3 Comments

A biochemist working for Novartis in Basel told me of an astounding finding recently. Of the top three prestigious journals. "Cell", "Nature", and "Science", 90% of their genetics articles have reported experimental results that are not repeatable. Among the leading causes: Cell lines get misplaced or mislabeled during the experimental work.

These sophomoric blunders cast serious doubt on the reliability of lab work that was pitched to the public with cultlike fervor and led to public denunciation of anyone questioning beliefs in the work.

When the Scientific Method vanishes white lab coats become cult vestments and so-called science degenerates into a cultlike worship of certain leaders.

Expand full comment

Incredible sloppiness - how can American universities have credibility when years of research go out the window? Boggles the mind...

Expand full comment

What's funnier is the guy responded with the same trick Angleton developed at the OSS during WW II The Big One: Bayesian algebra. The only way to sort the slop is by the number of correct predictions it makes. So he got the idea to start a computer company building storage area networks big enough to handle the amount of detail but with a small-enough form factor to fit in existing university data centers and a low-enough power consumption that they won't bankrupt the school that tries to run them. I started doing some tech writing for him. Meanwhile I improvised a way to do your spine flush routine with OTC substances: Alternate coffee with water. It seems to help a lot.

Expand full comment