There is a difference between being shamed for breaking the build because you didn't run the test suite (and well knew that you should run the test suite) and being shamed for making an innocent mistake because you didn't know better.

And honestly, even breaking the build, you get a pass on the first one or two times, everyone has to learn somehow, and tests can be heisen-buggy.

Not so much on pushing broken builds to production though.

You should not be able to push broken builds to production. Shame to your manager if you are.

Is there really an over-utilization of blood tests? MRIs, yes, but I'm not sure that right now we're really doing too many blood tests.

Regarding your point on Bayes Theorem, that's valid in some ways and invalid in some ways. Suppose someone gets a vitamin D test back at 8 ng/mL. There's not really a chance of a false positive. He's deficient and needs supplementation. Making the test cheaper, faster, and more accessible can really only improve the health of the population. It will likely also drive down the cost of healthcare overall, because vitamin D deficiency has all sorts of unpleasant effects.

Furthermore, the literature suggests that the vast majority of doctors do not understand Bayes Theorem. Source: http://yudkowsky.net/rational/bayes

You are demonstrably wrong about the vitamin d. What if it was a specimen swap, and the lab screwed up and tested someone else's blood instead of yours? This happens.

The cheaper and faster the tests, the easier it is to reorder a confirmation (or two or three).

And then what.... Rock Paper Scissors for which one is correct?

Eh, other than Adwords Google hasn't been able to monetize their web properties.

I think part of it is desperation. If someone is diagnosed with an incurable or untreatable disease or condition, then it's natural to try to look for some sort of treatment that will yield better results.

That means applying less skepticism towards "alternative" therapies.

Then you also have the following working against science: - Confirmation bias - Mistaking coincidence with causation - Overly aggressive pattern matching

To be fair, if that bug happened to me, even with 10-20 data points I don't think I would have figured out it was a Tuesday bug.

Seems like quite a low rate. I hope they have a good team/plan for anti-fraud, because merchant fraud is a different beast from e-tail fraud...

Hm. I think this is the real answer.

If people like the app then it is useful. I don't think Apple should be trying to censor apps that people like.

Depending on what niche the website is in and how much traffic you get, you can potentially make decent money with the right affiliate program.

You can also upsell consulting services (again, depending on the niche).