Hacker News Reader

It seems to me that many programmers suffer from some privacy blind spot where their brain literally freezes when they have to think about the negative consequences of their latest software gizmo.

It was decided in a court of law that employers can't spy on their employees. It's good that companies are being cautious. Not everything in this world revolves around efficiency and tracking everything.

I certainly do not suffer from "privacy blind spots". As a matter of fact, I've worked for one company in the telco space doing industry-leading work on setting standards about how to properly anonymize call detail records. And then, in this German telco that I mentioned, we were doing groundbreaking work setting up our data engineering in such a way as to comply with Germany's stringent privacy protections where consumers are concerned. I now work in another privacy-focused company.

But this one story I was telling:

Management: Please make a mathematical model that will predict if a radio unit has repeated and prolonged outages, using well over a hundred independent variables, ranging from technical information, to network topology, to information from the ticket system.

Me: Done. Mathematical model says: Radio units have problems if the people servicing them are lazy bums. Here is a system, based on natural language processing, that will predict people being lazy bums with huge accuracy.

Management: So who are the lazy bums?

Me: Can't tell you that. Because privacy.

I mean. Come on...

"Here is a system, based on natural language processing, that will predict people being lazy bums with huge accuracy."

What could go wrong... I was quite impressed until you got to this point too and thought I had been too harsh.

The patterns my system noticed were literally things like this: There's an error entered into the system by the NOC operator as an intermittent error on the station. Field technician drives there. Enters something like "Station on air upon arrival" into the ticket system. Does nothing else. (Like nothing to diagnose or fix the actual problem). Closes the incident. Drives home. Pockets the money for the time spent driving. A few days later (who would have thought that?) a new incident is opened for the same station. Repeat once a week for months and months.

I'm definitely not someone who is too naive to understand that technology like NLP has limits. But this was a perfectly appropriate usecase for NLP and it worked perfectly fine.

And people acting in this manner should not be immune from repercussions for reasons of data protection.

...also, it seems weird to me that workers representatives knew, before the project even started, that it was in their best interest to ask for this data to be anonymized. It was as if workers, collectively, knew the whole time that this sort of thing was systematically happening, and were actively working towards preventing management from finding out. That's what I mean by "not pulling on the same string".

It's legally their job to represent only the workers. And they legally have no way of knowing, what's really going on economically. They may be informed about the financial status, if the company is big enough for a economic sub-committee. But they can't make any decisions on that. They aare there solely to represent the interests of the workers. See paragraph 80 Betriebsverfassungsgesetz.

Your example only shows, that your workers council is unhappy with the provision based system, because it's unfair. Regional targetting would make it even unfairer. The solution is easy: abandaon the provisions and let the people do their jobs.

If they have seats on the supervisory board, then they have wide-ranging legal powers to know about everything that's going on in the company.

And thanks for pointing me to paragraph 80 of the Betriebsverfassungsgesetz. I read it just now, and "representing the interests of workers" is totally not what it's saying. It's saying: They should ensure that legal & contractual provisions that exist in order to further workers' interests are observed. That's totally not the same thing. -- Because a law doesn't become law unless the needs of both workers and employers have been considered. And a contract doesn't become a contract unless both workers and employers have agreed. And their role is simply to ensure compliance with those provisions. Where there is no provision, they have no mandate. For example, coming back to my example: There is no legal provision that protects workers from repercussions if there is documented evidence that they have been lazy, so the workers council has no mandate to try to protect lazy workers, yet for some reason they do.

The legal duty of the supervisory board is to watch out for the company's best interests, period. Even those board members who are sent there by the workers council are, in theory, bound by that duty.

I'm not against trade unionism. Trade unions exist so that workers don't suffer from reduced bargaining power when making contracts with employers. But still: Employers and employees are separate economic entities with separate interests who engage with each other over a market interface, with each looking out for their own bests interests (and no one else's).

Workers' codetermination is taking it one step further: You have a company that has certain interests. And you insert people into the decision-making process of that company who observe the best interests of an entity that is completely separate and has completely separate interests.

Imagine the relationship between General Motors and Microsoft as an analogy for the relationship between General Motors and General Motors Employees. With GM having lock-in for Windows software, MSFT has a lot of bargaining power. But each of GM and MSFT look out for their own best interests when they negotiate licensing. -- That's trade unionism.

Now imagine what it would be like if providing software licenses legally entitled MSFT to put people onto GM's board. This would mean GM would actually be hampered in its ability to make decisions that are in its own best interests (both as it negotiates with MSFT, but also as it does pretty much anything else). -- That's worker codetermination.

On the extremes, I see two ways of doing negotiation - or in general, reconciling conflicting interests: a competitive one, and a cooperative one.

What you're describing is the competitive one: you have sides focusing solely on their local, direct, short-term interests, in hopes that as they all fight each other, the aggregate state of the system will be a decent compromise. The cooperative way would be if all parties got together, did a global optimization across all their goals, and went along with the result until something changes and it needs to be readjusted.

For some reason, the competitive way is the most common one. It's in what you say about worker representatives vs. company. It's how the whole market works. The good side is, such system is very dynamic and quick to respond to shifting preferences (and shifting landscapes). The bad side: it's ridiculously wasteful, and it turns people against each other.

I suppose the reason the competitive model is more common is because you can't just get random people to do global optimization together - it's hard to ensure honesty and mutual understanding of all participants.

RE wastefulness, I like to compare it to the issue of active vs. passive stationkeeping in physics[0]. Imagine that your village wants a new fire tower. You want to have a few cameras 50 meters above ground for spotting fires. There are two ways to approach it. You could build an actual 50m tower and place cameras there. Or you could go to the store, buy a large quadcopter (and a lot of spare parts), tether it to mains power and keep it continuously flying at 50m. Building an actual tower has a much bigger up-front expense, but then it's essentially free. Getting a camera-equipped drone flying is cheap, but keeping it flying for days, weeks, years would waste absurd amounts of energy, very quickly exceeding all the energy put into building the tower.

Obviously, cooperative approach is like building the tower - the desired goal is made explicit and free to maintain. Competitive is like flying that drone - the desired goal emerges from opposite forces fighting, and requires the battle to go on forever.

--

[0] - it's not the real name of the consideration, it's just how I call it. I wonder if it has a proper name.

Interesting. This is a pretty common criticism of unions in the anglosphere as well -- often alongside a note that the problem has been "solved" in Germany via worker codetermination.

Have you found this situation to be universal in Germany or only at some employers? This is one of those things I'd actually expect "the market" to be good at regulating. Bad union or management kills a company, company with better management or union fills the market void etc...

This was the only company with worker codetermination that I've ever worked at, so I don't have anything to compare it to. Part of the reason is that, as the article says, companies will try at all costs to avoid it when there is legally a way it can be avoided, and I'm guessing that this in and of itself speaks to the fact that there's a problem there. Otherwise you might expect companies to volunteer into it for reasons of improving their employer brand and so forth. But the opposite is the case, with many large companies going as far as incorporating in other European jurisdictions and using European freedom-of-movement rules for legal entities to operate out of Germany. N.b.: This confers no tax advantages. It is done solely for the purposes of escaping Germany's companies' law. (Examples: The German department store "Müller" with over 700 retail locations in Germany and "Air Berlin", a major airline out of Germany, were incorporated in Britain).

Why would a company's board of directors volunteer to, basically, hand off a significant part of their power to somebody else? Somebody whom they know will definitely not optimize for upper executive compensation.

Actually, executive compensation is usually subject to board approval. The board is answerable to the shareholders. Executive compensation hits a company's bottom line, and therefore shareholders' pockets. Shareholders are perfectly well-placed to keep executive compensation low, if they feel that that's what gets them the best value, or allow it to go high (with that compensation coming from their own pockets) if they feel that that's what gets them the best value.

I really don't see where employees come into this equation, other than perhaps wanting to keep executive compensation low for reasons of petty jealousy.

This is extremely naive. One reason executives get more compensation is because they frequently sit on each-other's boards and are more than happy to scratch each-other's backs. At large companies, most shareholders are "institutional investors" (rich dudes managing other people's money) and are not going to be showing up to shareholder's meetings complaining about executive compensation.

Meanwhile those same shareholders are very happy to complain about worker's compensation[1], for reasons of petty greed.

[1] https://www.salon.com/2017/04/28/labor-is-being-paid-first-a...

If shareholders were indeed "rich dudes", then why do you have a problem with rich dudes paying too much money to other rich dudes for being executives? With that logic, "normal" folk don't even enter into the equation, do they?

Institutional investor is not a euphemism for "rich dudes". Institutional investors are overwhelmingly things like pension funds and insurance companies, whose clients are people like you and me. -- Even more so in Europe, where stock ownership by individuals and private equity makes up an even smaller proportion than it does in the U.S., and top-level executives in large corporates still make a lot of money.

If a pension fund wanted to, it could represent the interests of pension fund holders in general meetings, although I will grant to you that they don't usually live up to that duty. But that doesn't mean that such a thing as shareholder activism doesn't exist. We are seeing increasing activity by agencies who collect proxies from institutional investors to push for change around certain issues at shareholders' meetings.

...even with all of that discussion, I really don't see why employee activitsm is needed over shareholder activism to rectify the situation, if indeed there were something wrong with the situation.

Furthermore, I don't believe that there is anything wrong with the situation: A C-Suite executive at an S&P 500 corporate or a EuroStoxx 500 corporation should, in my opinion, make more money than a football player or Hollywood star, because what they do is actually more valuable. -- For some reason, the general public never complains about the level of compensation of the latter, even when, as is the case in Europe, it is the taxpayer who pays for this through mandatory contributions to public broadcasting (which, in turn, get used to pay for broadcast rights to sports events and so forth).