Research skepticism

[BootinUp]

Note: Baldur is a long time software developer and author and writes about tech like AI

[e-mail]
Baldur Bjarnason

A reminder that you can dismiss out of hand any research that comes out of corporate research departments if the study in question is of that corporation's products, target market, or even vaguely adjacent to either

"But X's research department is full of good people doing great work!"

Doesn't matter. Their incentives are aligned towards improving the company's products or sales and aligned against academic or scientific rigour. Anybody who can resist that incentive won't last long in the job.

Baldur Bjarnason
@[e-mail]

There are exceptions to this principle. If the research is documentation of work or new practices, then that has value, for example.

May 27, 2023, 10:07·Edited May 27, 10:08 · · 0 · 0

https://toot.cafe/@baldur/110441735497868255

[Warpy]

1. You need to read this stuff with a jaundiced eye

For instance, how big was the study? What variables were and were not considered? What sort of population were they looking at? How long did the study last? Most of all, what is the confirmation bias?

Asking those questions will make you the popular press's worst nightmare and keep you from taking things like dietary hysteria too seriously.

[BootinUp]

2. +1