What does that actually achieve though?

Ok, they know that you called them an idiot.
Does that make them less of an idiot?

Does that actually help the world in any way?

Correcting them, without insulting them, may....possibly, result in them learning something. Possibly, maybe.

Insulting definitely won't, it'll just make them write off all of your opinions.
Insulting people because you believe they're "idiots" and need to be called out as "idiots" only satisfies your ego, it doesn't provide any other benefit.

It's also about not operating with all the guardrails and governance and red tape and arse-covering that a giant corporate entity can't survive without.

The reason Startups succeed (and the reason they fail), is because they do a whole bunch of very sketchy stuff that they would never get away with in a big corporate.

You really can't get startup success without startup risk, and I just don't think a company the size of Amazon is actually willing to accept startup risk, no matter what a C-level says.

Unless they fire their entire legal, cyber security and HR departments, plus 90% of all the managers and just tell the senior engineers to go for it.

I'm missing the part where this law prevents under 16 year olds from hosting their own websites to express their opinions...

Maybe he's thinking they'll lose a leg or something?

If you're having difficulty understanding, you could possibly try the rather unconventional approach of reading the second sentence too.

Maybe even all of the sentences....

In Australia the difference is that your four weeks annual leave is part of your entitlement that you've worked towards.
The company always owes you that money.

So if you work ten years and never take annual (non sick) leave, then when you leave the company owes you 40 weeks pay.

Sick leave isn't paid out like that. It does usually accrue (probably to prevent people suddenly all getting "sick" at the end of the financial year each year), but it doesn't get paid out when you leave.

The initial unit tests can always be written first, because true unit tests test output given input.

You know what the output should be for a given input, if you have to write the function first before you discover the possible outputs then your function is poorly defined from the beginning.

Once you've got your tests for your success case(s) and your anticipated failures, then you can write your function, then _after_ that you might use coverage to see if there's a path in your function as it is written that's not covered.
And then you're able to ask yourself if that path should even exist - if not, then fix it, if so, then write a test for it.

But you'll definitely get better results if you write the tests as specification first, and then write the function, than if you just test what you wrote after the fact.

Drinking a gallon of seawater rarely ends well for anyone, radiation or not.

Bank transfers - easy, fast and free.
And even before they were fast they were still the standard way to do it.
I never paid rent with a personal cheque - before bank transfers became the common way to do it, your landlord would give you a payment booklet and you'd take that to a branch of their bank and deposit the money.

When I bought my last car a year or so ago, I paid $20k via a bank transfer - although I spent a good 15 minutes double checking the numbers before hitting "Send".

It does feel like there should probably be some other solution for the really big numbers - like house purchases (for me, that was done with a Bank Cheque, 7 years ago), but maybe we'll all just be comfortable enough with bank transfers by then for that to be the solution there too. It really only needs a guarantee from someone that you can get your money back if you accidentally send to the wrong account.

They're quoting a Monty Python song...

Stuck in 1999, I see...

....that's not living "paycheck to paycheck". It's the exact opposite.

"living paycheck to paycheck" means that if you didn't get a paycheck, it's game-over.
If you've managed savings of more than a couple of months worth of wages/salary, and you actually _have_ any sort of positive "net worth" to build, then you're not "living paycheck to paycheck".

ok, so you've got testing in place. And an updated version breaks the tests.

Then what?
Sit on an old version indefinitely?
No, you're going to have to go to a bunch of effort to replace the library.

Just because it doesn't result in production problems doesn't mean it's not a giant pain in the arse, and a real cost to anyone (not just businesses) that depend on it.

Isn't that the current situation though?