Captain Obvious: real world transactions don’t look anything like this.

Sometimes we needs to start with a simpler model, demonstrate concepts, and work our way up. Of course, it is important to acknowledge the model is simple.

How would you better model real world transaction?

To be fair, the answer to that question is probably someone's PhD thesis.

Quite possible, but he started with "Captain Obvious" so I hoped that meant he knows other models that are obviously better.

Even if they do know other economic simulations which are better, it isn't a given that a better tool exists for every tool that doesn't ideally meet a need. Of course, you did say "hoped" instead of "expected"...

He "sells" these ideas as if he was the first one to ever think of them.

I always thought there should have been an enormous evolutionary pressure for humans to have different skin tones in different geographies. If sunlight is bad for skin than we should all be black or tanned all the time. However, if you don't live in a sunny area and you have a tan you might not getting enough Vitamin D. Vitamin D must be very important in order to make us evolve to have different skin tones and to tan according to the seasons.

The article mentions we store vitamin D for use in the winter.

It also suggests pale skin may have evolved to allow us to produce way more vitamin D in the summer to compensate for cloudy winters.

So, the advice would seem to be, go outside in the sun during the summer if you have pale skin.

Windows 10 + Windows Subsystem for Linux did it for me.

There are now services specific for this that work very well. Example: https://transferwise.com/u/joaoi

Your denominators are wrong.

The denominator should be the probability of the classifier being positive: P(True Positive)+P(False Positive)

Since you assumed that P(False Positive) = P(False Negative) = (1-P(True Positive))/2 = (1-P(True Negative))/2

In the algorithm case the denominator is 0.8x0.04+0.1x0.96=0.128

In the human case the denominator is 0.6x0.04+0.2x0.96=0.216

In summary, if a human thinks someone is homosexual, the probability increases to 11.1%. If the algorithm makes that call the probability increases to 25%.

The best way to memorise hundreds of different algorithms (or anything else for that matter) is to understand the few underlying principles beneath them. I would be surprised if he knew them all and couldn't come up with intelligent ways of mixing them for achieving a new task.

The best way sure, not the only or even most popular way though.

More like learning rate decay. Learning rate is now zero.

The human reaction time is about 0.1 seconds. They shouldn't use a manual chronometer to measure these results 4.9+/-0.1 is the difference between a world record and a good score.

I believe WCA officially uses StackMats for the timing. Timer starts when you lift your hands from the touch pads and stops when you return them to the pads.

This isn't perfect either but it doesn't rely on someone else's reaction time.

Correct.

The guy with the stopwatch isn't the official timer. He's just duplicating the job for fun / backup.

They are giving poor people another choice they can have access to a very limited Internet or they can have access to none. If they would just loose lot's of money by giving people this choice it wouldn't be a reliable business and soon it would end. What's so bad about giving people another choice? No one is forcing them.. and it's way better than doing nothing.

Yeah like I said in my other comment - deplorable but not unethical. FB has the capital to provide the service AND make it open.