AM UTC

fake-name wrote: post #898

dcolkitt wrote: post #847

Earlier quoted context omitted.

You're stretching the definition of "political" to such a degree that it renders the word meaningless. You're basically claiming that any act that has any impact on anyone is political. That it's basically impossible for an organization to be non-political or even less political than any other organization. And that simply flies in the face of common sense. Only the most hardcore post-modernist would deny that the NR…

> to strive to be non-political. To strive to be "non-political" is an intensely political position. It's basically a complete support of the status quo. To not take a stand on a position is as much of a position as taking a stand. We live in this world. It's literally impossible to disengage. Not saying anything, or not disagreeing is as much a statement as saying something, or disagreeing. Neutrality is a political…

> It's basically a complete support of the status quo.

I've seen many people spout this off as if it's self-evident, and I can't for the life of me understand why.

The status quo isn't some neutral, natural resting state that will continue on until otherwise affected, it needs to be _constantly_ maintained.

Being non-political doesn't maintain the status quo, because maintaining the status quo is an _active_ endeavour, in the same way a plane doesn't keep flying forever just because its engines turn off.

Re: Gitlab: don't discuss politics at work

#925

Post by repolfx » Thu, Oct 17, 2019, 9:20 AM UTC

neonate wrote: post #386

repolfx wrote: post #147

Earlier quoted context omitted.

I don't think even Google actually encouraged people to discuss politics at work. They never did when I was there. It was more a side-effect of decentralising the power to create forums. Anyone could create discussion lists there without asking permission, this was highly useful for coordination, making it easy to spin up new projects and avoiding siloing of information into people's private mail accounts. The flip s…

What's an ITT? You seem to mean none of these? https://en.wikipedia.org/wiki/ITT

Experiments that ask people to fill out surveys as themselves, or as they imagine a liberal would, or as they imagine a conservative would, in equal proportion.

Jonathan Haidt did one of these experiments with about 2000 participants. It's social science so it should come with all the usual disclaimers, but the results can be found here and have apparently been replicated since (though I'm having trouble rapidly re-locating a study I found that replicated it, for some reason I can't find the right keywords to surface it).

You can find a writeup here: https://goodmenproject.com/featured-content/sympathy-for-the...

Re: Gitlab: don't discuss politics at work

#926

Post by repolfx » Thu, Oct 17, 2019, 9:20 AM UTC

cbuchanan wrote: post #237

repolfx wrote: post #225

Earlier quoted context omitted.

That merge appears to change the policy in ways that render this discussion largely irrelevant (given the current subject line). It used to say: We do not currently exclude anyone from being a customer based on moral/value grounds. followed by a list of reasons why not. It now says: In some circumstances, we may opt to not work with particular organizations, but we assess those on a case-by-case basis. Some reasons w…

We're a team that moves quickly and is receptive to ideas and opinions. I'm very proud of that.

It looks from the outside like the opposite - you changed the policy explicitly so you can refuse to deal with organisations whose ideas and opinions you're not receptive to.

Worse, it looks like you just reversed direction because the Register made fun of you. You went from being willing to work with all organisations to reserving the right to drop any organisation on a whim. I wonder, what does Gitlab believe exactly?

You aren't the CEO but apparently the actual CEO lets you set global sales policy, so perhaps you can illustrate with some examples how your new-new policy will work.

I ask because the edited policy looks more like a personal speech code than something designed for companies. Do software development organisations routinely "encourage violence towards protected groups"? They don't, in my experience. So which organisations did you have in mind when writing it? Who exactly will you refuse to do business with now who would have been served before? You list some categorisations that look quite vague and state it's not a full list.

For example, in recent times certain political factions have argued that immigration control is violence against protected groups: can you clarify if you do/would sell Gitlab to ICE?

Re: Gitlab: don't discuss politics at work

#927

Post by yxhuvud » Thu, Oct 17, 2019, 9:26 AM UTC

Gunax wrote: post #512

anarchodev wrote: post #331

Earlier quoted context omitted.

> I don't wan't some git hosting startup to be the arbiter of morality for society. The engineers, designers, and PMs shouldn't have an outsized voice in society because they have a specialized useful skillset and ended up on a successful product Refusing to serve a customer when you disagree with that customer's goal is pretty far from being a "moral arbiter for society." Keep in mind that anyone is free to use othe…

> Refusing to serve a customer when you disagree with that customer's goal is pretty far from being a "moral arbiter for society." Keep in mind that anyone is free to use other services (or roll their own) and gitlab can't do anything about that. Neither would it be overreaching for the workers building that product to request a say in how it's used. I think this is the crux of the issue. Anyone is of course free to…

> Facebook control ~90% of social media.

Source? That doesn't really match the numbers I've seen. Especially not for younger people.

Re: Gitlab: don't discuss politics at work

#928

Post by xurias » Thu, Oct 17, 2019, 9:30 AM UTC

I think it's concerning that everybody seems to think it's okay for companies to clamp down on political discourse. Where do people think we're supposed to talk about this shit? Who's supposed to champion these causes if not companies? Who has actual power in this world other than companies? Companies dominate the media, the way we consume media and everything that we see.

All this means is nothing changes, nothing improves and companies are free to profit off of suppressing rights in authoritarian countries without blowback from people using their platform (like Blizzard/Riot) or people that work for them. Companies are not and never have been neutral entities. If your goal is profit maximization to the detriment of the social fabric and moral principles, then sure, go ahead and pretend to be neutral. But I'm not sure that should be the goal and we shouldn't be helping companies pursuing this goal.

Re: Gitlab: don't discuss politics at work

#929

Post by repolfx » Thu, Oct 17, 2019, 9:42 AM UTC

lovich wrote: post #318

repolfx wrote: post #147

Earlier quoted context omitted.

I don't think even Google actually encouraged people to discuss politics at work. They never did when I was there. It was more a side-effect of decentralising the power to create forums. Anyone could create discussion lists there without asking permission, this was highly useful for coordination, making it easy to spin up new projects and avoiding siloing of information into people's private mail accounts. The flip s…

>I suspect the shift was caused by Google running out of experienced software engineers to hire, and switching to growing through recruiting new college grads at much greater numbers. Could this not have easily been a backlash against the companies no longer operating neutrally? You describe a timeline here that goes neutral Google -> hard left because of the employees. That doesn't jive with what I've seen in the ne…

I didn't describe a neutral Google internally: looking back, it's possible there were a bunch of closet Marxists who felt uncomfortable admitting that. I doubt it because in other offices I've never seen people on the left hold back from publicly announcing their views (it's called virtue signalling for a reason!), but I guess it's possible.

The companies policies were neutral though. There was no concept that helping people find information might hurt them, as you see in modern day Google.

During the time period I'm talking about Google got banned from China because they refused to continue censoring the search engine. That was circa 2010, I think. So consistent with a company having a very free speech, libertarian view. It made waves because so few companies were willing to sacrifice the Chinese market on the altar of free speech, but Google was. Building a new censored search engine only seems to have started after Pichai took over and Brin/Page checked out.

As for "started surreptitiously tracking everyone", that was a media attack that started around the time Google News took off. Nothing had actually changed about privacy policies of tech firms, and users were clearly very happy with the ads-for-services arrangement given the rapidly rising usage numbers. But the news industry was shrinking, and editors/owners in particular felt hugely threatened by Google News, which effectively replaced them with an algorithm and commoditised news almost overnight. The media went from being very positive about Google (great company to work for) to trying everything they could to attack and destroy it. What they wanted was money, pure and simple, and eventually as the attacks stepped up they stopped pretending otherwise. Hence the new EU Copyright directive and "link taxes" that started popping up.

The employee backlash phenomenon only seems to have started recently. There were always pointed questions at TGIF when I was there, but disagreement with management was largely congenial. Things started going downhill around the time Colin McMillen created Memegen. It was basically an internal Twitter but worse, I don't think you can ever fit 140 characters into an image macro. That introduced and encouraged Tweet-sized thinking/viral liking campaigns whereas previously disagreement had been surfaced in email debates (the famous "centi-threads") or through face-to-face discussion. Long email discussions did annoy some people, but they at least had an expectation you'd contribute by writing something useful. After Memegen got big I started seeing people unironically cite that they got highly voted memes as evidence of creating value in their internal CV and promotion packets, which was ridiculous.

Re: Gitlab: don't discuss politics at work

#930

Post by pishpash » Thu, Oct 17, 2019, 9:44 AM UTC

Mvandenbergh wrote: post #334

neonate wrote: post #280

Earlier quoted context omitted.

Wow, that is such a clear analysis that someone should transcribe it. Here's the main part of what Naval says: "If Google, Facebook, and Twitter had been smart about this, they would not have picked sides. They would have said "We're publishers. Whatever goes through our pipes goes through our pipes. If it's illegal, we'll take it down. Give us a court order. Otherwise we don't touch it." It's like the phone company.…

The problem with at least Facebook and Twitter (and possibly Google) in this analysis is that their attention-grabbing model requires them to analyse and interfere with the data which is coming through their systems. The only way not to do that would be for FB timelines and Twitter feeds to be strictly sequential feeds which at some point in the past they may have been but they are not now. Once you start basing your…

Publishers don't exercise editorial control? Come on now.