Thanks for posting. I wonder if the author's conclusion would differ depending on the subject's relation to this type of incident? For example:
- You provide a service in this domain and are aware from the news that this is getting more common
- You provide a service in this domain and your customers tell you this is annoying them, and the story has amplified that issue, making things uncomfortable for you
- You perceive that you can provide specific value in this kind of situation, as someone who can change the dynamics of the issue with an intervention- or diagnostic-type product or service
- You see this news as a way to highlight the additional value of your offering, for example your industry group or union provides additional leverage by vetting or screening
Just some thoughts & questions as to interest, scope, and leverage outside that particular office.
you're talking about how it might be different from a recruiter's point of view, thats what you mean by "a service in this domain", "your offering", etc? I'm guessing, not entirely sure what you mean, why not just say so?
See how quickly you tried to narrow it down? This is tricky because the scenario seems so easy to contain in a given example, the straw-example-man so to speak. But the point is not the job or intersection of job titles, it's the conversion of scenario into broader principle. The author has started this pathway for us in their post.
Abstraction allows one to theorize and stay big-picture more effectively, so if a different but still relevant individual or position comes into play later, the rules aren't instantly outdated.
I wouldn't limit a message or rule to a position like "recruiter" if I wasn't sure that's exactly who I wanted to speak to, because that just leaks leverage all over the place.
And also, I'm thinking of many different positions, possibly hundreds, not just the poor recruiter... They do get a lot of flak :D
I'm just trying to figure out what you're talking about ("leaks leverage" what? what even is a "services" in the "domain" of... interviewing?), but ok.
You used the word "just" in both replies. It's a word that often comes up when a new way of thinking or a new angle may be discussed. But "just" is almost always that word which demands that past understanding be brought forward. It will block new insight. (It is a great word for communicating, "if it's not simple to understand, and if it's even this inscrutable, then how can it be true" which is a kind of fantasy-fallacy.)
If you have a question about new information, and if it's out of legit interest, it's better to stay away from "just" unless you are firmly on the side of life being mansplainably simple in any given case.
At this point it's too weird to go into the rest though, as if to convince you that I have great ideas. I'm not exactly some kind of employment wizard and I'm not writing for an audience. "Leaking leverage" ought to be metaphorically simple enough to puzzle out with some interest in the problem...
> You don't seem to be writing to be understood, which is of course your right
Okay, that's quite a projection. I think it's better to admit that this is your assumption, maybe even informationally based on someone saying they aren't writing for an audience. This layer of interpretation carrying some emotive personal confirmation is getting in your own way. You feel pain, emotional suffering, look I'm doing my best but this whole topic was derailed by trying to take the original comment off into a single, assumptive concrete example way too fast.
IMO you are also writing for yourself here whether you know it or not. Read it over. Better to embrace that aspect. It may be a new and frustrating type of discussion for you, I get that. If you stay with the abstractions, build up from there without assumptions, I don't think this would be an issue.
> But this behavior – having someone else pretend to be you during a job interview – is so far outside anything that normal people would consider that it’s simply not measurable on the same axis as anything job-related.
I am not so sure that it is rare enough that you don’t need to think about it.
We do know that it also happened with the SAT and ACT college admission tests
Except that something was done to prevent it happening again - the person was almost immediately fired.
The author is arguing that prevention in the interview process would do more harm than good, particularly because there are already processes in place to deal with this issue.
I think the SAT/ACT example is different. With a new hire in a technical field, it would just be so obvious if someone interviewed with a fake.
I don't see how that is comparable. With SAT/ACT once the test taker steps out of the hall there is literally no way to face any consequences. In this case you have to actually show up to work and prove that you are fit for the job you interviewed for or you will quickly be fired.
In my early 20s, I thought technology could solve all of our problems.
In my late 20s, I thought people-processes were the real hard problem.
Now, in my early 30s, I think leaders who build both of the above, but have the wisdom and flexibility to know when to make exceptions, are perhaps the most fundamental key to success.
I wonder how my perspective will change as I complete this decade of my life.
The other logical step in your line of thought is to query what is a problem and what is a success. Finding the right problem to solve to bring success most painlessly is another way to define great leadership.
I've found that designing a programming language is a veritable mountain of compromises and exceptions. A language that rigidly follows rules tends to be intractable for users.
Well, cynicism isn't the total opposite of idealism. You can have an ideal that's compatible with cynicism, such as to create a revolutionary new product, which people will adopt because (here's the cynical part) they are interested in elevating their own status or quality of life.
“There are four reasons why the Cynics are so named.
First because of the indifference of their way of life, for they make a cult of indifference and, like dogs, eat and make love in public, go barefoot, and sleep in tubs and at crossroads.
The second reason is that the dog is a shameless animal, and they make a cult of shamelessness, not as being beneath modesty, but as superior to it.
The third reason is that the dog is a good guard, and they guard the tenets of their philosophy.
The fourth reason is that the dog is a discriminating animal which can distinguish between its friends and enemies. So do they recognize as friends those who are suited to philosophy, and receive them kindly, while those unfitted they drive away, like dogs, by barking at them.”
At some point after that, you'll realize that "leaders" and people who are seen to have the authority and capability to solve problems have a vested interest in perpetuating problems and making them worse.
I'm going to take a stab at one more iteration showing you that org building is the real hard problem. But what that eventually gets you to is that politics and community is the true hard problem, which is often where one starts off.
Going through the same journey and hitting 40 this year, I've learned problems are an illusion. It doesn't get systemically better. Nothing ever is really solved, and the industry is constantly seeing an influx of people in their early 20s who think technology can solve all of our problems.
I think that’s too pessimistic. Problems don’t get solved to the extent that they completely go away, but it is possible to make things better. That’s not to say we’re on a constant slope of monotonic improvement, but with effort and patience, progress can and does get made. And we need those new crops of fresh-faced 20-year-olds to carry on the fight.
I used to share this view, but bitter experience has taught me that people do not change in the aggregate. Individuals may change, but that doesn't affect society at large. There is virtue in expanding knowledge and abilities, such as improving the mortality rate or getting us to the moon, but without a corresponding improvement in societal ethics it's like handing a child a knife. The systems and technologies we improve only increase the potential for chaos like chekov's gun waiting to go off. That's been true throughout history and the last several decades are no different.
Early 40s: I hate all computers and want to send my phone into Lava.
Oh, sorry, real answer. Early 40s: So, I built and sold a consulting company starting in my early 30s. Business is hard. Computers are easy. Business is leadership and being able to build marketing, sales, etc as needed.
Slartibartfast:
Perhaps I'm old and tired, but I think that the chances of finding out what's actually going on are so absurdly remote that the only thing to do is to say, "Hang the sense of it," and keep yourself busy. I'd much rather be happy than right any day.
Arthur Dent:
And are you?
Slartibartfast:
Ah, no. Well, that's where it all falls down, of course.
I figured out clear but wobbly rules were the path to success. I also figured out that sometimes a spreadsheet or a piece of paper is the most advanced technology required. It's a matter of speed, complexity of task and performance.
Two more decades: technology is not the problem, it's a tool, but a lot of people are using that tool inexpertly and that leads to all kinds of problems, some of which can no longer be solved by the application of technology.
It’s sad the author even had to write this article, but a lot of devs do act and think like this. It’s also one of the reasons bullshit bureaucracy seems to be ever growing.
A lot of that is just developer/general employee incentives.
As an employee, it is far more important for me to avoid blame than to be efficient as I will get no credit for marginal improvements/no hit for marginal harms, but will get blamed if there is a very noticeable bad incident.
I am far more likely to have an annoyed boss from the wrong person showing up to work than driving away 100 candidates as the interview seemed hostile.
This situation (interviewing for someone else) is not uncommon in the IT/Software contracting field. Even more common is someone sitting across the room from the interviewee and giving answers. Sometimes through headphones hidden under hair. And I am taking about pre covid times where even if the interview was remote the job was more likely than not in office.
What does the manager do when this happens? I have been there. The answer is simple, I did exactly what I would do if the exact candidate I interviewed showed up but fell short of expectations (say used google extensively, or the questions were theoretical and they were good talkers - happens with QA and Project Manager jobs). If they are borderline, give them a task, see if they are able to complete it. A Manager should do it anyway, interviews are hard. Most companies have probation periods for this reason.
Why does it happen/what’s in it for the candidate/contracting firm?
For the candidate a few weeks or months of experience that they can add to their resume. Do it in 3 or 4 places and suddenly you are a mid level developer. Hopefully they also learn a bit of programming in the meantime.
For the contracting firm any money is money.
Sketchy IT contracting forms routinely do this and worse.
If you use sketchy IT contracting firms, you’re vulnerable to this grift.
It’s pretty common. A contractor attends some degree mill back home and is basically a human terminal. They get some training on how to function and send most of their work to a smart guy who does the work of a dozen contractors.
If you allow remote, it’s 10x worse - the folks are almost certainly working multiple contracts.
I've worked alongside some of the sketchiest of IT contracting firms and I've seen this kind of behavior.
One firm just decided to replace the trained contractors working on my BI team with a bunch of junior devs overnight, no explanation, no warning. The fact that BI management kept employing them says more for the suspected "hello money" I think was paid to get them in first.
There’s one reaction to this, my initial reaction: “that’s just common sense”.
But I think it’s a valuable post nonetheless: it’s succinct, it’s a punchy reminder of something we should all try to remember, and as a colleague of mine says: “common sense is not always common practice.”
Trying to fortify some system against some failure mode overwhelmingly often makes it more vulnerable to some other one. This is what keeps terrorists in business.
A lot of that is just developer/general employee incentives.
As an employee, it is far more important for me to avoid blame than to be efficient as I will get no credit for marginal improvements/no hit for marginal harms, but will get blamed if there is a very noticeable bad incident.
I am far more likely to have an annoyed boss from the wrong person showing up to work than driving away 100 candidates as the interview seemed hostile.
So the calculation is heavily weighted towards fortification for me as an employee, as I benefit not from the enterprise value but do from the value of my boss not being annoyed with me.
At the risk of asking a naive and perhaps dumb question, wouldn't this assertion invalidate the entire product class of P&C insurance, wherein that specific asymmetry is directly and entirely hedged against?
Transacting financial instruments is for most entities a negative sum game. There are transaction costs in the general case.
But we can get more conceptual clarity by modeling it as a zero-sum game: one actor wins, one loses. In general the actor with the asymmetrical relationship to forecasts about future price action wins. Sometimes the smart guy loses because he's undercapitalized.
Watching PHDs in friggin "steering outcomes under low gravity conditions" running home to Gaussian assumptions, the Central Limit Theorem, Fourier math, and VAR would be amusing if it wasn't going to put actual hunger on people who no one ever sent the link to HN. In light of the outrageous concrete human suffering, I think I can stifle my amusement.
This is not directed at the parent, I know nothing about the parent's situation.
But god damn am I sick of rich kids. I thought I would get some catharsis from kicking them around like a soccer ball on the biggest field in consumer technology, but it didn't really make me feel any better.
One guy stomping the shit out of legacies at Ivies for a few years doesn't change anything. This year, next year, ten years from now: paper-failures like @sama will still know the right people.
"I didn't make anything anyone wanted, and I'm still rich as fuck."
At the risk of over-reading a situation I know nothing about, perhaps you could use a break from the SV. I had a friend who grew up blue collar in Europe and with whom I worked with in NYC. He spent a couple years in SF for his career and just about ended up strangling people. I visited him in SF for a weekend and nearly ended up in the same place. Ultimately, he ended up back in NYC and back to his usual happy go lucky self.
For all the criticisms I could make about rich kids in the Bay Area, it is simply that they are _lame_. I'm nothing but a middle class public school educated child of immigrants myself, but I always keep in mind my mother's somewhat cynical advice -- "always remember you will work twice as hard for half as much in this country" followed by "and it will still give you a better life than where we immigrated from."
If I can give you a piece of advice (maybe advice is too strong a word, as this is really just a coping mechanism for me personally), it is to remember that the karmic gears of time tick slowly but inexorably. To be a paper-failure 10 years from now that didn't make anything anyone wanted and to still be "rich as fuck" is a certain kind of Dantean hell in and of itself. Your entire life is the real life version of the Chinese "heaven's ban" where you are surrounded by sycophants who always lie to you to get access to your resources, and anyone else worth knowing would never want to associate with you because they deem you a clown and a fraud. You never learned the skills during your formative years to stand on your own two feet, and now you are too old to do the work that brings at least its own intellectually stimulating and market validated reward, as well as the respect of people in the world who are most worth associating with -- never mind the respect of salt of the earth people!
Perhaps this is my NYC elitism creeping in, but I wouldn't switch lives with such a paper failure SV talking head if you paid me. They'll never have taste, or be able to truly enjoy the finer things in life, or stand on their own two feet, or live in a truly cosmopolitan manner; they'll never have the edge to be as good of an operator as me or the operators I respect. And they will always know that deep down inside they are inferior no matter how hard they posture -- in a sense, I am more free than they will ever be, and if the pinnacle of life is to "live free or die" then they are effectively the walking dead, while I am breathing my own air free and clear.
Where was I going with this rant? Ah yes. Perhaps you need a break from SV because it's only there where these sorts of folks receive any kind of societal acceptance. I'd recommend either the east coast (I love NYC and it's never been more fun), or southern Europe (Spain and especially Portugal have always been nice to me). If you stay long enough in the SV echo chamber, it will warp your thinking and make life seem a lot less sunny than it could be.
In my opinion (and this is having no idea about your personal situation), I view it as just a bubble where you happen to work to make your living, and you can always exit as you please when you need a break or you've had enough. Hope I haven't been too presumptuous with this post and that I've at least been a little helpful.
I'd like to really thank you for taking the time to thoughtfully comment on a kind of pissy remark I made. Just about everything you said ties out with my own intuition and experience.
For context, I spent ~10 years in SV, then another ~3 in NYC (but for an SV company most of it), and now I'm back in San Diego where I'm from. I think you're right on the money that I'm a more than a little over-rotated on the SV worldview. I'm generally a bit more balanced and...diplomatic about it, but when a rough week brings out some weapons-grade snarkiness? I'm clearly at least somewhat wrapped around the axle about it.
It's not so much that other people worked less hard for way more money that grinds my gears: I've lived comfortably on each of 3-ish orders of magnitude in terms of income, net worth, etc, and if I pulled out all the stops I could probably grind a bunch more cash out of my career than the perfectly reasonable amount that obtains. I work way less hard for way more money than a lot of folks around the world and feel a bit guilty about it in fact. Maybe humbled is a better word than guilty.
I think that I spent way too much time around people who attributed success to intelligence, insight and hard work, when it was actually due to intelligence, insight, hard work, and a lot of luck. I know some absurdly rich people who acknowledge that there were at least a few dice rolls along the way, and those folks don't get under my skin. Coincidentally or not, those folks seem to not have a bunch of opinions about how other people should live and work and a megaphone.
Andreessen is not far off with the "software eating the world" stuff, in some sense caring about the culture of the tech scene is a bigger and bigger part of caring about society in general every year. And I think that ultimately what bums me out is that YC/HN was one of my key/formative inspirations when I was getting serious about math and technology. Things like the T-shirt that says "I made something people want" that you got by selling a startup to people who enjoyed the experience really spoke to me.
For all I know sama and seibel are really smart, really solid people, and I regret calling them out on hearsay. I know at least one person who I trust and respect who knows sama personally and thinks the world of him.
But optics matter, and I'm probably not the only person who stopped believing in Santa Clause right around the time that pg handed the reins to the guy who needed a bailout on Loopt, and got kinda morose around the time that the Loopt-bailout guy handed the reins to the Socialcam-bailout guy.
I think ultimately I'm obviously very touchy/sensitive/negative on the topic because YC/HN was a childhood hero that ended up being the same damned nepotism that it was supposed to replace. At least MIT is available to anyone with a YouTube account and some time. YC is the new MIT and much, much more exclusive on the basis of who you know.
That's enough of the pissy part: I've got it insanely good for a guy whose grandparents worked in coal mines and gratitude around that is probably the right thing to focus on. It was a rough week.
Thank you again for such a compassionate and thoughtful reply, you've given me plenty to think about. Cheers.
Oftentimes things that are common-sense at the macro level are really hard to implement or remember day-to-day, and as you mention stories like this can be good reminders to keep perspective.
I.e. when investing in the stock market it's almost common knowledge that silently putting money in an ETF every month will outperform almost everyone, even professional traders.
Overreactions to day-to-day things like corrections or a 1-in-a-million hiring situation break the obvious macro strategy
The investment analogy is a pretty great one in my opinion, and very timely given what I think we're all sort of agreeing is a set of asset bubbles driven (mostly) by "accommodative" monetary policy.
When an asset class is going up and up and up, FOMO can get even serious professionals to go long at a (relatively) high price, and gloss over the risk management. When things correct a bit, a lot of folks realize they hadn't managed the risk and get short "before it gets any worse".
This is probably the most common way to buy high and sell low, and while retail investors probably do more of this than hedge fund managers, hedge fun managers also do it.
Dollar cost averaging into a diverse set of ETFs is the simplest and cheapest thing that gets you highly competitive returns (at least to date). But I think that this is more to do with how much it takes emotion out of the picture than that SPY is like, ideal. Buffet, and Michael Burry, and others have demonstrated that if you're willing to spend years to decades of 16-hour days reading public filings in a drab office, it's not "hard" to beat the S&P. But those people are a lot more dispassionate about their trades than I am, and I suspect than most people are.
What's very interesting to me is that the Byzantine, multi-day, "we can't afford to make a mistake" hiring practises are so normal in the US, most of which allows easy firing without cause. It's a truly bizarre combination.
I work at Google, and I understand it here. Yes, Google can technically fire people very easily, but culturally Google doesn't want to do this. Google doesn't want to be quick to fire full-time employees, generally speaking.
So in practice, policies are in place such that firing people is actually somewhat difficult or lengthy, even though yeah, legally Google could fire almost anyone instantly. So if they hire someone who sucks, getting rid of them would be somewhat challenging just because of those self-imposed policies.
This (designing processes for rare edge cases) is endemic not only in people processes, but in technology-enabled business processes. I've had numerous clients fall into the trap of trying to over-engineer their process flow with validations for every conceivable parameter and variation.
This sort of implementation inevitably becomes a mountain of technical debt that overfits to the current process and reduces operational agility when the current process changes. At some point with business applications you have to trust your people to do the work and view the technology as something that enables that rather than necessarily constrains every dimension of it.
This reminds me of “The Computer Won’t Let Me” problem with many companies’ customer support lines. You call about some edge case problem, like your modem’s MAC addressed was entered wrong by the installer tech (not a real example but whatever). The support agent would like to help you, but your problem is sufficiently off the normal rails that the agent can’t manually correct it. The field is not editable on his screen, and that’s that. The process won’t allow them to fix it. This happens a lot with billing errors. An exceptional outage happens and the agent wants to help but they can’t because the Holy Computer was programmed to not trust customer support to adjust people’s bills. I’m sorry, Dave, but I’m afraid I cannot do that.
True story. Happened over the last couple of weeks.
Me: Kraken, I'd like to recover my old account I created years ago.
Kraken support: Sure, just login from the same computer and IP.
Me: I can't, I no longer have either of those things.
Kraken: How did our customer service go?
Me: It didn't.
Kraken: We can fix it, just login to zoom from the same computer and IP...
Ad nauseam, ad infinitum.
On the bright side, you no longer have an account with a wildly incompetent company. Same IP? Unless you are on a campus with statically allocated addresses or something...
I hope you're making this up. My spidery sense tells me that a random company cannot verify the IP I access zoom from. (Unless kraken was zoom itself presumably)
Kraken is a cryptocurrency exchange so it's plausible that someone's years-inactive account actually is worth a small fortune (to the customer, and to any hacker).
And that's my concern. I can't remember if there's a fractional bitcoin sitting around in a kraken wallet. Even if it's 1/10th of a bitcoin, that's enough to spend a little time trying to get it back.
Some of this is a guardrail against social engineering. They can have an armada of minimally-trained level 1 support telling people they need to plug in their modem for it to work. Anything unusual they encounter can be appropriately reviewed by people who will understand the issue and not suffer from the confused deputy problem.
I blundered my home address when initially setting up my cable internet installation. The installer called from the other house and I redirected him. He told me just to call customer service to have it corrected.
Multiple attempts, there is no way to change the billing/service address of an open account. Web says to call, customer service agent says to use web..
I've set to paperless billing and figure for any service visit, the tech will call me confused from 2 doors down and I will redirect them. Easy enough. Two year going now, so far so good.
Have you tried telling them you are moving from the incorrect address to the correct address?
I can see someone overlooking the need to support address corrections because that is probably a rare case, but customers moving within the company's service area and wanting to have service at the new address surely is common enough that they cannot have overlooked supporting that.
> I can see someone overlooking the need to support address corrections because that is probably a rare case
In my experience working working in tech support at a company that encouraged us to fix whatever we could do like as, service address and billing address changes occured everyday,multiple times a day.
At the time I was there that ISP had fewer than 200,000 customer, wasn't even one of the big players at the time.
Intenode in Australia, based in Adelaide, circa 2004 through 2010.
Hi, I was an internode customer from 2009-2016. Great company, great service, even though being competent with tech I never really used support (I think like 2 times during the time). Still the times I did use it there was always a very positive attitude from support.
I'm still a customer today, although now my service is delivered via NBN FTTH, I don't really see much differentiation in the market.
When I first started working at Intenode, tech support was still solely on the third floor of 132 Grenfell Street, and you would occasionally bump in to Simon Hackett in the lift.
I was still there when iiNet acquired the company, but left before TPG acquired iiNet.
It's cheaper. If they don't let their customer service agents have any autonomy, then they don't have to hire for things like 'judgement' and can just throw warm bodies in a chair, which makes people easy to replace and cheap to employ. It also prevents your cheapness from causing issues; if you hire decent people and rely on them, then it causes issues if those people leave en masse or start siding with your customers over you. "Sure, Mrs. Johnson, I'll forgive last month's bill; they don't pay enough managers to check and I'm quitting tomorrow what do I care?"
I’ve seen it a lot in well-intentioned post-mortems. Some bad code gets pushed to production and causes an outage. The responsible team does a thorough review, and in a bout of self-flagellation proposes a long list of extra processes that add steps and checks essentially amounting to “don’t trust anyone to commit anything right ever.” If a level-headed manager doesn’t step in, this pile of process gets implemented, output drops, and people hate their jobs.
I’ve seen a similar thing with integration tests, as on a team with decent unit test practices production bugs are often a result of bad assumptions about another team’s API. Eventually a mountain of brittle end-to-end tests cripples a team until they get managerial buy-in to rip them all out and the cycle starts over.
This is the dark side / failure mode of checklist-based processes.
Yes, checklists and process standardisation will avoid numerous types of errors and issues. However one of your checklist processes must be to assess the checklist itself, weeding and refactoring the checklist itself from time to time. This includes recognising that the refactored checklist will itself have further issues and shakedown bugs.
Ultimately process is a compromise between what can be kept in working or near-at-hand memory, and what's both sufficient and required for the particular problem domain.
That … seems like a company that isn’t doing postmortems correctly. You’re not supposed to robotically implement every “what would have prevented this”. It’s just there to explore the space of such things and provide your options. “This is so rare and low impact that we change nothing” is a valid option!
The purpose of keeping a record of past postmortems is that you can potentially go back and say “oh gosh, it isn’t actually rare, and so maybe the more elaborate countermeasures are justified”.
I’ve seen the opposite extreme happen, where a devops guy was adamantly against writing a postmortem at all, because “it wasn’t our fault” and “it probably won’t happen again”. Facepalm.
Oh this is definitely me describing a failure mode of postmortems, not an indictment of the process generally. It usually happens with well intentioned junior teams that lack experience/perspective.
I’m a firm believer in learning from mistakes, and making the right trade offs between reliability and velocity. Postmortems are a fantastic tool for facilitating this. As you point out though, the implementation matters and you need to be conscious of the cost/benefit of the recommendations that come out of the exercise.
> The responsible team does a thorough review, and in a bout of self-flagellation proposes a long list of extra processes that add steps and checks essentially amounting to “don’t trust anyone to commit anything right ever.”
If that list of extra checks is automated, i.e. added to your deployment scripts rather than manually followed by a human, then that's great. Every bit of repetitive decision-making and tribal knowledge that you can eliminate from your team's day-to-day functioning is a win.
I see this come up more and more in calendar widgets. They don't allow any manual entry so if you are older than 30 and the year of birth picker is some ultra custom SPOS it takes old people like me literally 30 seconds to enter my birth year.
My guess, though it carries a bunch of other problems that people below listed:
Population density means zip codes are extremely concentrated in NYC, especially in Manhattan. Traveling down the length of Manhattan might have you cover ~25 zip codes in 13 miles. Basically, a zip code isn't a meaningful measure of distance like it might be elsewhere. A radius might also place you into NJ, which could be an issue depending on what you're trying to accomplish: Delivering something from Queens might be significantly easier than from NJ, even though Queens might be further.
That's my guess anyway, though it's odd because nowadays generally looking something up via zip code + radius is actually easier in Manhattan (with a map you can easily see whether it would require coming from Jersey, crossing the park, etc. etc.)
Back in ~2000, zipcode databases cost money, radius search was moderately difficult, and in the US Northeast, you can cross most states their edge-most "relatively major cities" in ~4-5 hours, putting an average distance of anywhere-to-anywhere in the state as ~2 hours. In the extreme Northeast (eg: NY, CT, MA, NH, VT, PA, NJ, DE, etc.) the whole state is ~1-2 hours "big", and likely concentrated in a single (or few) city centers.
Contrast to FL, TX, CA, and we had to immediately support zip-code-ish search (a non-trivial technical and operational burden, as zip-codes are rough-approximates for GEO's, and change pretty frequently) b/c "Couch for Sale in TX" means something completely different from "Couch for Sale in MA".
...you can see that "just tell me what state it's in" hits within ~1hr of a single city center up until you hit TX or FL. So in 2000 it was a reasonable shorthand for "east-coasters" to say: "MA == Boston of course", and "CO == Denver" b/c where else was the internet going to be? ...either relatively close-by, or at your states major population center.
This is totally non-scientific and anecdotal, but I worked for a P2P sales startup out of college, implemented the zip-code search, radius, and update code, and distinctly remember noticing that some competitors (and dating sites!) definitely didn't support zip + radius searching, and they were usually ones that were funded from NY. CA was probably a bit more technically advanced, geo-aware, and wouldn't usually launch w/o zip + radius capabilities... again b/c "Couch in NV/CA", or "Singles in NV/CA" is a completely different scale from "Couch in NY/NJ, or Singles in NY/NJ"
(edit: in the 2000-era!! barely dial-up, and computers in general were decidedly non-rural devices, no cellular networking, and a motorola "two-way" was peak connectivity).
This used to annoy me when I lived in Hoboken NJ. Often “nearest store” searches etc would suggest that Manhattan (or even Brooklyn) were the closest, and as the crow flies that is true. However, as anyone who has tried to get from Hoboken to Brooklyn will attest, even much further distance in the same state is far easier to get to, and was the option I was looking for, since I could not (reasonably) fly to Manhattan.
That said, I’d take Suffern NY over south Jersey, so by state isn’t ideal either.
I think this might nicely dovetail with Taleb's anti-fragility. You get anti-fragility not by planning for every edge case (you'll miss some) but by building self-correction mechanisms and controls in that allow to adjust without waiting for the whole behemoth to move. Give people enough room to maneuver to handle exceptions. If you cannot trust people to do that you might to find different people
Edge cases is where every legislatively mandated process I've ever administered has failed hard. The laws and regulations I was administering worked well for 99% of clients.
But if you were in that 1%, shit sucked, sorry.
Two examples:
1) You're a single parent working part-time as a legal secretary, and in receipt of some form of social welfare payment to ensure you and your children don't end up in hardship.
You are offered full-time hours! Yes! Except because you hit the boundaries/limits/edge cases in the legislation, your net income will only increase by $30 a week, because your income tested supplementary assistances will decrease as your earnings do.
Now, the legislation in play here, ensures social welfare and works okay for 99% of people (when I say "works okay", I mean, gives them money, and doesn't disincentivise work). But that income testing bites you as you approach the edge of the income curve.
2) You want to go hunting in a national park managed by the Dept of Conservation. The usual hunting permit requires that you only use centrefire rifles - rimfires and shotguns are totally banned (to prevent hunting of native birds, especially the tasty fruit fed kererū/kūkupa/wood pigeon[0]). However, you want to hunt red deer and chamois, with a crossbow.
The policies at the time don't account for crossbows (and this was before bow hunting was as common place as it is now). So the department defaults to "No".
In both situations, the broad rules work well for 99% of users. But for the 1% it sucks. The key to the 1% is human discretion and agency.
Good systems, whether legal, business process, or algorithmic, always allow for a human to override. Bad ones don't. "Computer says no" style.
(On the crossbows, a ranger (my Mum, I'll be honest) went into bat for the hunter in question, to determine what was needed for "no" to become "yes", and well, now there's a policy that determines the minimum draw weight for a crossbow to be considered humane - and no barbed or exploding bolts allowed! [1])
Cheers to your mum for that! One of the trickiest parts of changing the rules in any bureaucracy or medium-to-large organization is if the org are lucky enough to still have people willing to stick their necks out to try to make things better for someone else. I admire the courage it takes to do something like she did.
Half way through your first example I was thinking it sounded a lot like New Zealand. And then, it was! In that case, it's also affected by (from numerous anecdotes on twitter et al) that a lot of case managers either don't bother, or aren't empowered to work the system for their clients.
When I was a case manager for WINZ, you tried your best, but ongoing discretionary assistance (Special Purposes Benefit, now called Temporary Additional Supplement) above a certain time period required a manager's sign-off, and then you hoped that your manager wasn't a dick.
One weird edge case administrative process that worked in my (well, my partner's) favour:
The UK introducted legislation stating that people on a specific type of visa would only be eligible for permanent residence if it was issued on or before April 2010. The same legislation said, if this type of visa was issued after April 2011, it oculd only be extended to a maximum of 6 years, and then the visa holder would have to go back to their home country.
Well, there were obviously a bunch of these visas that were issued in the interim period, where the holders wouldn't be eligible for permanent residence, but if the employer company was willing to do so, could keep extending these visas indefinitely.
This seemed so baffling to us that we spoke to multiple immigration solicitors and the Home Office/UK Visas & Immigration support reps (or whatever outsourced org they were then -- Capita? Sopra Steria?) and they confirmed that this was true -- the visa could be indefinitely extended unless the government changed the rules.
The government eventually patched this bug, but IIRC, not before most such visa holders used the extra time to sort their employment situation out. (Opinion, not facts, based only on forum discussion anecdata)
> You are offered full-time hours! Yes! Except because you hit the boundaries/limits/edge cases in the legislation, your net income will only increase by $30 a week, because your income tested supplementary assistances will decrease as your earnings do.
My mother at one point made $25/mo too much for her childcare assistance because she worked some overtime. She lost her assistance for months and ended up several thousand bucks in the hole. Plus joy for me, since when parents can't afford childcare, guess who gets to do it instead? If you guessed usually the oldest kid, here's a prize!
Reminds me of a company where a committee (with just one technical member) decides what JIRA tickets are allowed to be worked on. To minimise $arr_past_fuckups.map(will_happen)
What does that actually get you, beyond cursory verification of your TIN and the customary background check? I look nothing like my valid photo ID, which was taken just over nine years ago. I now have a nose ring and a shaved head, I've lost my facial hair,and I carry forty pounds more muscle than I did in 2013. Lots of folks with nothing to hide don't identify as their gender in their photo ID. How can you tell over Zoom that the ID is real and undoctored?
Other than added effort, it's also highly sensitive data. My barrier to showing an ID is far higher than jumping on a zoom call - even more so since it's so rare, the question would immediately raise suspicions.
It makes me also think of the poorest in our populations. Governments that have support systems for struggling families have put in hurdles to accessing those funds. It has the benefit of preventing abuse of the system but we need to be careful how it may prevent access to those who need it.
Thomas Sowell's theory was that the primary reason most of those systems were put in place is not so much to prevent abuse but to guarantee a level of security/power/income for the bureaucrats.
Hence most of these systems are so shitty, like that both sides of the aisle benefit. Bus shows up only every 120 minutes? One side says "we told you it wouldn't work", and the other side says "at least we tried".
Every stupid process in social services is driven by the NY Post and similar outlets.
Q: “Why did nobody notice the baby is sick?” A: Babies are weighed in every interaction, which is humiliating to parents.
Q: “Why didn’t you stop Sally from enrolling for benefits in 2 counties?” A: Fingerprint indigent beneficiaries, at their expense.
Q: “Why should people on drugs get welfare?” A: A family loses crucial benefits because mom smoked weed.
Bureaucracy builds walls out of process to remove individual agency. But politicians design programs to meet different goals, and social services law was no different.
Very much so, these systems are often needlessly and unfairly adversarial. And I assume quite often the system is not even saving money overall by doing so.
Excellent point. Some people/media/politicians are so terrified that someone might get money they're not entitled to, that they design paranoid systems that are likely to deny people money that they are entitled to.
Netherland recently had the long-running scandal of the "toeslagenaffaire", where the tax service had a tendency to assume people from foreign ethnicity or double nationality might be committing fraud, and gave them additional checks and hoops to jump through, leading to a lot of people to be treated as criminals and have to pay back money that they did have every right to. Was incredibly poorly handled, government fell over the scandal, but the new government consists of exactly the same parties, with the same PM, and I have no faith at all that this won't happen again.
While I agree with the author in principle about human processes, as someone who has done a lot of virtual interviewing lately, I think it's worth rethinking the virtual interview and hiring process from first principles. Much of the virtual interview experience is built on the traditional interview (so much so it's always being referred to as a "virtual on-site"). You think asking for an ID is hostile? I'd argue virtual interviews are already hostile in many ways.
I believe the "Ask for ID" is a strawman that wouldn't be helpful, but there are likely non-invasive steps we can take if we spend more time thinking about it. If nothing else, it may result in a less hostile experience for all involved.
Long ago I worked tech support for a company that supported some mainframe equipment.
Eventually after some incidents they introduced “special instructions” for customers who we had to gather extra troubleshooting data for because of some complex issues they had.
Then a few more customers had special instructions… and a few more.
Eventually these involved things like “Send the account manager an email for each call, this customer is very sensitive.”
After just a year EVERY account had special instructions.
Many people oriented instructions were impossible to follow “Page Jim.” Jim didn’t carry a pager anymore and when he did he would never respond so you couldn’t do step 2.
Some involved gathering loads of memory dumps for issues long solved. Others would have you follow program pointers to memory that didn’t exist so you would start over and over until you gave up and decided failing to do your job was preferable to losing your mind at 4am….
Some instructions were multi page word documents expected to be read every time the customer called. After reading it a dozen times a week it almost became impossible to notice if it had a slight change or not.
One clever account manager trying to make sure tech support “noticed” his special instructions tried using a <blink> tag… entirely ignoring the issue that upon noticing the instructions they would be hard to read. Thankfully the tag failed to work.
As you can imagine eventually there were exasperated executive meetings about how nobody follows the special instructions.
This is the world of protecting against corner cases.
Human systems are not computer systems. In computer systems, edge cases are usually worthy of investigation and rethinking. In human systems, edge cases are to be dealt with when they appear
Those disagreeing with your comment might want to consider that it's very close in spirit and meaning to what JKM wrote in TFA:
Computers don’t have emotions; I don’t need to worry insulting the vast majority of S3 objects when I defensively check integrity every time. But humans are different; when we design a human system around uncommon cases, we do need to consider the ramifications on the majority.
Complexity is a much larger issue in human centric systems than it is in computers.
We can hide the complexity in computer systems even engineer the risks of it out, this is not the case in human centric systems. That complexity will endanger everyone involved.
KISS is a good principle generally, but it's the most important principle in humans systems, sometimes even ahead of completeness.
> when we design a human system around uncommon cases, we do need to consider the ramifications on the majority
This is how I feel every time some new service asks me which pronouns I prefer. I'm happy that some people get to choose the pronouns they feel represent them (or avoid the pronouns they feel disservice them), but it seems like it is both pushing an agenda and adding additional friction to the process. Just leave the option setCustomerPronoun() available without making it a necessary step.
I've chosen a contentious example, but there are dozens of others. Just for another example, if I was born in 1977 (above age of consent), what does the forum software care my exact date of birth and err when I don't want to provide it?
It's not really an agenda to leave the option open - people not identifying with male or female (or fluidly between the two) doesn't harm anybody. Surely for it to be an agenda, it has to be putting somebody else out?
I don't get the issue. Making gender-nonconforming people's options the same level in the UI as male and female is a bad thing? Seems like pretty standard egalitarian design to me.
I don't celebrate Christmas, why not ask me my religion in the sign up form? I certainly don't want to see the Christmas themed website or get the "holiday greetings" newsletter. Is making non-Christian people's options the same level in the UI as Christians a bad thing?
Same with the Last Name field. I happen to know someone who doesn't have one. Why isn't he accommodated?
How about colour blindness? Why isn't there a colour-blind accessibility option in the sign up form? Is making colour-blind people's options the same level in the UI as regularly-sighted people a bad thing?
Wait, so because other people aren't accommodated, we also shouldn't accommodate gender non-conforming people?
Having a separate first-last name is often brought up as a UX failure for exactly this reason, and UX designers often design websites so that colourblindness doesn't hamper usability (in my experience working with designers at big companies). The Christmas one isn't as much of an issue because AFAIK very few people are meaningfully put out by seeing Christmas decorations on websites from majority-Christian countries. In other countries these things often /are/ turned off depending on cultural sensitivities.
We should be trying to design our processes to fit the world around us, not rejecting the parts of the world we don't like.
To be fair, the equivalent solution for the name field example would be solving it by insisting everyone specify how many names they have before letting them enter their name - there's a reason that isn't the recommended solution, people don't want to deal with minutia that doesn't matter to them, one field works fine for everyone.
The obvious solution is to just not use pronouns - it's a messy part of the language that is currently in flux, so why wade in?
Obviously, some sites are dedicated to these issues, so they can justifiably ask on sign up, but if you don't _need_ to care about people's personal details, better not to ask at all.
There was a time when asking for religous affiliation was fairly standard, though also often associated with prejudicial practice.
Forms of address --- Mr., Mrs., Miss., Ms., and often professional titles (Dr., in German Ing. (engineer), esquire (lawyers), Reverand, etc.) is at least fairly common if not entirely standard practice.
I suspect, again, some motivation on the part of a requestor. A magazine's circulation department, for example, might want to know the number of lawyers and doctors among its subscribers as a proxy for advertising value.
Many business information request forms provide detailed rosters of who you are, what you do, and your company title. For similar reasons, I suspect.
(I also feed those bogus information as a matter of course.)
For me - it becomes an agenda when they’re collecting the information for no other reason than to collect the information - or - they’re using the field to show how progressive they are. Unless gendering language is somehow critical to the operation of the software - why collect it at all? I have the same feeling about birthdays - I’m automatically assuming it’s because you have “sell customer data” on the 5-year business plan.
Some of us are still old enough to remember when "Ms." was a contentious or at least novel title, though few are old enough to have seen its original proposal in 1901.
The New York Times formally adopted use of "Ms." as a title, distinct from "Mrs." (a married woman, often referred to by her husband's full name, e.g., "Mrs. John Q. Smith"), or "Miss" (an unmarried woman or girl, addressed by her given and maiden names, "Miss Jane Q. Jones"). Ms. Magazine was a direct and deliberate challenge to that practice, and was launched in 1971 by Gloria Steinem. Interestingly, all three words, "miss, "missus", and "mizz" originate from "mistress", which was at one time the single title applied to any woman, adult or child, married or not.
(For men, the terms "Master" (unmarried child) and "Mister" were both represented as "Mr.".)
And then as now, "Ms." resulted in much gnashing of teeth, changing of forms, and updating of databases.
An issue is that many systems presume pronouns, and for some people, this is decidedly uncomfortable and significant. For others ... not so much.
There's virtually always the option to leave the option blank, or to make up a garbage or meaningless value. The first system on which I recall the option being offered was Google+. My response was "trans-krell", playing of my pseudonym's character.
For age, I usually try to find the earliest possible birth year acceptable to the system. For Google this seems to be about 140 years prior to the present date. Again, I avoid providing this information if possilbe (most of my various little-used Google accounts have either no value provided or a ridiculously early age).
I'm of the view that we don't need to be formulated, sprawling on a pin, within some global surveillance database(s). If I can evade classification and feed garbage to the system, I will, for as long as that is viable, and probably for some time after that point.
The gendering or nongendering agenda concerns me far less than the Total Information Awareness agenda. Services demanding anything from me other than some random username and password (I tend to use password generators to create both values), and possibly a contact email address ... tend not to get used.
As I just commented to a friend a few days ago, I can't remember the last time I did create an account, with the exception of some recent Mastodon and Diaspora* migrations in the past year or two.
For my most recent Android device (the Android aspect of it being among the least attractive characteristics), I bailed out of Google Play Store registration, which requires creating a Google account. (Even if not formally associated with other identities I have, those could all but certainly be trivially linked.) Instead I'm relying on F-Droid, APK-Mirror, and the Aurora Store. I've kept actual app installations to a bare minimum, and most of those through F-Droid. There are I think three apps with actual accounts associated to them, though only one has been so configured.
My use of the Internet dates to the 1980s. I've seen a lot. And am disliking increasing amounts of it. Read Dan Geer if you haven't recently.
> There's virtually always the option to leave the option blank, or to make up a garbage or meaningless value.
Right, but it's an uncommon option that relates to a controversial subject. That's why I mentioned to leave the option setCustomerPronoun() available without making it a necessary step: show it in the UI options but it doesn't have to be front and center in the sign-up form.
> The gendering or nongendering agenda concerns me far less than the Total Information Awareness agenda.
Sure, without a doubt. I may have inadvertently picked a flamebait example!
For some reason I'd always been under the impression that the 'F' in 'TFA' was a different word. This makes so many exchanges less hostile in retrospect. Oops.
This, right here, is why I never use acronyms with the sixth letter of the English alphabet anywhere in them. Heck, as you might already have noticed, I don’t use said letter at all, unless it is part of a word.
This is one “human corner case” I have no shame working around. Wouldn’t want to convey hostility where none exists.
The F-word (I'll spare your virgin eyes), or even the letter F, is not capable of conveying hostility on it's own, so there is no sense in omitting it.
The acronym was coined with the non "fine" meaning for that letter. It was much later that internet culture got more polite and people retroactively replaced the meaning on RTFM and RTFA.
I'm almost certain what you're thinking is the original usage. TFA (disparaging, implying the person you are responding hasn't read it) was co-opted as TFA (neutral)
I prefer "TFA" to "OP" as the latter may be ambiguous as to whether it refers to the article or, more typically in my experience, commentary on it. Even here, "OP" might reference either a thread root or the parent of the post immediately being replied to.
"TFA" is a well-known and long-standing usage within the hacker community. It came to prominance at Slashdot, which I'd read back in the day, and is at least some decades old.
It refers specifically to the submitted article, and carries an overtone, sometimes tongue-in-cheek, sometimes subtle, sometimes more blunt, that the person being referred to ought to read the specific submitted work more closely. Perhaps at all. And without crossing HN's guidelines against specific accusations.
That is, the meanings are similar but different. "TFA" is more concise and specific. All of which make it the more fabulous ;-)
It turns out I do use both terms fairly frequently, as my comment history shows. The distinction is largely as I've described above. "The article" usually refers to some additional or other reference rather than the HN submission. "TFA" in the context if "you/I should have read that closely".
I presume they mean that out in the real world, there are lots of edge cases where reversing a transaction is a better outcome than chest-thumping about "Code is Law".
Yeah and the way I described it every "skeptical instructions" was a catch ... that triggered on every call.
Unlike code that triggers (hopefully) only when something weird happens, the cost of the special instructions and the dysfunction that followed was ultra high.
This sounds insane and I genuinely don't understand why anyone would work here or put up with this.
I understand the main point, I think, of 'make sure that people read instructions instead of ignoring them', but this is taken to bizarro levels which I don't understand why anyone would tolerate. The 'account managers' designing these 'special instructions' are either sadists or just assholes, and if management is not willing to protect its own staff from this bullshit, I see no reason for anyone to continue to work there. I don't really understand people staying in abusive job situations like this. There are many other job opportunities out there.
> I don't really understand people staying in abusive job situations like this.
I'm working to get out of a job that turned into this once our (lovely) manager bailed and exposed our team to the incompetence of our current leadership.
I'm having to HEAVILY fight my gut instinct by leaving, because I was raised in poverty and grew up as a disabled lesbian back when those things created more issues than they do now. I'm also from a rural area and neither of my parents have college degrees.
I may logically know that I (as someone with credentials and a decent skill set) have every right to leave and it's in my best interest to do so, but it conflicts with literally decades of messaging I've got telling me I'm lucky to be allowed where I am at all and seeing those around me be punished for standing up for themselves.
I was also abused by my parents, and that's another angle: People who are used to abuse don't think it's that bad, and I'd wager there's a fair number of people who have just never had a decent work situation. If you're going to be abused, at least pick an abuser you know and can work around.
Humans are pattern recognition/social machines, and if you only expose humans to dysfunctional patterns, they'll assume the problem is them. If you design a society, it's pretty easy to get yourself a group of people who are open to being exploited.
I can relate to this. Neither of my parents have college degrees and we were poor growing up. For a while they just pushed me to get any job I could. I worked in a call centre (horrible), door to door sales (more horrible), a butcher in supermarket (slightly horrible), working in the NHS (terrible pay, lovely people).
I've had a lot problems myself struggling around guilt about the pay and general lifestyle in tech. But I'm now working in a tech company with lovely people, doing interesting work. For what it's worth from a stranger: I wish you luck in overcoming your inner critic and that you are worthy and deserving to get a job with people you like at a job you enjoy.
> For a while they just pushed me to get any job I could. I worked in a call centre (horrible), door to door sales (more horrible), a butcher in supermarket (slightly horrible), working in the NHS (terrible pay, lovely people).
This is a large part of it, especially as someone whose community is likely to have those sorts of jobs. If everybody around you hates their job and you grow up in a culture where people actively despise their workplaces/management, you assume that's just how jobs are.
EVERYBODY I grew up around hated their jobs, so it's really hard for me to know what to put up with. To people like my stepdad, my displeasure at how I'm treated by leadership is complete stuck up whining: I personally make more than him + my mother and their households ever have, I'm not in physical danger (unlike him doing HVAC work, my mom doing warehouse work, my siblings' SOs who work landscaping, etc), etc. I was taught growing up that basically unless your boss put you in the hospital, you were lucky to not be homeless or starving and work just sucks.
This is a particularly hard mindset to get out of because in white-collar environments, you can't talk about how growing up blue collar or working class makes you anxious because a lot of 'professionalism' boils down to 'don't let the nice people know you're a peasant or you'll be kicked out for not being a culture fit'.
The pandemic has been very helpful, ironically! I'm more than willing to blame myself for all my life problems, but when I see OTHER people being treated like I was, it raises my heckles and pisses me off.
I'm so glad to hear that you found a job you like. I'm hoping to move from a small-shop/journeyman dev to working in a dev team and I'm trying to maintain optimism, so stories like that help!
It's scary how two different people can have such similar backgrounds (minus the lesbian part, that's playing the rural dating game on hard mode for sure). For me I just kept living like a college student and saved most of my paycheck until the adult part of me was able to convince my inner child we were, in fact, safe.
Feeling safe let me leave bad employers in the past. Thankfully my current one is pretty awesome.
Yeah, I was on the track to do what you did, and then I got multiple sclerosis during my last semester of graduate school. Disability is a major player in my own personal negative patterns: My MS came out of nowhere, I can't make good choices to make it go away, and I can't just ignore it. I also was fucked over when I tried to plan for things like student loans: Over half my undergrad debt is from my last year because my dad lost all the money he was using to help me in 08.
My personal experiences, outliers though they are, tell me that change is likely to screw me over. I would love to save my paychecks but when your meds cost 300k+ a year for your entire life, your life is dictated by health insurance and care. I lived like a college student and saved and was STILL fucked by change, so I'm very change and risk averse.
>I genuinely don't understand why anyone would work here or put up with this.
While the special instructions incidents were horrific, particularly to someone like me who wants to do the right thing and their brain seizes up with such insanity, they were eventually resolved.
Finally they instituted a policy where instructions were reviewed by a group of senior account managers every 90 days. Whomever wrote them had to justify them, or they were simply deleted. That actually turned out to be a great process as many out of date instructions were auto deleted without anyone doing anything.
As for the job, it was actually a great job. The support team was amazing, supportive, it paid great, and the company weathered a lot of economic downturns easily. If anything the special instructions incident demonstrated how even a great organization can go bonkers insane.
I remember learning that the reason school busses don’t have seat belts is that as a transportation method, busses have such ridiculously low mortality rate that you don’t want to change anything. Almost anything you might do could only make things worse.
I feel like I've experienced similar situations with post mortem and sprint retrospective processes. Sometimes shit goes wrong in a sort of unique way and people go too far with wanting to have the outcome always be "we will do X to make sure Y never happens again", even though X is more costly than Y in terms of engineering effort and business value. So I think this is also applicable to software design.
> We could ask candidates on video (or in person) to see a photo ID and match the ID against the resume. But this would seem very weird. It starts an interview off in a hostile manner, and send the a strong message of distrust. Honest candidates – which are, remember, the vast majority – will wonder why the heck this company is acting so weird, and will rightly see this as a red flag about the company culture. There will be negative consequences for your hiring practices.
This sort of verification is already being used for some KYC processes by financial institutions (i.e., take a live video of yourself holding your ID that’s uploaded for verification). It’s probably a matter of time until this is so normalized with virtual KYC that people wouldn’t care much when asked to do the same by a potential employer. The process already excludes certain people (like the transphobic process mentioned in the article).
There is always a cost to the company for an incorrect hire, regardless of the underlying reason. So I would expect company HR teams to read that (widely shared) post and formulate strict identity checking and recording mechanisms so that they can absolve themselves from what may seem like an oversight or error. It also fits well with the typical HR style that’s more about controlling employees than about enabling/helping them (apologies to any good HR folks who are out there who struggle against the weight of the systems).
It’s probably a matter of time until this is so normalized with virtual KYC that people wouldn’t care much when asked to do the same by a potential employer.
Though as it becomes more common, it's probably also a matter of time before software is available that lets someone hold up a green card and it's replaced with the document of their choice on the video feed.
Of course it's an arms race, but they already have countermeasures in place: asking you to partially obscure the ID with your hand, asking you to show how the holograms change as you tilt the ID, etc.
I'd think that's all relatively easy to compensate for, just film (or create in CGI) the card held in different directions, then merge into the green screen card.
Seems like it'd be better if chip card readers became ubiquitous on home computers and mobile devices, then we can scan our drivers license or other government ID to prove that we have possession of the physical card (and when making online purchases we can scan a credit card instead of typing in a number that can be stolen).
When I read the original article, my first thought was – there are so many companies that don't mandate the hiring team interview their own candidates. Applicants are interviewed by random engineers from anywhere in the company, get accepted, and are then matched with a team. How do they even detect this case? And I'm talking the likes of Facebook and Google, not random small companies.
As long as results happen, things move forward, are we not all happy?
What if the remote role was ostensibly being done by one person, but in fact they delegated to a team of people? A team of people that themselves come and go but all get managed into the same bodyshop? Does that matter as long as the work's getting done? What if, to mitigate potential loss of IP, a third-party information security policy needed to be agreed? Is it then turtles all the way down?
Trust matters in relationships. Finding out that a working relationship started off with a lie is going to poison that relationship. I’d forever be wondering what other as shortcuts and risks they are taking. Are they taking shortcuts like incorporating code we don’t have rights to into the product? Will they some insecure shortcut and leave me vulnerable? More importantly, what would my customers think of they knew that the company they entrust their valuable data with was willing to continue to employ a person with such ethics?
The article makes some good points, but it also shows that the author doesn't have a lot of hiring experience on the closing end (after the candidate has passed interviews).
"For example, anyone who goes by a name that doesn’t match their government ID could be forced into an uncomfortable explanation"
I've seen this happen more than a few times, and the explanations are seldom anything you'd have sympathy for when discovered during a background check. People with good reasons will tell you up front.
Why not just take a picture? If you're interviewing so many people that there's a genuine risk of forgetting who is who, a picture seems useful. I'm sure I'm not the only person who will forget your name in 3.5 seconds but never forgets a face.
Maybe this is my time at Pivotal speaking, but your first day on any team you get your picture taken with a polaroid. It goes on the pairing board. It's not weird.
"Thank you for interviewing with us! Mind if we take a picture so we can keep our who's who straight when the team is making the decision?"
All that jazz with IDs sounds unfriendly, sure. A snap isn't. Hell, make it a selfie with a team member.
That seems like it would be incredibly hard to perform convincingly and if you could, you've found your true calling in hosting/announcing/commentating live events not sucking at software gigs.
I think in this case it was all remote. But yeah, I don't think it would be that hard to add protection against this fraud in ways the candidate wouldn't even notice.
This is what we do. We literally take a screenshot when interviewing a candidate and ask them to turn on their cameras during meetings (at least the first week or so)
While I sort of agree, the potential impact of someone smuggling in a bomb in their shoe onto a plane is obviously overwhelmingly larger than the impact of hiring a fake cheater guy for your corporation.
Part of why the answer for hiring fakes is 'do nothing' is because the downside there just isn't that bad, it's several orders of magnitude less destructive than "plane with dozens of people on board explodes".
So why stop at airplanes? Why not subway trains, elevators, Starbucks?
Is it because someone tried in an airplane once? Tried and failed? Would you change your mind about Starbucks if someone tried and succeeded?
The whole point of that article is that you can’t design around extreme edge cases. If we used shoe bomb security theater logic around all aspects of flying, there would be no airline industry.
Sadly, for some people out there, any part of the process not swaddled in shoe bomb security theater is an opportunity to kill dozens of innocent people just waiting to be taken.
I've interviewed at places where the hiring manager would give your name to the front desk, and the front desk would ask for ID when you showed up and signed in. This never struck me as weird, maybe because the front desk checking your ID was standard practice in NYC by that point.
It's part of standard building security in numerous large cities.
There were a number of incidents in which attackers gained access to office towers with deadly effect. One that comes to mind is the 101 California shooting in San Francisco in 1993.
That's likely a building security practice, and the hiring manager never sees your id. That's a bit different from the hiring manager themselves asking for it before even being willing to talk to you. And building security is likely to let you through with a convincing story if you know what name they're looking for, after noting the name on your id for reference in case there's a problem.
At any rate, the article is right, this probably causes problems for some candidates (especially, but not exclusively, trans people). But if it's common practice in NYC it's likely most local candidates are used to navigating it. That's less likely elsewhere.
The advice here seems to correspond fairly strongly with those of statistical process control, the method championed by Walter Shewart and W. Edwards Demming.
Processes are modeled as having "normal" and "special" causes of variance. In managing processes, there are risks involved in overmanagement of either, though the Wikipedia article linked below doesn't seem to mention these.
For special causes, building too many special-case checks can create an ossified process. For normal causes, a problem may arise that in attempting to manage what is essentially normal random variation, additional variance is added to the system, or management processes themselves inject further variation or failure modes.
For an example I only just ran across, the Fermi 1 nuclear reactor meltdown incident near Detroit in 1966, and event which gave rise to a book We Almost Lost Detroit, and a Gil Scott-Heron song of the same title. The cause of the meltdown was determined to be "zirconium metal plate that was installed in the reactor as a safety measure", according to a Detroit Free Press article on the incident. Not the first or last time safety equipment has contributed to an accident --- think of the 737 Max and its MCAS system failures, or the thermal insulation cladding of the Grenfall Tower which precipitated a disasterous fire.
I may be mis-recalling or mis-understanding Demming's points on process control, and if anyone could help nudge this the right way I'd appreciate it. I'm pretty certain there is a connection however.
311 comments
[ 3.1 ms ] story [ 282 ms ] thread- You provide a service in this domain and are aware from the news that this is getting more common
- You provide a service in this domain and your customers tell you this is annoying them, and the story has amplified that issue, making things uncomfortable for you
- You perceive that you can provide specific value in this kind of situation, as someone who can change the dynamics of the issue with an intervention- or diagnostic-type product or service
- You see this news as a way to highlight the additional value of your offering, for example your industry group or union provides additional leverage by vetting or screening
Just some thoughts & questions as to interest, scope, and leverage outside that particular office.
Abstraction allows one to theorize and stay big-picture more effectively, so if a different but still relevant individual or position comes into play later, the rules aren't instantly outdated.
I wouldn't limit a message or rule to a position like "recruiter" if I wasn't sure that's exactly who I wanted to speak to, because that just leaks leverage all over the place.
And also, I'm thinking of many different positions, possibly hundreds, not just the poor recruiter... They do get a lot of flak :D
If you have a question about new information, and if it's out of legit interest, it's better to stay away from "just" unless you are firmly on the side of life being mansplainably simple in any given case.
At this point it's too weird to go into the rest though, as if to convince you that I have great ideas. I'm not exactly some kind of employment wizard and I'm not writing for an audience. "Leaking leverage" ought to be metaphorically simple enough to puzzle out with some interest in the problem...
You don't seem to be writing to be understood, which is of course your right, but I feel foolish for trying to understand now.
Okay, that's quite a projection. I think it's better to admit that this is your assumption, maybe even informationally based on someone saying they aren't writing for an audience. This layer of interpretation carrying some emotive personal confirmation is getting in your own way. You feel pain, emotional suffering, look I'm doing my best but this whole topic was derailed by trying to take the original comment off into a single, assumptive concrete example way too fast.
IMO you are also writing for yourself here whether you know it or not. Read it over. Better to embrace that aspect. It may be a new and frustrating type of discussion for you, I get that. If you stay with the abstractions, build up from there without assumptions, I don't think this would be an issue.
I am not so sure that it is rare enough that you don’t need to think about it.
We do know that it also happened with the SAT and ACT college admission tests
https://www.nytimes.com/2019/04/12/us/college-admissions-sca...
Where parents hired a person to take tests on behalf of their children.
Now that word is out about this technique, if you don’t do anything to try to prevent it, it will happen more.
The author is arguing that prevention in the interview process would do more harm than good, particularly because there are already processes in place to deal with this issue.
I think the SAT/ACT example is different. With a new hire in a technical field, it would just be so obvious if someone interviewed with a fake.
In my late 20s, I thought people-processes were the real hard problem.
Now, in my early 30s, I think leaders who build both of the above, but have the wisdom and flexibility to know when to make exceptions, are perhaps the most fundamental key to success.
I wonder how my perspective will change as I complete this decade of my life.
First because of the indifference of their way of life, for they make a cult of indifference and, like dogs, eat and make love in public, go barefoot, and sleep in tubs and at crossroads.
The second reason is that the dog is a shameless animal, and they make a cult of shamelessness, not as being beneath modesty, but as superior to it.
The third reason is that the dog is a good guard, and they guard the tenets of their philosophy.
The fourth reason is that the dog is a discriminating animal which can distinguish between its friends and enemies. So do they recognize as friends those who are suited to philosophy, and receive them kindly, while those unfitted they drive away, like dogs, by barking at them.”
Oh, sorry, real answer. Early 40s: So, I built and sold a consulting company starting in my early 30s. Business is hard. Computers are easy. Business is leadership and being able to build marketing, sales, etc as needed.
Which is a useful concept to grasp!
If you get it, maybe you learn to drop your ergo and focus.
Nothing really matters, but feelings are real in the moment. So maybe you learn to do the right thing and have the courage to be a decent person!
Or you get fat and complain about the kids of today!
Does a individual thread's identity count? May be it does may be it doesn't.
Arthur Dent: And are you?
Slartibartfast: Ah, no. Well, that's where it all falls down, of course.
reminds "the severity of Russian laws is alleviated by the lack of their enforcement."
As an employee, it is far more important for me to avoid blame than to be efficient as I will get no credit for marginal improvements/no hit for marginal harms, but will get blamed if there is a very noticeable bad incident.
I am far more likely to have an annoyed boss from the wrong person showing up to work than driving away 100 candidates as the interview seemed hostile.
Every time there’s a fire, they want to build a new fire engine.
What does the manager do when this happens? I have been there. The answer is simple, I did exactly what I would do if the exact candidate I interviewed showed up but fell short of expectations (say used google extensively, or the questions were theoretical and they were good talkers - happens with QA and Project Manager jobs). If they are borderline, give them a task, see if they are able to complete it. A Manager should do it anyway, interviews are hard. Most companies have probation periods for this reason.
Why does it happen/what’s in it for the candidate/contracting firm? For the candidate a few weeks or months of experience that they can add to their resume. Do it in 3 or 4 places and suddenly you are a mid level developer. Hopefully they also learn a bit of programming in the meantime. For the contracting firm any money is money. Sketchy IT contracting forms routinely do this and worse.
It’s pretty common. A contractor attends some degree mill back home and is basically a human terminal. They get some training on how to function and send most of their work to a smart guy who does the work of a dozen contractors.
If you allow remote, it’s 10x worse - the folks are almost certainly working multiple contracts.
One firm just decided to replace the trained contractors working on my BI team with a bunch of junior devs overnight, no explanation, no warning. The fact that BI management kept employing them says more for the suspected "hello money" I think was paid to get them in first.
Don't such short employment stints speak against a candidate?
But I think it’s a valuable post nonetheless: it’s succinct, it’s a punchy reminder of something we should all try to remember, and as a colleague of mine says: “common sense is not always common practice.”
Trying to fortify some system against some failure mode overwhelmingly often makes it more vulnerable to some other one. This is what keeps terrorists in business.
EV[failure] = P(failure) * cost-of-failure.
As an employee, it is far more important for me to avoid blame than to be efficient as I will get no credit for marginal improvements/no hit for marginal harms, but will get blamed if there is a very noticeable bad incident.
I am far more likely to have an annoyed boss from the wrong person showing up to work than driving away 100 candidates as the interview seemed hostile.
So the calculation is heavily weighted towards fortification for me as an employee, as I benefit not from the enterprise value but do from the value of my boss not being annoyed with me.
But we can get more conceptual clarity by modeling it as a zero-sum game: one actor wins, one loses. In general the actor with the asymmetrical relationship to forecasts about future price action wins. Sometimes the smart guy loses because he's undercapitalized.
Watching PHDs in friggin "steering outcomes under low gravity conditions" running home to Gaussian assumptions, the Central Limit Theorem, Fourier math, and VAR would be amusing if it wasn't going to put actual hunger on people who no one ever sent the link to HN. In light of the outrageous concrete human suffering, I think I can stifle my amusement.
But god damn am I sick of rich kids. I thought I would get some catharsis from kicking them around like a soccer ball on the biggest field in consumer technology, but it didn't really make me feel any better.
One guy stomping the shit out of legacies at Ivies for a few years doesn't change anything. This year, next year, ten years from now: paper-failures like @sama will still know the right people.
"I didn't make anything anyone wanted, and I'm still rich as fuck."
For all the criticisms I could make about rich kids in the Bay Area, it is simply that they are _lame_. I'm nothing but a middle class public school educated child of immigrants myself, but I always keep in mind my mother's somewhat cynical advice -- "always remember you will work twice as hard for half as much in this country" followed by "and it will still give you a better life than where we immigrated from."
If I can give you a piece of advice (maybe advice is too strong a word, as this is really just a coping mechanism for me personally), it is to remember that the karmic gears of time tick slowly but inexorably. To be a paper-failure 10 years from now that didn't make anything anyone wanted and to still be "rich as fuck" is a certain kind of Dantean hell in and of itself. Your entire life is the real life version of the Chinese "heaven's ban" where you are surrounded by sycophants who always lie to you to get access to your resources, and anyone else worth knowing would never want to associate with you because they deem you a clown and a fraud. You never learned the skills during your formative years to stand on your own two feet, and now you are too old to do the work that brings at least its own intellectually stimulating and market validated reward, as well as the respect of people in the world who are most worth associating with -- never mind the respect of salt of the earth people!
Perhaps this is my NYC elitism creeping in, but I wouldn't switch lives with such a paper failure SV talking head if you paid me. They'll never have taste, or be able to truly enjoy the finer things in life, or stand on their own two feet, or live in a truly cosmopolitan manner; they'll never have the edge to be as good of an operator as me or the operators I respect. And they will always know that deep down inside they are inferior no matter how hard they posture -- in a sense, I am more free than they will ever be, and if the pinnacle of life is to "live free or die" then they are effectively the walking dead, while I am breathing my own air free and clear.
Where was I going with this rant? Ah yes. Perhaps you need a break from SV because it's only there where these sorts of folks receive any kind of societal acceptance. I'd recommend either the east coast (I love NYC and it's never been more fun), or southern Europe (Spain and especially Portugal have always been nice to me). If you stay long enough in the SV echo chamber, it will warp your thinking and make life seem a lot less sunny than it could be.
In my opinion (and this is having no idea about your personal situation), I view it as just a bubble where you happen to work to make your living, and you can always exit as you please when you need a break or you've had enough. Hope I haven't been too presumptuous with this post and that I've at least been a little helpful.
For context, I spent ~10 years in SV, then another ~3 in NYC (but for an SV company most of it), and now I'm back in San Diego where I'm from. I think you're right on the money that I'm a more than a little over-rotated on the SV worldview. I'm generally a bit more balanced and...diplomatic about it, but when a rough week brings out some weapons-grade snarkiness? I'm clearly at least somewhat wrapped around the axle about it.
It's not so much that other people worked less hard for way more money that grinds my gears: I've lived comfortably on each of 3-ish orders of magnitude in terms of income, net worth, etc, and if I pulled out all the stops I could probably grind a bunch more cash out of my career than the perfectly reasonable amount that obtains. I work way less hard for way more money than a lot of folks around the world and feel a bit guilty about it in fact. Maybe humbled is a better word than guilty.
I think that I spent way too much time around people who attributed success to intelligence, insight and hard work, when it was actually due to intelligence, insight, hard work, and a lot of luck. I know some absurdly rich people who acknowledge that there were at least a few dice rolls along the way, and those folks don't get under my skin. Coincidentally or not, those folks seem to not have a bunch of opinions about how other people should live and work and a megaphone.
Andreessen is not far off with the "software eating the world" stuff, in some sense caring about the culture of the tech scene is a bigger and bigger part of caring about society in general every year. And I think that ultimately what bums me out is that YC/HN was one of my key/formative inspirations when I was getting serious about math and technology. Things like the T-shirt that says "I made something people want" that you got by selling a startup to people who enjoyed the experience really spoke to me.
For all I know sama and seibel are really smart, really solid people, and I regret calling them out on hearsay. I know at least one person who I trust and respect who knows sama personally and thinks the world of him.
But optics matter, and I'm probably not the only person who stopped believing in Santa Clause right around the time that pg handed the reins to the guy who needed a bailout on Loopt, and got kinda morose around the time that the Loopt-bailout guy handed the reins to the Socialcam-bailout guy.
I think ultimately I'm obviously very touchy/sensitive/negative on the topic because YC/HN was a childhood hero that ended up being the same damned nepotism that it was supposed to replace. At least MIT is available to anyone with a YouTube account and some time. YC is the new MIT and much, much more exclusive on the basis of who you know.
That's enough of the pissy part: I've got it insanely good for a guy whose grandparents worked in coal mines and gratitude around that is probably the right thing to focus on. It was a rough week.
Thank you again for such a compassionate and thoughtful reply, you've given me plenty to think about. Cheers.
I.e. when investing in the stock market it's almost common knowledge that silently putting money in an ETF every month will outperform almost everyone, even professional traders.
Overreactions to day-to-day things like corrections or a 1-in-a-million hiring situation break the obvious macro strategy
When an asset class is going up and up and up, FOMO can get even serious professionals to go long at a (relatively) high price, and gloss over the risk management. When things correct a bit, a lot of folks realize they hadn't managed the risk and get short "before it gets any worse".
This is probably the most common way to buy high and sell low, and while retail investors probably do more of this than hedge fund managers, hedge fun managers also do it.
Dollar cost averaging into a diverse set of ETFs is the simplest and cheapest thing that gets you highly competitive returns (at least to date). But I think that this is more to do with how much it takes emotion out of the picture than that SPY is like, ideal. Buffet, and Michael Burry, and others have demonstrated that if you're willing to spend years to decades of 16-hour days reading public filings in a drab office, it's not "hard" to beat the S&P. But those people are a lot more dispassionate about their trades than I am, and I suspect than most people are.
So in practice, policies are in place such that firing people is actually somewhat difficult or lengthy, even though yeah, legally Google could fire almost anyone instantly. So if they hire someone who sucks, getting rid of them would be somewhat challenging just because of those self-imposed policies.
This sort of implementation inevitably becomes a mountain of technical debt that overfits to the current process and reduces operational agility when the current process changes. At some point with business applications you have to trust your people to do the work and view the technology as something that enables that rather than necessarily constrains every dimension of it.
Their financial incentives aren't aligned with customers.
I blundered my home address when initially setting up my cable internet installation. The installer called from the other house and I redirected him. He told me just to call customer service to have it corrected.
Multiple attempts, there is no way to change the billing/service address of an open account. Web says to call, customer service agent says to use web..
I've set to paperless billing and figure for any service visit, the tech will call me confused from 2 doors down and I will redirect them. Easy enough. Two year going now, so far so good.
I can see someone overlooking the need to support address corrections because that is probably a rare case, but customers moving within the company's service area and wanting to have service at the new address surely is common enough that they cannot have overlooked supporting that.
In my experience working working in tech support at a company that encouraged us to fix whatever we could do like as, service address and billing address changes occured everyday,multiple times a day.
At the time I was there that ISP had fewer than 200,000 customer, wasn't even one of the big players at the time.
Intenode in Australia, based in Adelaide, circa 2004 through 2010.
When I first started working at Intenode, tech support was still solely on the third floor of 132 Grenfell Street, and you would occasionally bump in to Simon Hackett in the lift.
I was still there when iiNet acquired the company, but left before TPG acquired iiNet.
Good times.
It's cheaper. If they don't let their customer service agents have any autonomy, then they don't have to hire for things like 'judgement' and can just throw warm bodies in a chair, which makes people easy to replace and cheap to employ. It also prevents your cheapness from causing issues; if you hire decent people and rely on them, then it causes issues if those people leave en masse or start siding with your customers over you. "Sure, Mrs. Johnson, I'll forgive last month's bill; they don't pay enough managers to check and I'm quitting tomorrow what do I care?"
I’ve seen a similar thing with integration tests, as on a team with decent unit test practices production bugs are often a result of bad assumptions about another team’s API. Eventually a mountain of brittle end-to-end tests cripples a team until they get managerial buy-in to rip them all out and the cycle starts over.
Yes, checklists and process standardisation will avoid numerous types of errors and issues. However one of your checklist processes must be to assess the checklist itself, weeding and refactoring the checklist itself from time to time. This includes recognising that the refactored checklist will itself have further issues and shakedown bugs.
Ultimately process is a compromise between what can be kept in working or near-at-hand memory, and what's both sufficient and required for the particular problem domain.
The purpose of keeping a record of past postmortems is that you can potentially go back and say “oh gosh, it isn’t actually rare, and so maybe the more elaborate countermeasures are justified”.
I’ve seen the opposite extreme happen, where a devops guy was adamantly against writing a postmortem at all, because “it wasn’t our fault” and “it probably won’t happen again”. Facepalm.
I’m a firm believer in learning from mistakes, and making the right trade offs between reliability and velocity. Postmortems are a fantastic tool for facilitating this. As you point out though, the implementation matters and you need to be conscious of the cost/benefit of the recommendations that come out of the exercise.
If that list of extra checks is automated, i.e. added to your deployment scripts rather than manually followed by a human, then that's great. Every bit of repetitive decision-making and tribal knowledge that you can eliminate from your team's day-to-day functioning is a win.
(Spoiler: NY was "Search by State", CA/TX was "Zip + radius")
Population density means zip codes are extremely concentrated in NYC, especially in Manhattan. Traveling down the length of Manhattan might have you cover ~25 zip codes in 13 miles. Basically, a zip code isn't a meaningful measure of distance like it might be elsewhere. A radius might also place you into NJ, which could be an issue depending on what you're trying to accomplish: Delivering something from Queens might be significantly easier than from NJ, even though Queens might be further.
That's my guess anyway, though it's odd because nowadays generally looking something up via zip code + radius is actually easier in Manhattan (with a map you can easily see whether it would require coming from Jersey, crossing the park, etc. etc.)
Contrast to FL, TX, CA, and we had to immediately support zip-code-ish search (a non-trivial technical and operational burden, as zip-codes are rough-approximates for GEO's, and change pretty frequently) b/c "Couch for Sale in TX" means something completely different from "Couch for Sale in MA".
Contrast to the midwest (eg: kansas, montana), and this random-population-density-map I found: http://ecpmlangues.u-strasbg.fr/civilization/geography/US-ce...
...you can see that "just tell me what state it's in" hits within ~1hr of a single city center up until you hit TX or FL. So in 2000 it was a reasonable shorthand for "east-coasters" to say: "MA == Boston of course", and "CO == Denver" b/c where else was the internet going to be? ...either relatively close-by, or at your states major population center.
This is totally non-scientific and anecdotal, but I worked for a P2P sales startup out of college, implemented the zip-code search, radius, and update code, and distinctly remember noticing that some competitors (and dating sites!) definitely didn't support zip + radius searching, and they were usually ones that were funded from NY. CA was probably a bit more technically advanced, geo-aware, and wouldn't usually launch w/o zip + radius capabilities... again b/c "Couch in NV/CA", or "Singles in NV/CA" is a completely different scale from "Couch in NY/NJ, or Singles in NY/NJ"
(edit: in the 2000-era!! barely dial-up, and computers in general were decidedly non-rural devices, no cellular networking, and a motorola "two-way" was peak connectivity).
That said, I’d take Suffern NY over south Jersey, so by state isn’t ideal either.
But if you were in that 1%, shit sucked, sorry.
Two examples:
1) You're a single parent working part-time as a legal secretary, and in receipt of some form of social welfare payment to ensure you and your children don't end up in hardship.
You are offered full-time hours! Yes! Except because you hit the boundaries/limits/edge cases in the legislation, your net income will only increase by $30 a week, because your income tested supplementary assistances will decrease as your earnings do.
Now, the legislation in play here, ensures social welfare and works okay for 99% of people (when I say "works okay", I mean, gives them money, and doesn't disincentivise work). But that income testing bites you as you approach the edge of the income curve.
2) You want to go hunting in a national park managed by the Dept of Conservation. The usual hunting permit requires that you only use centrefire rifles - rimfires and shotguns are totally banned (to prevent hunting of native birds, especially the tasty fruit fed kererū/kūkupa/wood pigeon[0]). However, you want to hunt red deer and chamois, with a crossbow.
The policies at the time don't account for crossbows (and this was before bow hunting was as common place as it is now). So the department defaults to "No".
In both situations, the broad rules work well for 99% of users. But for the 1% it sucks. The key to the 1% is human discretion and agency.
Good systems, whether legal, business process, or algorithmic, always allow for a human to override. Bad ones don't. "Computer says no" style.
(On the crossbows, a ranger (my Mum, I'll be honest) went into bat for the hunter in question, to determine what was needed for "no" to become "yes", and well, now there's a policy that determines the minimum draw weight for a crossbow to be considered humane - and no barbed or exploding bolts allowed! [1])
[0]: https://www.doc.govt.nz/nature/native-animals/birds/birds-a-...
[1]: https://www.doc.govt.nz/parks-and-recreation/things-to-do/hu...
But I suspect a) was the main driver :D
The UK introducted legislation stating that people on a specific type of visa would only be eligible for permanent residence if it was issued on or before April 2010. The same legislation said, if this type of visa was issued after April 2011, it oculd only be extended to a maximum of 6 years, and then the visa holder would have to go back to their home country.
Well, there were obviously a bunch of these visas that were issued in the interim period, where the holders wouldn't be eligible for permanent residence, but if the employer company was willing to do so, could keep extending these visas indefinitely.
This seemed so baffling to us that we spoke to multiple immigration solicitors and the Home Office/UK Visas & Immigration support reps (or whatever outsourced org they were then -- Capita? Sopra Steria?) and they confirmed that this was true -- the visa could be indefinitely extended unless the government changed the rules.
The government eventually patched this bug, but IIRC, not before most such visa holders used the extra time to sort their employment situation out. (Opinion, not facts, based only on forum discussion anecdata)
My mother at one point made $25/mo too much for her childcare assistance because she worked some overtime. She lost her assistance for months and ended up several thousand bucks in the hole. Plus joy for me, since when parents can't afford childcare, guess who gets to do it instead? If you guessed usually the oldest kid, here's a prize!
Q: “Why did nobody notice the baby is sick?” A: Babies are weighed in every interaction, which is humiliating to parents.
Q: “Why didn’t you stop Sally from enrolling for benefits in 2 counties?” A: Fingerprint indigent beneficiaries, at their expense.
Q: “Why should people on drugs get welfare?” A: A family loses crucial benefits because mom smoked weed.
Bureaucracy builds walls out of process to remove individual agency. But politicians design programs to meet different goals, and social services law was no different.
Why is this humiliating?
Netherland recently had the long-running scandal of the "toeslagenaffaire", where the tax service had a tendency to assume people from foreign ethnicity or double nationality might be committing fraud, and gave them additional checks and hoops to jump through, leading to a lot of people to be treated as criminals and have to pay back money that they did have every right to. Was incredibly poorly handled, government fell over the scandal, but the new government consists of exactly the same parties, with the same PM, and I have no faith at all that this won't happen again.
I believe the "Ask for ID" is a strawman that wouldn't be helpful, but there are likely non-invasive steps we can take if we spend more time thinking about it. If nothing else, it may result in a less hostile experience for all involved.
Eventually after some incidents they introduced “special instructions” for customers who we had to gather extra troubleshooting data for because of some complex issues they had.
Then a few more customers had special instructions… and a few more.
Eventually these involved things like “Send the account manager an email for each call, this customer is very sensitive.”
After just a year EVERY account had special instructions.
Many people oriented instructions were impossible to follow “Page Jim.” Jim didn’t carry a pager anymore and when he did he would never respond so you couldn’t do step 2.
Some involved gathering loads of memory dumps for issues long solved. Others would have you follow program pointers to memory that didn’t exist so you would start over and over until you gave up and decided failing to do your job was preferable to losing your mind at 4am….
Some instructions were multi page word documents expected to be read every time the customer called. After reading it a dozen times a week it almost became impossible to notice if it had a slight change or not.
One clever account manager trying to make sure tech support “noticed” his special instructions tried using a <blink> tag… entirely ignoring the issue that upon noticing the instructions they would be hard to read. Thankfully the tag failed to work.
As you can imagine eventually there were exasperated executive meetings about how nobody follows the special instructions.
This is the world of protecting against corner cases.
Computers don’t have emotions; I don’t need to worry insulting the vast majority of S3 objects when I defensively check integrity every time. But humans are different; when we design a human system around uncommon cases, we do need to consider the ramifications on the majority.
We can hide the complexity in computer systems even engineer the risks of it out, this is not the case in human centric systems. That complexity will endanger everyone involved.
KISS is a good principle generally, but it's the most important principle in humans systems, sometimes even ahead of completeness.
I've chosen a contentious example, but there are dozens of others. Just for another example, if I was born in 1977 (above age of consent), what does the forum software care my exact date of birth and err when I don't want to provide it?
Same with the Last Name field. I happen to know someone who doesn't have one. Why isn't he accommodated?
How about colour blindness? Why isn't there a colour-blind accessibility option in the sign up form? Is making colour-blind people's options the same level in the UI as regularly-sighted people a bad thing?
I think that my point is made.
Having a separate first-last name is often brought up as a UX failure for exactly this reason, and UX designers often design websites so that colourblindness doesn't hamper usability (in my experience working with designers at big companies). The Christmas one isn't as much of an issue because AFAIK very few people are meaningfully put out by seeing Christmas decorations on websites from majority-Christian countries. In other countries these things often /are/ turned off depending on cultural sensitivities.
We should be trying to design our processes to fit the world around us, not rejecting the parts of the world we don't like.
The obvious solution is to just not use pronouns - it's a messy part of the language that is currently in flux, so why wade in?
Obviously, some sites are dedicated to these issues, so they can justifiably ask on sign up, but if you don't _need_ to care about people's personal details, better not to ask at all.
Forms of address --- Mr., Mrs., Miss., Ms., and often professional titles (Dr., in German Ing. (engineer), esquire (lawyers), Reverand, etc.) is at least fairly common if not entirely standard practice.
I suspect, again, some motivation on the part of a requestor. A magazine's circulation department, for example, might want to know the number of lawyers and doctors among its subscribers as a proxy for advertising value.
Many business information request forms provide detailed rosters of who you are, what you do, and your company title. For similar reasons, I suspect.
(I also feed those bogus information as a matter of course.)
The New York Times formally adopted use of "Ms." as a title, distinct from "Mrs." (a married woman, often referred to by her husband's full name, e.g., "Mrs. John Q. Smith"), or "Miss" (an unmarried woman or girl, addressed by her given and maiden names, "Miss Jane Q. Jones"). Ms. Magazine was a direct and deliberate challenge to that practice, and was launched in 1971 by Gloria Steinem. Interestingly, all three words, "miss, "missus", and "mizz" originate from "mistress", which was at one time the single title applied to any woman, adult or child, married or not.
(For men, the terms "Master" (unmarried child) and "Mister" were both represented as "Mr.".)
And then as now, "Ms." resulted in much gnashing of teeth, changing of forms, and updating of databases.
https://en.m.wikipedia.org/wiki/Ms.
https://www.nytimes.com/2009/10/25/magazine/25FOB-onlanguage...
There's virtually always the option to leave the option blank, or to make up a garbage or meaningless value. The first system on which I recall the option being offered was Google+. My response was "trans-krell", playing of my pseudonym's character.
For age, I usually try to find the earliest possible birth year acceptable to the system. For Google this seems to be about 140 years prior to the present date. Again, I avoid providing this information if possilbe (most of my various little-used Google accounts have either no value provided or a ridiculously early age).
I'm of the view that we don't need to be formulated, sprawling on a pin, within some global surveillance database(s). If I can evade classification and feed garbage to the system, I will, for as long as that is viable, and probably for some time after that point.
The gendering or nongendering agenda concerns me far less than the Total Information Awareness agenda. Services demanding anything from me other than some random username and password (I tend to use password generators to create both values), and possibly a contact email address ... tend not to get used.
As I just commented to a friend a few days ago, I can't remember the last time I did create an account, with the exception of some recent Mastodon and Diaspora* migrations in the past year or two.
For my most recent Android device (the Android aspect of it being among the least attractive characteristics), I bailed out of Google Play Store registration, which requires creating a Google account. (Even if not formally associated with other identities I have, those could all but certainly be trivially linked.) Instead I'm relying on F-Droid, APK-Mirror, and the Aurora Store. I've kept actual app installations to a bare minimum, and most of those through F-Droid. There are I think three apps with actual accounts associated to them, though only one has been so configured.
My use of the Internet dates to the 1980s. I've seen a lot. And am disliking increasing amounts of it. Read Dan Geer if you haven't recently.
As a human, can you please tell all of us what these abbreviations are?!
...TFA: The Featured Article
a.k.a. (also known as) OP (original post(er))
This is one “human corner case” I have no shame working around. Wouldn’t want to convey hostility where none exists.
I prefer "TFA" to "OP" as the latter may be ambiguous as to whether it refers to the article or, more typically in my experience, commentary on it. Even here, "OP" might reference either a thread root or the parent of the post immediately being replied to.
In the second case, if I were to say "OP" here, I'd be referring to your comments parent, by tchalla, https://news.ycombinator.com/item?id=30345056
It refers specifically to the submitted article, and carries an overtone, sometimes tongue-in-cheek, sometimes subtle, sometimes more blunt, that the person being referred to ought to read the specific submitted work more closely. Perhaps at all. And without crossing HN's guidelines against specific accusations.
That is, the meanings are similar but different. "TFA" is more concise and specific. All of which make it the more fabulous ;-)
It turns out I do use both terms fairly frequently, as my comment history shows. The distinction is largely as I've described above. "The article" usually refers to some additional or other reference rather than the HN submission. "TFA" in the context if "you/I should have read that closely".
TFA: https://hn.algolia.com/?dateRange=all&page=1&prefix=true&que...
"the article" https://hn.algolia.com/?dateRange=all&page=1&prefix=true&que...
The real problem with crypto is that usually what is asked for is a ledger, and the crypto and distributed parts are just bloat.
Unlike code that triggers (hopefully) only when something weird happens, the cost of the special instructions and the dysfunction that followed was ultra high.
This sounds insane and I genuinely don't understand why anyone would work here or put up with this.
I understand the main point, I think, of 'make sure that people read instructions instead of ignoring them', but this is taken to bizarro levels which I don't understand why anyone would tolerate. The 'account managers' designing these 'special instructions' are either sadists or just assholes, and if management is not willing to protect its own staff from this bullshit, I see no reason for anyone to continue to work there. I don't really understand people staying in abusive job situations like this. There are many other job opportunities out there.
I'm working to get out of a job that turned into this once our (lovely) manager bailed and exposed our team to the incompetence of our current leadership.
I'm having to HEAVILY fight my gut instinct by leaving, because I was raised in poverty and grew up as a disabled lesbian back when those things created more issues than they do now. I'm also from a rural area and neither of my parents have college degrees.
I may logically know that I (as someone with credentials and a decent skill set) have every right to leave and it's in my best interest to do so, but it conflicts with literally decades of messaging I've got telling me I'm lucky to be allowed where I am at all and seeing those around me be punished for standing up for themselves.
I was also abused by my parents, and that's another angle: People who are used to abuse don't think it's that bad, and I'd wager there's a fair number of people who have just never had a decent work situation. If you're going to be abused, at least pick an abuser you know and can work around.
Humans are pattern recognition/social machines, and if you only expose humans to dysfunctional patterns, they'll assume the problem is them. If you design a society, it's pretty easy to get yourself a group of people who are open to being exploited.
I've had a lot problems myself struggling around guilt about the pay and general lifestyle in tech. But I'm now working in a tech company with lovely people, doing interesting work. For what it's worth from a stranger: I wish you luck in overcoming your inner critic and that you are worthy and deserving to get a job with people you like at a job you enjoy.
This is a large part of it, especially as someone whose community is likely to have those sorts of jobs. If everybody around you hates their job and you grow up in a culture where people actively despise their workplaces/management, you assume that's just how jobs are.
EVERYBODY I grew up around hated their jobs, so it's really hard for me to know what to put up with. To people like my stepdad, my displeasure at how I'm treated by leadership is complete stuck up whining: I personally make more than him + my mother and their households ever have, I'm not in physical danger (unlike him doing HVAC work, my mom doing warehouse work, my siblings' SOs who work landscaping, etc), etc. I was taught growing up that basically unless your boss put you in the hospital, you were lucky to not be homeless or starving and work just sucks.
This is a particularly hard mindset to get out of because in white-collar environments, you can't talk about how growing up blue collar or working class makes you anxious because a lot of 'professionalism' boils down to 'don't let the nice people know you're a peasant or you'll be kicked out for not being a culture fit'.
The pandemic has been very helpful, ironically! I'm more than willing to blame myself for all my life problems, but when I see OTHER people being treated like I was, it raises my heckles and pisses me off.
I'm so glad to hear that you found a job you like. I'm hoping to move from a small-shop/journeyman dev to working in a dev team and I'm trying to maintain optimism, so stories like that help!
Feeling safe let me leave bad employers in the past. Thankfully my current one is pretty awesome.
My personal experiences, outliers though they are, tell me that change is likely to screw me over. I would love to save my paychecks but when your meds cost 300k+ a year for your entire life, your life is dictated by health insurance and care. I lived like a college student and saved and was STILL fucked by change, so I'm very change and risk averse.
While the special instructions incidents were horrific, particularly to someone like me who wants to do the right thing and their brain seizes up with such insanity, they were eventually resolved.
Finally they instituted a policy where instructions were reviewed by a group of senior account managers every 90 days. Whomever wrote them had to justify them, or they were simply deleted. That actually turned out to be a great process as many out of date instructions were auto deleted without anyone doing anything.
As for the job, it was actually a great job. The support team was amazing, supportive, it paid great, and the company weathered a lot of economic downturns easily. If anything the special instructions incident demonstrated how even a great organization can go bonkers insane.
Try Auftragstaktik instead:
https://en.m.wikipedia.org/wiki/Mission-type_tactics
Weird edge case escalation happens, engineering team spends a couple days handling it.
In sprint planning/retro – "we have to make sure we never waste time over this again".
Team spends weeks coming up with an optimal solution to address this very specific case.
Original problem was rare enough that it doesn't happen again.
A different edge case happens. Rinse and repeat.
They don’t. But some days they are complete arseholes and you do begin to wonder.
> We could ask candidates on video (or in person) to see a photo ID and match the ID against the resume. But this would seem very weird. It starts an interview off in a hostile manner, and send the a strong message of distrust. Honest candidates – which are, remember, the vast majority – will wonder why the heck this company is acting so weird, and will rightly see this as a red flag about the company culture. There will be negative consequences for your hiring practices.
This sort of verification is already being used for some KYC processes by financial institutions (i.e., take a live video of yourself holding your ID that’s uploaded for verification). It’s probably a matter of time until this is so normalized with virtual KYC that people wouldn’t care much when asked to do the same by a potential employer. The process already excludes certain people (like the transphobic process mentioned in the article).
There is always a cost to the company for an incorrect hire, regardless of the underlying reason. So I would expect company HR teams to read that (widely shared) post and formulate strict identity checking and recording mechanisms so that they can absolve themselves from what may seem like an oversight or error. It also fits well with the typical HR style that’s more about controlling employees than about enabling/helping them (apologies to any good HR folks who are out there who struggle against the weight of the systems).
Though as it becomes more common, it's probably also a matter of time before software is available that lets someone hold up a green card and it's replaced with the document of their choice on the video feed.
Seems like it'd be better if chip card readers became ubiquitous on home computers and mobile devices, then we can scan our drivers license or other government ID to prove that we have possession of the physical card (and when making online purchases we can scan a credit card instead of typing in a number that can be stolen).
What if this was based on telephone interview?
What if remote, and one would be none the wiser?
As long as results happen, things move forward, are we not all happy?
What if the remote role was ostensibly being done by one person, but in fact they delegated to a team of people? A team of people that themselves come and go but all get managed into the same bodyshop? Does that matter as long as the work's getting done? What if, to mitigate potential loss of IP, a third-party information security policy needed to be agreed? Is it then turtles all the way down?
Honesty and integrity do matter. Trust matters.
"For example, anyone who goes by a name that doesn’t match their government ID could be forced into an uncomfortable explanation"
I've seen this happen more than a few times, and the explanations are seldom anything you'd have sympathy for when discovered during a background check. People with good reasons will tell you up front.
Why not just take a picture? If you're interviewing so many people that there's a genuine risk of forgetting who is who, a picture seems useful. I'm sure I'm not the only person who will forget your name in 3.5 seconds but never forgets a face.
Maybe this is my time at Pivotal speaking, but your first day on any team you get your picture taken with a polaroid. It goes on the pairing board. It's not weird.
"Thank you for interviewing with us! Mind if we take a picture so we can keep our who's who straight when the team is making the decision?"
All that jazz with IDs sounds unfriendly, sure. A snap isn't. Hell, make it a selfie with a team member.
Less so in smaller start-ups, but part of the process.
If you don't have ID cards, they could be part of, say, a company directory, wiki, or other system.
This isn't inherently adversarial and can be integrated into proceedures reasonably smoothly.
Again, this fits into workflows and isn't directly confrontational, but does preserve a record for comparison.
I really, really wish this writer had been the guy in charge of the TSA back when the “shoe bomber” incident had gone down.
Part of why the answer for hiring fakes is 'do nothing' is because the downside there just isn't that bad, it's several orders of magnitude less destructive than "plane with dozens of people on board explodes".
Is it because someone tried in an airplane once? Tried and failed? Would you change your mind about Starbucks if someone tried and succeeded?
The whole point of that article is that you can’t design around extreme edge cases. If we used shoe bomb security theater logic around all aspects of flying, there would be no airline industry.
Likewise the "no liquids" thing because of a risk of "binary explosives".
Now weight the potential impact of the downside by the cost to society/taxpayers.
There were a number of incidents in which attackers gained access to office towers with deadly effect. One that comes to mind is the 101 California shooting in San Francisco in 1993.
https://www.sfgate.com/news/article/10-YEARS-AFTER-101-Calif...
At any rate, the article is right, this probably causes problems for some candidates (especially, but not exclusively, trans people). But if it's common practice in NYC it's likely most local candidates are used to navigating it. That's less likely elsewhere.
Processes are modeled as having "normal" and "special" causes of variance. In managing processes, there are risks involved in overmanagement of either, though the Wikipedia article linked below doesn't seem to mention these.
For special causes, building too many special-case checks can create an ossified process. For normal causes, a problem may arise that in attempting to manage what is essentially normal random variation, additional variance is added to the system, or management processes themselves inject further variation or failure modes.
For an example I only just ran across, the Fermi 1 nuclear reactor meltdown incident near Detroit in 1966, and event which gave rise to a book We Almost Lost Detroit, and a Gil Scott-Heron song of the same title. The cause of the meltdown was determined to be "zirconium metal plate that was installed in the reactor as a safety measure", according to a Detroit Free Press article on the incident. Not the first or last time safety equipment has contributed to an accident --- think of the 737 Max and its MCAS system failures, or the thermal insulation cladding of the Grenfall Tower which precipitated a disasterous fire.
I may be mis-recalling or mis-understanding Demming's points on process control, and if anyone could help nudge this the right way I'd appreciate it. I'm pretty certain there is a connection however.
https://en.wikipedia.org/wiki/Statistical_process_control
https://en.wikipedia.org/wiki/Enrico_Fermi_Nuclear_Generatin...
https://www.freep.com/story/news/local/michigan/2016/10/09/d...