The problem with ferreting out bullshit in scientific papers even in areas that I’m pretty familiar with is that even not-that-difficult publication take a long time to read, even longer to fully understand, and to do the calculations all over again is nearly impossible (for me).

How does the average person even stand a chance? They don’t. They really don’t.

This paper for instance is 45 pages long.

The paper’s premise is that 100% of the decline in the US labor share of income is due to intellectual property. This doesn’t even pass the sniff test, but I’m willing to give the paper a fair shot at convincing me. (So far it has not.)

However, even to make sense of one not-that-hard paper without any complex math, it has taken me about two hours. And I think I know where and why the authors went wrong. But it’d take me another 10-15 hours of work to prove it, which I’m not planning on doing because I don’t have that kind of time.

My point is that 90% of science could be utter BS (even though I don’t think it is) and no one would ever know. Or at least not many people.

Damn, we humans need an upgrade. This software just isn’t cutting it. It is buggy, slow, prone to failure at the worst possible times and can’t easily be improved.