Hacker News Reader

Comment by alphazard | Hacker News Reader

Don't unions work best in fields where quality is very uniform and expertise counts for very little? I'm thinking about low skilled jobs that can be mastered in months or years. Unions create salary progression, and benefits for seniority where there is otherwise no reason for them to exist. When no one does anything better than anyone else, it makes sense to negotiate as a block, individuals have no power.

Now compare that to tech where the 80/20 rule is patently obvious to anyone in the field. The best engineers are worth literally 10x the worst engineers, but often get paid only 2-3x as much. 20% of the field, with 80% of the negotiating power, wants employers to find better and better ways to pay proportional to value-add, not to increase the floor salary, or working conditions across the board.

View on HN

>The best engineers are worth literally 10x the worst engineers

Uh, I'm pretty sure the best employees in any field are worth 10x what the worst employees are, whether you're writing code, assembling a car or gutting fish.

The economics of labour suggest unions are most valuable where there's a monopsony buyer of labour. Big Tech might fit that bill, especially when we know they colluded to depress salaries.

The best airline pilot lands the plane the same as the worst.

With a few notable outliers...

99.999% of the time I agree with you.

Sure, just the integrity of the plane after landing vary sometimes.

Yeah the best developer uses the same compiler of the worst, guess the difference is not what you do at the end, but how you manage the trip, and how you write code before

>The best airline pilot lands the plane the same as the worst

Have you considered there's more to being a pilot than landing the plane?

Plenty more, but the economic impact to the airline is the same from any individual pilot.

Not really. The worst pilot only lands once, and never again.

Now what qualifies as landing?

Collect a large number of pilots, divide them into two pools based upon their performance (low and high). Which pool do you think will incur more expenses to the airline? Keep in mind that some of those costs range from reducing consumer confidence (poor landings or damaged cargo), to planes requiring more maintenance (increased costs of maintenance, planes taken out of service more often), to planes being put out of commission (regardless of whether everyone walks away or not).

A lot of people seem to be measuring value in landings per day, widgets per minute, lines of code per week without acknowledging that the quality of work does have an impact.

There's only acceptable piloting and not. There's no pilot that makes the airline 10x more per flight.

The best airline pilots get the best routes for the best airlines and are mostly retired military. At the other end you have cargo pilots and regional flights that carry a brand of major airline but are operated by a different company.

Look at salary differences.

https://www.thepilotguys.com/blog/airline-pilot-salary

>The best airline pilots get the best routes for the best airlines and are mostly retired military.

Curious about this. I'd have thought going the civilian route of flight school should put one on equal footing on this regard? Or are you saying that just statistically more long-tenured pilots are retired military?

In general, these are due to union rules. The union decides what the compensation is, and it's largely by seniority. The airline makes no more money from retired military pilots than they do from any other pilot.

If you are on an assembly line and you have 45 s to do your job you either do it, or don't. If you do it fast and wait for 40 s you don't add any value.

>If you are on an assembly line and you have 45 s to do your job you either do it, or don't. If you do it fast and wait for 40 s you don't add any value.

Have you worked on an assembly line? I have.

You may not add value by doing it faster, but you certainly cause problems if you do it slower.

So there’s a cap on the maximum value (always doing it on time), and a cap on the minimum value (never doing it).

The above poster takes the position that the ratio between the maximum value, and minimum employable value is less than 10.

I think the distinction is the relationship between the median performer and the 90th percentile performer. The two are very different employees for software development but not on an assembly line. As you point out, there is a difference between the median performer and the 10th percentile performer on an assembly line, but that's not the comparison the parent comment introduced.

>As you point out, there is a difference between the median performer

But I didn't say that, and I don't agree.

I think the 90th percentile never causes delays on a particular station, while the median may cause a few per hour, and the 10th percentile might delay every 2nd or 3rd cycle.

There are tangible costs associated with those levels of performance.

Generally you design your cycle times so that the median performer never is late. Remember if each median performer has 1 delay per hour, on a line with 100 workers that mean each cycle there is a delay somewhere and so the line always moves slower than expected. How much slack to add is a complex question, but you have to assume there will be delays. The 10th percentile cannot be allowed to delay the line, so you have to deal with it (training, fire them, give them faster tools, or switch to a station they can do in time come to mind, there may be more)

This is a completely different argument though and irrelevant to the comment chain you replied to. Everyone involved may even agree with you - why did you frame it as an argument against the claim that there is less capacity for 10x performers?

> You may not add value by doing it faster, but you certainly cause problems if you do it slower.

This makes his point though. There is no capacity to be 10x more productive.

With all due respect, why are you both talking about yourselves as if you were machines? If your job is so mechanical it can be automated, you're screwed. If it's not, but your boss wants to downsize it bc "savings", a union would help you.

Unions add the same kind of value whether you're writing a Hollywood screenplay or assembling a Tesla because you have the same relationship to capital either way.

The same divide/conquer tactic was done 100 years ago when owners of mines in Pennsylvania pitted white and black miners against each other.

The actual distinction doesn't matter to the owners provided it divides labor. Anything that makes you feel special and separates you from people in the same boat as you is fair game - whether that's the intellectual needs of your job or just plain old racism.

Do I have the same relationship to capital? A miner’s relationship to capital seems pretty plausibly adversarial to me: capital buys a bunch of tools and shafts and scaffolding, asks miners to use them to pull coal out of the ground, then pays the miners a fraction of what the coal sells for. My relationship with capital is that rich investors send money to my company’s bank account and they forward it directly to my bank account.

Rich investors send money to your companies bank account. Upper management then pays as little as they can get away with. You might have “shares” in your company that gain value in an “exit” event, but have you looked at how those payout structures are arranged? Mostly the investors and “founders” are paid out first with great returns. What little is left gets divided up amongst the rest.

There is value in an investor paying for the shafts and scaffolding. The investor also takes the risk that coal prices will fall (eventually the miner is laid off, but the investor takes a lot of the risk)

Mines also had investors and they also paid wage bills before the company was profitable. Did you think that they did not?

That depends on the assembly line. The assembly the car line works like that: you get 45 seconds (or whatever time) to put your part one, then the car moves to the next station.

However I know of assembly lines where things spend several hours at each station. The people (often as many as 10 people per station) all work together for that time putting different parts on. If you can do the job faster than the time in station you can take an extra break to do what you want. Such lines are generally planned so that if you work as expected you get a 20 minute break when the lines move, most teams work faster and get a 30 minute break, almost never does some team get less than the 20 minute break planned.

The car model assembly line is the most efficient, but it has high costs. When you make a lot of something it is well worth those costs, but if you are only expecting to sell 100 "things" per year it isn't worth setting up the line.

This idea crops up so frequently that Ive started calling it the "Apparently Tom Cruise Works in a Factory" fallacy.

Is Tom Cruise represented by a union?

Yes. He's a member of the Screen Actors Guild. He stopped shooting MI:8 in solidarity with the union.

The film industry is highly unionized. At some point in his career, he absolutely was - possibly multiple.

I'm pretty sure there are plenty of fields where the output of your best and worst employees don't differ dramatically. Worst is generally some passable minimum not someone who refuses to do the job entirely. For plenty of assembly line jobs, your worst employee has similar productivity to your best employee because there is no ability to make the assembly line faster for the more productive person.

In code on the other hand, your best employee can straight up do things your worst employee can't. The differences are much larger. There is a reason people hire Principal Software Engineers, but most IC roles in other fields don't have the same range of seniorities.

I've been doing software professionally for about thirty years now, and what's interesting to me is that the conversations I had with my grandfather about factory work involved all of the same dynamics. Prototyping new solutions versus consistent reproduction of known solutions? People who only understand their own tasks versus those who have a holistic understanding of the components, the process, and the final product? Troubleshooting unexpected failure states in complex systems and fixing mission-critical problems that can make or break the business, versus checking out when The System Breaks? Conflicts between book-smart college grads and grizzled vets who know how things "really work" under the hood?

Those are the stories he told me about working in a factory, and frankly the lessons I learned from him were just as valuable as any "fifteen lessons for software architects" books our industry has produced. Sure — software is unique! So is food, so is metal, so is film, so is concrete…

People don't deserve unions because they're 10x workers. People deserve unions because they're workers.

I think a working model of a technology union would have to work closely to professional sports unions or acting unions that support wide disparities in pay. I think historical tech opposition to unions is people not wanting the idea of a standardized wage and extremely high job security (which acts as a downward force on wages). Many people look for something like the Netflix model where you're expected to move on if you're not producing lots of value, but you get compensated more for it. The thing about having 10x engineers being possible is that everyone thinks they are above 1x and deserve above average pay.

> Uh, I'm pretty sure the best employees in any field are worth 10x what the worst employees are, whether you're writing code, assembling a car or gutting fish.

I've worked at FedEx as a package handler before. That is absolutely not the case there. The absolutely best employees might be 3x more productive than someone who's mediocre with 2 weeks of experience. When the job is picking up boxes off of a conveyor belt and putting them in trucks, your walking speed is one of the key limiters. And no one can walk 10x faster than someone else. Not for a 6-8 hr shift.

Some jobs - especially jobs that involve heavy manual labor - just don't have a wide range between the most and least productive employees.

>Some jobs - especially jobs that involve heavy manual labor - just don't have a wide range between the most and least productive employees.

Obviously we're all just throwing around numbers to make a point, this isn't hard data.

But I assure you there are zero jobs where there is no difference between the best and worst employees. Zero.

3x seems like a lot in a business that lives or dies on thin margins.

Specifically in the software development field, a good engineer increases the productivity of all other engineers in a way that allows engineers to increase the productivity of each other in a compounding way.

A “10x” engineer is worth a compounding amount more than their peers.

I don’t think that holds for every field.

Jorge Luis Borges colleagues at the National Library forced him to go slower because he could catalogue books in a fraction of the time they needed.

So yeah, if you're an absolute genius, you don't need a union. Everybody else does. Your common people colleagues do.

Yeah, that’s the trade off with many collectivist ideas.

They optimize for equality, which necessarily hurts the highest performers (in one way or another) and benefits the lower ones (in many ways)

Like anything, it’s a trade off

You would depress the wages of 10x engineers with unions. Look at teacher unions for an example.

Actors and athletes don't seem to have that problem despite having unions.

> Uh, I'm pretty sure the best employees in any field are worth 10x what the worst employees are, whether you're writing code, assembling a car or gutting fish.

Or presumably anything else (because you said "any field"). But why?

It seems like the less skill or expertise something takes the less ability there is to deviate from the average. In which case unions, if most useful for low-skilled labor like OP described, are most useful for industries that tend to outsource their labor to other countries. And thus reduces the need for unions. Seems consistent with the decline of unions in the US.

Jobs aren't all technical skills. You can be the best coder in the world but if you're a colossal asshole you're not going to get put in charge of other people. If you show up to work drunk or high, you're definitely not going to advance.

Never said it was. We're talking about productivity. Being such an asshole that your performance is crippled is baked into that.

Big tech might, but unlike steelworking or auto manufacturing, the barrier for entry for tech is very low, and so big tech aren't the only employers out there. They're not even the only employers paying high wages (you can get good salaries at startups as well)

> Uh, I'm pretty sure the best employees in any field are worth 10x what the worst employees are, whether you're writing code, assembling a car or gutting fish.

This is a big misconception. The difference is whether 10x talent is more valuable than 1x talent. You could be the worlds best fish gutter, but if you can't gut 10x more fish than the next guy, then your salary is capped and you're wasting your talent. In software, you can be 10x more productive than the next guy.

> low skilled jobs that can be mastered in months or years

Routinely on this site, there are job postings for "senior" engineering roles for that do not require even 5 years of experience. (By contrast, the recent headlines about UPS drivers making $170k are for drivers with 5+ years of tenure.) There are frequent discussions in this space about whether engineers need college or other pre-employment education. And many/most "tech" workers are not programmers; they contribute to tech companies in other ways. By and large, our industry as a whole does not place a huge value on long experience. I'm not sure how different tech really is.

> When no one does anything better than anyone else, it makes sense to negotiate as a block, individuals have no power.

How much power did the individual employees at Meta, Google, etc. have when they were being laid off en masse recently?

> By contrast, the recent headlines about UPS drivers making $170k are for drivers with 5+ years of tenure.

They do not make 170k, that was a puff piece from UPS. That number was total compensation + monetary value of all benefits. Which not the way most people/postings talk about compensation.

> Which not the way most people/postings talk about compensation.

This is a tech site. People in tech routinely include the value of e.g. RSUs in compensation numbers. Point taken, but a) it's not too far afield from how tech comp is discussed and b) even if you back out the non-cash components, the drivers still make a lot of money. (Possibly more even than the median US software developer.)

I agree that things like RSUs are generally included, but the 50k of benefits does not include RSUs. The closest this they talk about is the pension plan, which they quote as 23k.

Looks like UPS quotes above 95k avg, and 42$/hr avg after 4 years. I have no idea how that compares to average tech salary, but I agree that it is good money.

That last thing worth mentioning, also from UPS' site is that almost half of their full-time drivers have been at the company for >15 years. That's a long time.

[0]: https://about.ups.com/us/en/our-company/great-employer/real-...

> 50k of benefits

First cut at this would be $23k of pension, plus health insurance. (Average policy for a family of 4 is running around $25k; UPS may have a better plan than average.) It's fair for them to include these because comparable employers do not include either. (Though it make the compare to tech workers less direct.)

> almost half of their full-time drivers have been at the company for >15 years

There's a documented correlation between good benefits, low turnover, and higher profits.

> It's fair for them to include these because comparable employers do not include either.

It is fair, but not under the phrase "UPS workers make". That phrasing usually only refers to (pre- or post-tax) salary.

> There's a documented correlation between good benefits, low turnover, and higher profits.

I would believe it. All I meant to say is that the 95k average salary will be skewed by having half their full time employees being there for >15 years.

I think you and I agree for the most part, I am pushing to keep a clearer distinction between take home pay and total compensation.

> I am pushing to keep a clearer distinction between take home pay and total compensation.

I agree, but even this is squishy. For example; pension benefits eventually are taken home by the employee.

It includes health insurance premiums and stuff paid by the employer. No one uses that for TC.

> health insurance premiums and stuff paid by the employer.

As indicated above, most blue-collar jobs don't carry similar benefits. Totally fair to include in the TC number. It's also a higher percentage of notional wages for the specific type of job. Health insurance premiums for a family of 4 are averaging approximately $10/hr; it would be silly not to include that in a comp discussion where employees might be looking at other jobs that pay $15/hr.

Again, even backing out the compensation you don't want to call "compensation," these are still very good jobs.

This nuance is missing from most discussions about unions in tech; People have a high-level idea that unions have increased wages and improved working conditions in other industries and they’ve made a crude extrapolation that a tech union would therefore give them them more money and fewer working hours. All with no downsides or strings attached!

A more accurate mental model is to understand a union as trading some (or most) of your individual negotiating leverage and advancement opportunity out for a collective negotiating structure through the union. You have less opportunity to negotiate and advance on your own, but you would, in theory, get various protections in return.

This does not mean wages go up! In fact, companies are now incentivized to withhold as much compensation as they can get away with so they have room to negotiate in the next round of union talks. You also have to be prepared to strike to force those negotiations, which means giving up your compensation for that period.

Job hopping is huge in the software industry. It’s well known that job hopping is an easy way to get a pay increase when you’re in bad conditions. I would be surprised if software engineers as a whole would accept long strikes and strict compensation structures when they realize they can actually earn more by jumping out of a hypothetical union job where they’re not getting paid due to strikes and the recruiters come calling with job offers that are conveniently for 5-10% more than their union’s demands anyway.

Many tech workers already effectively have made the salary compromises required by collective bargaining. That’s what salary bands are. Those who think they’re worth more than that become contractors, while the FTEs mistakenly think they’re getting paid the same as everybody else and sleep well at night imagining they have more job security than the contractors do. Creating actual unions won’t change much. The suckers will still self-select into getting paid less, and will be very happy to do so.

At least here in (highly unionized) Finland unions set only the floor on the wages and other working conditions, and especially in white collar jobs practically everyone is paid above the floor.

Some, mostly public sector, contracts have union negotiated "wage tables" suggesting pay based on position and individual performance, but similar exist in almost all large organizations. And the unions don't care if someone is paid more than this, and at the moment they wouldn't even know, because individual pay isn't public information.

I think we're still in such a good position relative to everyone else in society that it's difficult to see the need for collective bargaining. The skills are in demand and not easily replicated. There are enough tech jobs that the solution to almost any workplace issue is to quit and work somewhere else - often for more pay.

> The best engineers are worth literally 10x the worst engineers

Lots of people will claim to be 10x engineers, few people will claim to be average engineers, even though statistically there are more of the latter. I suspect 80% of engineers think they're in the top 20%. This is bad for solidarity.

(Many of the people likely to face discrimination issues are also at risk of being discriminated against by their union, sadly; it's not an entirely simple solution. Again, poor solidarity.)

> pay proportional to value-add

As we're seeing in the Unity debacle, who gets to capture value is about power rather than simply distributing the total value-add among all those that add value. And power in this case is determined by the "BATNA" for both sides. So long as tech workers are difficult to replace, they already have considerable power.

>I think we're still in such a good position relative to everyone else in society that it's difficult to see the need for collective bargaining.

Corps have been pushing down wages for decades. Here's some methods:

  - Offshoring
  - Automation
  - Skeleton teams
  - Long hours
  - Removal of bonuses for anyone but managers
  - Removal of signing bonuses (this was a thing in the late 90s)
  - Recruiters posing mid level people as seniors.
  - No or inadequate inflation raises
  - Coordinated layoffs (flood the market, hire back what you need at deflated prices)
  - Little or no on site skill training (this was also big in the late 90s)
  - A few large, entrenched tech companies.

The math is: total developer hours needed - total developer hours available = surplus and higher wages. When you automate your daily grind, you get more done. When you get more done, the total developer hours needed is less. Multiply this by around 2 million developers and and pushes down wages. They're boiling frogs here.

>So long as tech workers are difficult to replace, they already have considerable power.

You may think that, but the phrase, "no one is irreplaceable," and "the graveyard is full of irreplaceable people," is still true. Go piss off your CEO, call him incompetent or something, and see how replaceable you really are. He will ditch you and make everyone else pick up the load.

This is true; I myself have had a below-inflation pay rise and some colleagues made redundant. I'm still paid more than 95% of people in my country.

What that highlights is just how bad the inequality curve really is, that the 0.1% right on one edge have exceptional power to make decisions which make the rest worse off.

What this means is that any tech union is going to have to extend a bit of solidarity out to lower paid and ""lower skilled"" people against whom we have traditionally been snobbish, in order to be successful. See the Hollywood millionares out donating to the rest of their picket lines, for example.

You can't just look "up" the curve at your salary getting ever higher. You have to look "down" as well, without contempt.

> "anyone is replaceable," is still true

Was it Napoleon who said "the graveyards are full of irreplaceable men"? Also true.

>I'm still paid more than 95% of people in my country.

What makes you think you aren't worth much, much more? What would happen to your company if all the tech people went on vacation for a month? CEO's don't think, "We make more than 95% of businesses in our industry, I guess we can relax now." We have to act like the CEO of our careers. Also, most tech workers in the US aren't in the top 5% of earners. Unless you are in a high position at a large company, you're probably top 15-25%. Might wanna check.

>Was it Napoleon who said "the graveyards are full of irreplaceable men"? Also true.

I remember Jobs saying it, but he was known for adopting other people's sayings.

>You can't just look "up" the curve at your salary getting ever higher. You have to look "down" as well, without contempt.

Yes, tech people's elitism isn't a beneficial trait. It's good when selling yourself, but once that's done, we should drop it.

FWIW, a tech union, if done right, would probably be the strongest union in history. Think Dune and Ix.

> most tech workers aren't in the top 5% of earners

No, I really did check https://www.gov.uk/government/statistics/percentile-points-f... before posting that stat.

(I guess this is where Americans start discovering how badly paid the UK as a whole is)

>I guess this is where Americans start discovering how badly paid the UK as a whole is

Holy moly. In the US you have to earn just over $240K USD/yr to break the top 5%. At least you don't go bankrupt when it's time for elder care.

https://www.cnbc.com/2022/01/24/how-much-money-you-have-to-e....

> Don't unions work best in fields where quality is very uniform and expertise counts for very little? I'm thinking about low skilled jobs that can be mastered in months or years. Unions create salary progression, and benefits for seniority where there is otherwise no reason for them to exist. When no one does anything better than anyone else, it makes sense to negotiate as a block, individuals have no power.

No, this is a common misunderstanding. Note that all professional sports have unions for athletes.

The point of labor unions is that no matter how much power you may think an individual worker has, it's always relatively small and vastly surpassed by the power and money of the business ownership.

> Note that all professional sports have unions for athletes.

Except tennis, where players are basically independent contractors—and not coïncidentally they have the lowest average earnings of any major sport:

* https://www.vox.com/videos/2023/9/12/23870760/tennis-wages-s...

I would not expect a union for tennis players. Why would they even have an employer? They play by themselves, for themselves. I have never heard of a tennis player “owner”, other than obviously the tennis player themself.

And the coincidence that they earn less might have something to do with how nobody can name anyone outside of the top handful of players.

> They play by themselves, for themselves.

The same can be said of actors, and yet they have a union/guild.

> And the coincidence that they earn less might have something to do with how nobody can name anyone outside of the top handful of players.

Few can name most of the benchwarmers in the NBA/NFL/etc, and yet they have a minimum salary: NHL is $775K, NFL is $800K.

Carlos Alcaraz is the top player and made US$ 10M (gross, not net) in 2022: he would be the 147th top-paid athlete in the NBA, 202nd in the NFL, 126th in the MLB.

> The same can be said of actors, and yet they have a union/guild.

Good point. I guess the big difference is that selling the tennis game itself simply does not bring in the revenues that acting and the NFL does.

Top tennis players earn from brand endorsements, not from simply playing the game.

Sports are not typical businesses. You can’t just start a team and hire union workers. Current owner management need to vote on expansion (this is a simplification) it’s an artificial situation.

Unions exist in other fields. They work best when quality is uniform and expertise counts for little. That doesn't mean they do not work when quality is not uniform and expertise matters, just that they don't work nearly as well.

UFC, NASCAR, golf, tennis, WWE don't have unions.

> Note that all professional sports have unions for athletes.

Terrible example though because professional sports leagues are monopolies.

We're not sports players or actors. How about listing unions that contain actual engineers in the US and explaining how successful or not they are at serving their members? There's SPEEA and...?

> Don't unions work best in fields where quality is very uniform and expertise counts for very little?

Actors in the US are in SAG (currently on strike). AFAICT, there is a 'base contract' that most people get ("John Smith as Taxi Driver", "Jane Doe as Waitress #2"), but folks can negotiate above and beyond that ("Tom Cruise as Ethan Hunt").

So there's a floor, but no ceiling.

> Don't unions work best in fields where quality is very uniform and expertise counts for very little?

SAG-AFTRA (the big acting union) blows that argument out of the water. The disparity in compensation between the SAG extras and the big name stars can be several orders of magnitude. Their union works great, as it's mostly about dictating labor-favorable working conditions and contract terms.

There are all sorts of things that the vast majority of software developers could get behind that we often don't see eye-to-eye with management about. Flexible working hours, having blocks of interruption-free time, on-call sucks, etc.

> Don't unions work best in fields where quality is very uniform and expertise counts for very little?

I will do my best to take this as the result of decades of anti-union sentiment across the USA; in places like Sweden this is not "common knowledge" as it seems to be stated here.

The best counter-examples I can think of, within the US even is: the screen actors guild (which is a union) and the writers guild. Both of which cannot (and definitely do not) represent uniform quality and lack of talent or special skill.

Unions have a lot of different forms. SAG-AFTRA is in a field with extreme striation for instance. There's still plenty of individual negotiation.

Why would a 10x engineer object to improved working conditions for all? Some general hatred for coworkers?

Yes. Thinking of yourself as a 10x engineer, you run the risk of seeing everyone else as 0.1x engineers, mere inconvenient ants that you're forced to tolerate. You can see a bit of that in this thread.

I didn't say they would object, just that they aren't incentivized to care about it. It doesn't help the 10xer's bottom line, and the conditions at the "bottom" of tech are some of the best of any profession.

They actually are, unless the 10x engineer thinks good customer service agents, customer deployment engineers, assistants, etc. somehow are not worth having. Usually, the product is way more than just some software engineering piece.

Because Union agreements can bind pay to non meritocratic ways such as time spent in the role or when a higher position opens up. Those top players will rise to the top easier in an open market. It's the same with waitstaff in a tip-based system. The bottom half wants fair wages while the top 20% will leave because they know they can earn without those rules in place.

First, that means there is an open, two-sided market and, second, citation needed that that is really the case. Btw., employers can and do also pay based on non-meritocratic KPIs (but depends on your definition of that term).

I'm not a scientist nor am I trying to get into an actual debate on hackernews. I'm just spouting my opinion and feelings on the matter like conversations were done preinternet when someone could just go digging around to find a study to support their argument. Of course employers reward people for loyalty but as someone who has worked in a union environment before, I've felt stifled. I come from a family of tradesmen and unions do wonders to protect the health and future of employees but they are just as corruptible as corporations. They're just corrupt for the people they protect which happens to benefit the union workers most of the time.

I'm just not interested.

And those 20% won't even be compensated 2-3x higher under a union: likely just in lockstep with their years of service. A union might figure out a couple of tiers within the payscale for specialized or managerial work, but I expect those to be gamed as well by shrewd but mediocre people.

Following up on what you say, the main reason for unions is to help labor extract more of the wealth they create. While I'm sure there are lots of people here who think devs should make more of it, they live pretty comfortably. It'd be like saying doctors should unionize to get a lockstep salary from the hospitals where they work. There might be some marginal benefits to it, no doubt, but they don't as a class seem any more interested in it than devs do.

> doctors should unionize to get a lockstep salary from the hospitals where they work

.. have you actually checked how much power the AMA has, even though they're not technically a "union"?

This year UK doctors went on strike against their monopsony employer because they were getting below-inflation pay rises: https://www.bma.org.uk/our-campaigns/junior-doctor-campaigns...

A union would protect everyone. And nobody is saying that a senior engineer should get paid the same as a junior.

You live under the impression that better workers will earn more.

In practice the skill that earns higher salary is to pretend to be very competent by using jargon and parroting stuff.

Even the terrible developers that are a drain on the company and hence compensation? No thanks

Do terrible NBA players get to keep their jobs? The pithy response is to point at like, James Harden, but on the whole professional basketball is very meritocratic and yet they have a union.

Pro sports in general. Netflix said they run their company like a pro sports team, why not include that players have a union?

> Do terrible NBA players get to keep their jobs?

Sports is different because there’s only a certain number of spots on the field. This changes the dynamics and keeps the crab mentality at bay. The same is not true for most workplaces.

How so? The limited number of spots seems like it would enhance negative behavior, not curb it.

> Even the terrible developers that are a drain on the company and hence compensation? No thanks

Nobody forced them to hire them?

> "In practice the skill that earns higher salary is to pretend to be very competent by using jargon and parroting stuff."

Remember that people who believe things like this would be in the same union as you, be paid just as well as you thanks to union contracts, and expect you to strike on their behalf. Do you really want to be in solidarity with people like this?

So? Everybody else would be too.

Did you also renounced your citizenship because there was someone you didn't like in your same country?

Uh, did you miss that Hollywood is on strike? That Marvel's 3D artists are unionizing? I wouldn't call those "low skilled" jobs, or one where "quality is uniform".

edit: Artistic endeavors where people are passionate about the work to the point where they can be exploited are an excellent place for unions. Within computing the obvious comparative is the gaming industry. The "EA Spouse" scandal in the '00s demonstrated how much people were willing to destroy themselves to get into the gaming industry, and how an employer could abuse that passion.

We seem to have based our ideas of on the excesses of the 1970's.

The idea of a tech union could simply be establishing the very basics. The expectations of the folks I know in the industry include more than 40 hour weeks, significant unpaid overtime during "crunch" periods. Managers promising the world to clients without even consulting the devs. Illegitimate H1-B hiring, and unspoken threats to H1-B workers when they are voluntold to do extra, unpaid work.

Again, I understand the hesitancy, but a tech union could be much more akin to the Screen Actors Guild, where the vast majority of the negotiation is about how the lowest-ranking get a fair shake, while the superstars barely notice the union. The benefits to the superstars, however, are obvious. When you have major players on record of egregious crimes of illegal non-competes a decade ago, which probably cost the superstars literally millions of dollars each over their careers, I would hope they would want to be on board too.

Unions can suck. They don't have to suck. I would think a bunch of habitually online, highly-skeptical, analytic, and systemic-focused folks like us could actively make them not suck.

Unions don’t exist so that people you consider unskilled will get raises. Unions exist so that people who sell their labor can cooperate to influence the conditions they work in.

If you consider nothing more important than your own individual compensation, and have absolute faith in your own irreplaceable genius, it’s easy to make a case against unions — sort of like “there is no prisoner’s dilemma, why would I ever cooperate?”

It's a decent take.

But taking a view on markets outside of the US and SV/bigtech in particular, those 10x devs aren't usually getting the 2-3x as much, and taking a view of industry in total, profits are runaway but wages are flat and falling, especially when considering inflation.

Capital is winning. Perhaps unions are required. But I'm not a union member and likely never will be.

>Capital is winning.

When hasn't it won in the western world?

> When hasn't it won in the western world?

In the US, post-WW2 until Reagan got elected.

They still won just not as much as now.

I agree, us non-capitalists took more of the cake back then.

Unions in tech won't recreate the 1950s. The reason capital always wins is because it can be moved around more easily than specialized workers can change their skill set. If you extract more of the pie in one area then capital (and thus growth) will just move to a different areas. Could be different industries or companies in different countries. They still win while your occupation and growth in earnings stagnates.

Whenever unions are strong.

(Was that a rhetorical question?)

Is that true? I am only a dataset of one, but while my wage at a job may not rise quickly, my raise at a job jump rockets up. We can say that employers should reward employees who stay, but in an industry where getting your next job is trivial (in my experience anyway), I just say sucks to be those companies when I take my institutional knowledge with me. Also it's been my experience that remote work has put real pressure on local companies to pay higher than they did before the pandemic.

I work in the Atlanta Metro, so it's not like it's a place you think of when you think tech.

> my wage at a job may not rise quickly, my raise at a job jump rockets up

There's a well-known plateau in tech wages that affects most tech workers. Look at what your current employer pays people 10 or 20 years older than you for the type of work you do and you will be able to observe the plateau.

> > my wage at a job may not rise quickly, my raise at a job jump rockets up

> There's a well-known plateau in tech wages that affects most tech workers. Look at what your current employer pays people 10 or 20 years older than you for the type of work you do and you will be able to observe the plateau.

I've been doing this for over 10 years. When is the plateau coming? Even lateral moves for me bump my paid up to whatever market is at the moment. Maybe that's just a 5K bump if i left in a hurry, but it's not like I lost money to inflation with a jump. I demoted myself during the pandemic due to family stuff and even taking a lesser title I still got a pay bump.

I'm still of the opinion that enough money exists, but I just haven't ever experienced that my salary has gone down with a job jump I've experienced meager annual raises and skipped bonuses, but I've never not been reset to market rate however many thousands of dollars that is. I think my record for a salary jump was 40K? I stayed at a jump I liked for like 5 years. Even still, I never felt my lifestyle decrease in those 5 years. I was always paid enough and I remember once there was a 12% raise in there to help that we were too far behind market rates.

I sort of guessed you were earlier in your career. The plateau broadly is around $250k in TC, with engineers having to move into management (e.g. change jobs) to earn much more. (Yes, there are some companies where this is not true. But most tech workers do not work at pure tech firms and so this is more aligned with their reality.)

Let's approximate a non-plateau career ladder this way:

- 20s: $150k

- 30s: $250k (+ ~65%)

- 40s: $410k (+ ~65%)

- 50s: $670k (+ $65%)

Does this look roughly like the compensation path for median engineers (people who spend most of their time writing code) at your company? If there are not a lot of people there in their 50s, it is likely that they are not paying median engineers with 25+ years of experience $670k. (This hints at the other plateau, which is that very experienced engineers are often pushed out of the field entirely. Would not want to be a 50+ engineer looking for work.)

To me those numbers look insanely high. Like even the 250k that you're using as a baseline for the plateau, that's around double what is expect for that experience.

I'm about 25 years into my career, and I can definitely say there is a plateau. Yes, my first job switch resulted in almost a whopping 50% increase. Wow! Second job switch was about +25%. Third was +10%. Fast forward a decade, my last two job changes were pretty much even. Maybe 0.1%+ if that. You will get to that point if you don't go into the executive leadership suite.

> Capital is winning.

and a reactionary union isn't really the way to fix "capital winning". Because the idea that capital "wins" is flawed imho. It's not that capital is winning, it's that capital is the part of the equation which has bargaining power.

The reason for that is hard to say, but i am guessing that most workers are not really unique or special enough to warrant having them the same bargaining power as capital. After all, capital is required to establish the job in the first place.

Forcing a union, to prevent capital from "winning" merely makes capital less effective at providing their function - investment and growth of the economy. This makes capital leave.

Capital will always have information asymmetry.

There’s also the fact that the worst case scenario for a capitalist is that they become an employee. The worst case for an employee is they end up homeless.

A “capitalist” can also become homeless. Small business owners have to secure loans with personal assets and bad bets / legal liabilities are abound.

Established business shareholders probably do not have this problem.

If they fail at being an employee, yes.

In addition to the writers’ and actors’ unions, many professional athletes are also unionized, where there is also a great deal of variance in pay and skill between the best and the worst athletes.

Both actors and athletes face a different dynamic, due to the fact that there’s a limited number of spots on the field / the set. This keeps crab mentality at bay to some extent. The same is not true for most workplaces.

Companies do commoditize software development by for example pushing technologies that cater to the least experienced common denominator.

Bondage and discipline languages are a good example of this. It's typically the main rationale for their advocation, although stated perhaps a bit differently.

It follows quite obviously from the profit motive that companies want to minimize salaries, labour bargaining power and dependence on individuals. These, especially the latter, apply to the "10x" as well.

Don't unions work best in fields where quality is very uniform and expertise counts for very little?

Actors are unionized.

Unions work best when you have 3 qualities of the job: a feeling that you cannot be better than someone else (that is quality/expertise don't count past a short training period); a uniform wage scale (that is no way to say one person is better than another); and enough skill that you can't be easily replaced.

Note that the above is how they work best. Unions do work in situations lacking some of the above, but because of that lack there is incentive for someone to leave the union and so they have troubles they must address somehow.

The above mostly lifted from https://pedestrianobservations.com/2016/07/16/public-transit...

Unions work best in industries that have captive audience, that's why public sector, and public-sector-adjacent (e.g. infrastructure construction) unions are the ones that stick around more; no matter how bad they are, the taxpayers can't choose to shop around. Stuff either becomes more expensive (infrastructure), or worse with consumers holding the bag (public schools; consumers do shop around, but it doesn't matter much to the unions because taxpayers still pay).

Or ports are a good non-public example, it's hard to "disrupt" a port. To the extent that industry is competitive, unions would basically destroy their companies... Big 3 automakers are a great example.

Skilled trades have unions.

The old-fashioned way to level this accusation is to say that the union is "Procrustean", named after the mythical Greek King who would chop people's limbs off or stretch them on a rack so as to fit the exact dimensions of his bed.

    For THERSITES & Co., for the weakly and small,
    Who in free competition must go to the wall,
    The plan of PROCRUSTES has obvious charms:
    "Cut 'em down to our standard, chop legs, shorten arms!
    Bring us all to one level in power and pay,
    By the rule of a legalised Eight Hours Day!"

Even if it were true that there are fields where quality is uniform and it were true that unions worked better for these fields, that doesn't mean that unions are bad for other fields.

>Don't unions work best in fields where quality is very uniform and expertise counts for very little?

Absolutely not.

I'm a software engineer in a research & development department (with a headcount of ~200 in an organisation of about 20k). Roles here are very varied, from people with multi-decade experience and very deep specialist technical knowledge to recent graduates, in domains that range from RF engineering to ethnography to project management.

Our union is indispensable, both within our department and across the whole organisation.

> The best engineers are worth literally 10x the worst engineers, but often get paid only 2-3x as much.

If you want to see what happens in a collective bargaining system you can come to Sweden. Here the best engineers make about 1.25x what the worst make. Also, everybody makes a lot less than in the US.

> 20% of the field, with 80% of the negotiating power…

In a free market those 20% have 80% of the negotiating power, true. But in a union they only have 20% of the votes. So what will happen? See above.

It makes sense to negotiate as a block when management has more power than you. Now more than ever, tech companies have more cash, more lobbying and more power. They use this to play games with employees, like attrition through a return to office program so that they can skirt around the WARN Act wrt layoff requirements.

With a union backing them up individuals can more easily:

- Get better at salary negotiation and interviews through having union training - Share how much they earn with other workers, even distributing statistics, reinforcing salary negotiation power and the ability for workers to ensure that businesses doesn't engage in illegal wage fixing - Have union back up when HR are being a pest to protect the company - Ensure that fair severance is distributed when an individual is let go - Negotiate on-call scheduling as a group - Negotiate all other parts of their contract, for example removing pesky no competes - Engage in lobbying

Individuals, even ones with great competence within their field of expertise, have a lot less information regarding the game they're playing with other companies. Unions can help with that.

Note: You can still be in a union without having your workplace be 'unionized' or having a collective agreement with the union. You can still engage in salary negotiations individually, and you may still skip out on being in the workplace's union.

Volkswagen workers are unionized. Their expertise counts for a lot.

Were you unaware that there is a writers strike on right now?

This thread is full of idiots who think they're geniuses.

The NBA Players Association would like a word.

Could you possibly be more elitist?

Unions protect against the power imbalance between the capitalist class (in the classical definition -- those who control the capital) and the labor class (again, in the classical definition -- those who sell work for wage). The skill disparities you describe exist in literally every field; software is not special in this regard.

Reply on news.ycombinator.com