Hacker News Reader

> I'd guess we here on HN are the people best equipped to prevent the worst.

I agree that individuals should be conscious of which companies they support through their labor and/or spending habits.

But I also think that HN overestimates the ability for single engineers to upend the goals of an entire organization. In the real world, if a company wants to automate anything and they’re paying well, there will be a line of engineers out the door applying to get it done.

> So the next time you help someone automate their customer support

I think this strikes at a false dichotomy that occurs frequently in these conversations: There’s an idea that if we simply removed the automated solutions then companies would be forced to replace them with the idealized solutions that we want. In practice, companies know very well that automated customer support and similar solutions aren’t comparable to having a well-paid, highly-trained person pick up the phone. But they weren’t going to pay for the expensive solution anyway.

And the real driving factor isn’t just the companies, it’s the customers. If given the choice between two identical products where the only difference is automated versus human customer support (and associated higher price for human customer support), the majority of customers will choose the cheaper option every time. There are a few people who will proudly pay more for the better CS, but they are a tiny minority. Overall, customers want the cheaper option even if it comes with tradeoffs, and they will vote with their wallets.

I know this may come off as cheeky, but I’m being very sincere: if not us, then who? Sure I alone can’t upend Amazon, but we aren’t all alone and we aren’t all working for the juggernaut that is the Bezos machine.

Maybe I sound naive but every bit matters, especially given how many of us probably work at start-ups. You never know when your decisions will have an impact on a multibillion dollar company that simply hasn’t emerged yet.

The people? Like it or not developing stuff like this is not illegal, and expecting people to grow a conscience when their livelihood depends on not having one is pointless.

The unfortunately cynical answer is that we need to have representatives that will put laws in place to prevent this. It's both simple and complex due to the political landscape, but that's the only real long-term solution.

I think it would be like earlier during industrial revolution when people would vandalize train tracks or factories. It can slow down a bit at particular place and time but otherwise these things are inevitable.

I don’t think it’s particularity fair to compare directing a company in a productive direction to vandalizing/destroying the tech that’s threatening your livelihood.

> I agree that individuals should be conscious of which companies they support through their labor and/or spending habits.

Agreed, but this is a lot of work. After doing it myself, I find that the options are almost always choose the least evil, choose something expensive, or don't buy that thing at all.

Choosing something expensive is a sacrifice I'm willing to make to support a company I believe is doing good. The hard part becomes not buying anything at all when there are no good choices. When I had a car, filling it up with gas felt like a crime against humanity's future.

I feel similarly, and recently I've been focusing on acquiring tools that enable me to disengage from our rotting culture by making things for myself. Of course not everything can be done, but a lot can be, especially if you start letting other people use the tools.

Instead of fast fashion, I'm getting a sewing machine and fabric and just letting several people use it in return for making me clothes (I have money, they have time). Likewise, I'm investing in knitting/crocheting stuff and lending it out. Instead of plastic doohickies, I'm getting a 3D printer: What I make may be cheap crap on par with what I'd get at Walmart, but at least I won't be exploiting or enslaving anyone.

Also buying used and repairing is an option. Basically I like to ask myself 'will this help me need less from big corporations?' before I buy something. Not that small business is inherently more moral, but it's easier to hold to account as an individual consumer. I can actually sue/raise enough of a ruckus to put a local business out of business if they hurt me, whereas Amazon doesn't care.

> The hard part becomes not buying anything at all when there are no good choices

That pretty much describes me and my Android phone right now. I don't want to use an iPhone, but I also want a phone that still fits into my hand and my pocket. Sadly, there's nothing comparable to the 2016 iPhone SE on the market. I could easily get a new phone for free from work, but I would hate to support a product that I actively dislike. So I wait and hope for a small Android phone to be released in the future.

> But I also think that HN overestimates the ability for single engineers to upend the goals of an entire organization. In the real world, if a company wants to automate anything and they’re paying well, there will be a line of engineers out the door applying to get it done.

There's a similar effect in journalism, where the loudest voices saying that individual journalists have the duty and ability to maintain the integrity of journalism are often other journalists.

It's a survivorship thing - the journalists/engineers who stay employed as such are the ones who probably honestly believe in the same things that their employers do, or at least have an output high enough in the things that their employers won't feel the need to heavily edit or nix entirely to have nearly the same effect. It creates an illusion of control.

But ultimately the author of anything is the person whose yeses and nos cannot be overridden, and that's the boss.

> In practice, companies know very well that automated customer support and similar solutions aren’t comparable to having a well-paid, highly-trained person pick up the phone.

And even if it's not particularly expensive in itself, the fact is that having bad customer service reduces complaints. Having anonymous/automated customer service even reduces complaints about customer service - because there's no way to identify or isolate a particular bad customer service experience, like a name. Avoiding the complaints probably creates so much savings that it would be worth automating customer service even if that cost more than having good, live customer service.

We do the same thing in means testing/bureaucracy in welfare/social services. The more difficult or confusing it is to prove one's qualification, the fewer qualified people will have to be serviced. In that case, spending more on "customer service" (i.e. adding more steps) ultimately reduces costs.

> There are a few people who will proudly pay more for the better CS, but they are a tiny minority.

There are also very few ways to know that the customer service on a product will be terrible before you need it. People aren't actively picking bad customer service any more than people choose to watch bad movies at theaters. It's something you find out about after you've already bought in.

Who expects "single engineers to upend the goals of an entire organization". Maybe I am missing something but I could not detect that suggestion anywhere in the parent's comment. Nor did I detect any suggestion that companies would ever adopt "idealized solutions". The comment appears to ask others to refrain from helping companies automate customer support, whereas the reasoning of the reply is something along the lines of "If you do not help the company automate customer support, then whatever customer support the company provides will not be "ideal", therefore, you should continue to help the company automate customer support." WTF.

What is perplexing about these type of replies, which seem to occur anytime there is suggestion of taking personal responsibility for one's actions, is that people even bother to make them. Because if we accept the premise as true that one person's actions can have no effect on organisational change, then why would anyone else care about, let alone try to discourage, someone acting according to some ethical principles. What prompts people to respond.

Perhaps there is something else going one here, especially when these replies often (a) try to point to some other focus, e.g., management, customers, price, etc., besides their own decision-making and (b) do not pressent any alternative courses of action. "It will not make a difference. Thus, keep doing what you are doing." If it does not make a difference then why would anyone care about someone doing it.

Whatever the motivation, maybe someone else can explain it, something about a person on HN who suggests following ethical principles triggers a counter reply from someone who generally tries to discourage this line of thinking/behaviour, alleging something along the lines of "it will not make a difference."

There are plenty of little things that you can add to make even an automated system more humane.

For example, assign a unique ID to every automated decision and display it in the automated replay email. That way, affected humans can at least sue to get the basis of their decision disclosed.

Or why should noreply-support@company.com actually be a no-reply? Let's forward it to the manager who is responsible so that if people are genuinely upset, their "feedback" reaches the correct person.

Or attach a footer to those emails, which contains an email address for legal matters. Make sure to say "Please only contact this address for urgent legal matters" so that everyone correctly infers that it'll reach an actual human.

As for those gig apps, it's probably good enough to just include the actual reason in the JSON reply that the server sends, without ever displaying it in the app. Of course, that's for internal debugging only, but nobody's going to get the production deployment flags set correctly.

Or you can make the decision really obvious, by instantly loading it as an AJAX. Customer fills out a form, the spinner starts, and the result is displayed. "We're terribly sorry but your claim has been denied." But keep the inputs visible, so the customer can tweak the values and submit again. They'll surely figure out what makes the decision flip.

And I wouldn't worry too much about colleagues who'd do anything for money. They will probably go to FAANG for better pay anyway. And if all non-FAANG companies have significantly better service, the overall public will notice.

Bonus tip: If you ever want to reach a human yourself, there are services like hunter.io that'll procure the email address of a human for you. And then you just circumvent all the automation and directly CC the CTO to say "f* - oops - thank you for your great customer service!"

> the majority of customers will choose the cheaper option every time

This is ultimately behind everything that goes to zero. Airlines must compete almost exclusively on price, so everything gets worse to make that happen. We buy goods mostly on price, and so are reliant on China.

Maybe the only choice then is to regulate the profession so that every Tom, Dick and Harry with a two-week bootcamp can’t walk in the door to replace you.

Engineers in other professions are required to behave ethically. Those who can’t get there need to stop calling themselves engineers, until they qualify to use the designation (see regulation above).

> But I also think that HN overestimates the ability for single engineers to upend the goals of an entire organization. In the real world, if a company wants to automate anything and they’re paying well, there will be a line of engineers out the door applying to get it done.

In a lot of countries with socialized medicine, the way to get better treatment for those who can't afford to pay twice (taxes + the cost of private care) is to have doctor friends. Suddenly there's always availability.

Sounds like in the future there will be people for whom technology will just work because they know the real engineer inside capable of taking care of their case, vs the unnetworked who will be stuck dealing with the black-box algorithm.

> But I also think that HN overestimates the ability for single engineers to upend the goals of an entire organization. In the real world, if a company wants to automate anything and they’re paying well, there will be a line of engineers out the door applying to get it done.

So my experience may not be the norm because I've only worked at smaller orgs, but my experience with all my product managers have been collaborative. Personally, I wouldn't be able to work at a place where I can't affect product decisions. But then again, maybe that mythical $300k total comp is really about "shut up and just follow the specs."

Want to argue this point

>>But I also think that HN overestimates the ability for single engineers to upend the goals of an entire organization. In the real world, if a company wants to automate anything and they’re paying well, there will be a line of engineers out the door applying to get it done.

Either the work is difficult, and few people can do it (hence the price tag on engineering salaries) or it isn't. But if it is, then even a few refusals can effectively shut down a project. 'oh theres a line out the door' gets given as a weak excuse from people who have zero threat to their security, and I'd rather never hear it again.

The irony is that it’s middle management that will be the first to be laid off. What globalization did for blue collar, management automation will do with middle management.

Making this a “workers’ issue” won’t lead to anything, specially when it’s gig economy workers. Obviously these AI will make their way to non-gig economy sector, and from there spread like wildfire.

This will only exacerbate socio-economic inequalities to a point of no return (revolution of the precariat aside)

Creepy how it’s all too similar to the short story Manna[1] (substituting fast food for gig economy)

I firmly state my disapproval to engineers working on such systems. Doesn't really work, you always find someone that prostitutes himself.

Production surveillance can be essential for quality control, but you don't need individual surveillance for that and line managers are better at evaluating people.

It is mostly useless managers that want to show of a nice Excel sheet.

> So I got the same wrong parcel again. After waiting a while, I opened up the two parcels, each roughly 10x5x5 inches (25x10x10 cm) large. It was two single pencil erasers.

The monstrosites which reach this scale always amaze me. In case of individuals, when sending an item the main concern is packaging, shipping costs and time. Amazon and others? They just furiously and repeatedly ship oversized boxes filled mostly with fillers, while eco-shaming the customer.

> I'd guess we here on HN are the people best equipped to prevent the worst.

Best positioned, but how well equipped are you? Are you organized enough to collectively refuse such requests, or will you fall into the "If I don't do it someone else will" trap?

> Best positioned, but how well equipped are you?

More importantly: how willing? The HN crowd is known for workshipping tech companies.

why is that a trap?

the "trap" seems to be instead the stable equilibrium state.

It's only a trap if you consider doing things for money that you, yourself, think are bad things, as a bad thing.

You often don’t even know how your tools will be used eventually. Most technology is pretty neutral. How it’s being applied defines whether it’s positive or negative.

> So the next time you help someone automate their customer support…

In my experience people working on these things are well intentioned. So it comes down to a need to shift priorities. Maybe you need to launch something now, and the PM tries to make you delay some important thing for the next release, launching without it at first. That’s the kind of thing we need to change. No more promises of “We are working to make this better” while launching without crucial pieces first. We’re used to making that kind of choice when it’s not people’s livelihoods and working conditions at stake, and we have to make those choices differently when the stakes are people.

Any automated process always needs human fallback.

Or will we live in a world where the computer is always right.

It's unreal the number of times a support person has said they "can't" do something because they don't have an option to do it in their user interface. And like, I understand that they don't have the authority to do it and it's not their fault, but someone at HQ basically decided that something was too expensive or inconvenient for the company and made it quite literally "not an option" and it just becomes a law of the universe that the customer can't get that remedy.

Google, Amazon, and the rest clearly disagree

Yes, and it's a kafkaesque nightmare.

I try to avoid Google for this reason.

Or else what? It's not like companies lose revenue or reputation for relying on cheap frustrating automation. If anything, it's in their best interests to keep doing so (i.e., to dissuade people from unsubscribing or complaining).

Or else we the people make it law.

“I'd guess we here on HN are the people best equipped to prevent the worst. So the next time you help someone automate their customer support, ask yourself: How would I feel if my well-being depended on this? Because for some poor soul, it might. Is there a clear fall-back in case the AI fails horrible? Because, you know, they always do.”

I don’t think we are that well equipped to prevent the worst. Maybe some entrepreneurs are but most engineers develop tools that can be used for good and bad. I think automation provides enormous benefits to society and we should keep pursuing it. To make sure it doesn’t end up with an abusive system is a political problem and should be addressed by rules and laws.

In the last few decades the mantra was that the more money the psychopaths can make the better for society and screw the victims. That has to change. Society overall suffers if we have a few billionaires lording over the working masses.

I am pretty hopeful that we can create systems that balance technological progress with concerns of social welfare. We have made big progress with environmental regulation and I hope we will do the same in other areas.

I think Amazon support forms are difficult to reach by design. It helps keep the traffic down. They are simply not at all interested in people who are not customers, and not especially interested in those who are. There was a time when I got a lot of email from Amazon about things that someone else had purchased. How my email address became associated with their account I have no idea. I tried to notify Amazon but there just wasn't any proper way to do it and after sending an email or two I just gave up. It was more than six months later that I stopped getting such email

I'm rather amazed people purchase single erasers from amazon.

The support links are two clicks away in the footer. “Customer Service” : “Get help with something else”

Or if it was your order, you could nav through “customer service” and then click on the order. Amazon may have some issues, but having a hard time reaching A2Z care isn’t one of them IMO.