Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

Ask HN: What has your work taught you that other people don't realize?

news.ycombinator.com

Post reply on HN
Print view View original

371–380 of 757 posts

Re: Ask HN: What has your work taught you that other people don't realize?

#371

Post by Psyladine »

muzani wrote: post #114

Everything is a system. The economy, society, relationships, nature, traffic. You don't need math to reverse engineer a system. You just need to pay attention to it. You can say the right words to make a date happy. You can figure out which lane is the fastest route, better than Google Maps can. You don't need an app or data - your brain is a wonderful data processing machine. Don't be angry at the people who are ben…

>If you don't fix a system, few will. Most people are reactive to it and try to live with it as background noise.

I'd add: If you want to know why a system remains broken, look to who profits by its brokenness.

Re: Ask HN: What has your work taught you that other people don't realize?

#373

Post by csours »

undebuggable wrote: post #104

Computers don't work, internet doesn't work, software doesn't work. If something functions flawlessly it's because there existed a group of people who figured out a rock solid setup at one point in time, and another group of people maintaining it actively right now. Many people seeing something that "just works" and it's "free as in beer" demand a similar setup for their own business assuming that creating it require…

A small example: A company develops a software product for a client. The company tests the product, and tests pass! No crashes! It just works.

It works for years with no problems. Then you get a flurry of tickets in a pretty short window: the software is crashing! What happened? It passed tests, it worked for years! What changed in the environment? Are disks failing or filling up? No. Is the network failing? No. Is another program hogging resources? No. Is it a database maintenance window? No.

Well the patch window changed, but smoke tests pass after patching, so it's not patches...

Hmm, the patch window went from 2 weeks to 4 weeks.

Ok, when did the software actually crash? About 25 days after patching. Is that consistent across sites? ... Yes it is.

What happens 25 days after patching??? [Insert scramble to figure out what happens 25 days after patching]

Is there anything special about 25 days? Well 2^31/86,400,000 is 24.85 ... Aka Signed Integer/Number of milliseconds in a day

So all along there's been 24.85 day fuse in the code, but before the extension of the patch window we just happened to defuse the bomb every time. The supplier never had any reason to run a test for more than a week, so they never saw the crash either.

----

So, yes, software doesn't work. It happens to work right now, on this computer, with this hard drive, with these network settings, with these dependencies, with this database .... etc etc etc.

Re: Ask HN: What has your work taught you that other people don't realize?

#374

Post by lbriner »

"Everybody is a little bit right"

I was taught it by a former colleague but it was probably coined by someone else. We often think that if two views do not agree, one must be right and the other wrong when in most cases, every idea is a trade-off and everyone is therefore a little bit right. We might be able to find a middle-ground solution or just accept that our idea isn't always chosen.

Re: Ask HN: What has your work taught you that other people don't realize?

#375

Post by dsfyu404ed »

mannykannot wrote: post #316
dsfyu404ed wrote: post #298

Earlier quoted context omitted.

>The degree to which everything runs on this sort of system is kind of horrifying, once you're exposed to enough things like this for that to sink in. So? It clearly works. People want everything to be high quality master craftsman type work but almost nothing is and it almost never matters. People like to think that the auditors are going over the papers with a magnifying glass and that their landscaper is applying…

It works right up until it doesn't: see 737 Max, mortgage crisis... Perhaps the best thing that can be said for pro-forma processes is that they usually put people on notice that they may be held accountable for their actions -- though, of course, that is not always the case.

And look at all the planes that don't fall out of the sky and all the financial products that don't cause crisis.

The "even rare failures are unacceptable" line of reasoning is simply too conservative to permit our modern society to exist. The resources we would have to dedicate to preventing these kinds of things (don't forget the 90/10 rule) in every case would be astronomical. Imagine if every bit of software had to be written the way NASA rights their software for human carrying vehicles and if every financial transaction had to be strictly scrutinized. Software would be so expensive that it would be used far less than it is today and all the good that brings would not exist. Credit would be far, far less accessible to just about everyone you could fill a library with books that have been written about why that is bad. And those are just two industries(!!!!).

It's really east to say "But the MAX" or "but the mortgage crisis" but occasional failures like that are simply the price of a reality where normal everyday people can travel thousands of miles in a day and have access to credit.

Re: Ask HN: What has your work taught you that other people don't realize?

#376

Post by m_fayer »

nkrisc wrote: post #327

How important it is to be able to inhabit another person's viewpoint and see things from their perspective. So much time and effort is saved in doing so. Much of my job has become translating and mediating between stakeholders going around in circles, because each one believes their viewpoint represents the entirety of the issue. They can not understand one another because they assume the other understands everything…

Absolutely. I would add: remember that we can all be right. That problem or dysfunction may have 3 root causes, not just 1, and that engineer, lawyer, and BA can stop debating about who diagnosed it right - they all did.

Re: Ask HN: What has your work taught you that other people don't realize?

#378

Post by lotsofpulp »

ufmace wrote: post #339
tpxl wrote: post #318

Earlier quoted context omitted.

> It clearly works. Clearly it doesn't and when it does, it's despite this, not because of it. The economic crisis of 2008/9 could have easily been prevented. Misinformation about WMDs led to useless wars. The war on drugs is a huge waste of resources because why? Ticking checkboxes is a big cause of our problems.

Kind of ironically, pulling out 3 anecdotal examples of well-known system failures doesn't make a very high-quality argument that most things being done sloppily most of the time doesn't usually work out fine. The real argument is, that while those 3 things were happening - and the War on Drugs goes back ~100 years - roughly 3 billion other things were also sloppily done, and most of them worked out well enough. If i…

tpxl and you are using different definitions of "work out fine".

Incarceration and systemic discrimination against an entire class of people does not count as "well enough", in my opinion.

And it's specious reasoning to conclude that if a society made up of untrustworthy actors committing fraud isn't starving and dodging lions, then it's worked out fined. One can go live in a country with low societal trust to see what that's like (Brazil, India, Pakistan, Somalia, etc).

I posit that it's the proportion of trustworthy actors in the system, along with a healthy dose of conveniently timed technological advances as well as luck providing resources at the right time that leads to a prosperous society. There are countless examples in history of a society doing well enough, and every time there is a tipping point where sufficient trust is lost and it starts degrading, or in some cases, collapses.

Re: Ask HN: What has your work taught you that other people don't realize?

#379

Post by csallen »

cambalache wrote: post #306
muzani wrote: post #165

Earlier quoted context omitted.

I've done all kinds of jobs. I'd take a 14 hour physical labor job over a 8 hour mental labor job anytime. Mental labor is painful... in the form that stalling the main thread of something is. But that's data entry. Development is much less mental labor, and I'd agree with the 4-5 hours/day assessment. Most people can't actually do this, which is why a lot of people hate math. Math class is dealing with intense focus…

I have done all kinds of jobs too, and I will take a programming/middle management job (no matter how soulless) over a 14 h, or even 8, physical job, every single time. Most of the blue-collar workers will tell you the same. Your skin suffers, your knees suffer, your lower-back suffers, honestly I doubt you have done it for a long time, it is people in that category who naively fantasize about how cool is to work 14…

Second this. I barely made it a week at a construction site when I was younger, and I was convinced the experienced shaved a few months off of my life expectancy. I was also just as exhausted coming home from work as I would be from an equivalent amount of time spent thinking about hard mental problems. At least for me, it wasn't even in the same universe as sitting down in an air-conditioned office and writing code all day.

Re: Ask HN: What has your work taught you that other people don't realize?

#380

Post by prometheus76 »

muzani wrote: post #292
anotheryou wrote: post #262

Earlier quoted context omitted.

how do you think about politics and capitalism than? can the system still be simplified enough to grasp it well enough to change it? where would an individual who has grasped it find the power to change anything? capitalism seems super stable for now one of the most sad thoughts I heard: if you try to change the system to the better you might keep it from hitting the wall and actually stabilize it

It's clear it doesn't work perfectly - depression rates in capitalist countries are very high, salaries are still a problem, even when GDP improves. The Communist revolution took off all over the world, but later everyone learned that communism isn't better. All change takes time and energy. I think we're still missing a great economic thinker who can try to solve it. We've put a lot of our energy into things like el…

I think it's going to be impossible to improve on price being determined atomically by buyers and sellers agreeing on an individual basis what something is worth. That's the heart of what makes capitalism so effective. What this individual is willing to pay for that individual item is almost impossible to predict down to more than a pretty broad range of pricing. So many variables factor into what one person will pay for an item. And that determination changes even for that person depending on timing. I might be much more likely to make an impulse buy on payday than I am on the day before payday. The decisions around how much something is worth is too unpredictable to be centralized or predicted.

"Let the market decide" is a misdirection from what makes free markets great: individuals deciding, for themselves, whether to make a purchase after a seller has suggested (or demanded) a price.

No central planning algorithm can take the place of that, which is why attempts at centralizing an economy are always undermined by black markets.

Post reply on HN