I used to think peer review served an important purpose.

I was wrong.

Peer review serves no meaningful purpose, and could be abolished without anyone (except academic's whose livelihoods depend on it) noticing. 🧵 Some arguments for peer review:

1. Without it, there would be no verification that published research is accurate. This is esp important in science and medicine.

2. Without it, it would be harder to make decisions about promotion, tenure and awards for the best researchers.

🧵on making dumb mistakes

I get really excited whenever I learn about a dumb mistake I've been making. The dumber the better.

Why?

The dumbest mistakes are the most harmful and also the easiest to fix.

By fixing them I'm guaranteed to be so much better off than I was before. Inspired by this, I've just posted an article about

How Loser's Lose: 8 surefire ways to lose at anything you do

harrycrane.substack.com/p/how-losers-l…

Some highlights below

Important life skill

You can get any answer you want, if you ask the question the right way.

Not just true for #ChatGPT Reminds me of an interesting lecture by @fbinegotiator about framing questions so that the answer you want has an answer of "No", not "Yes".

A thread illustrating the limits of machine learning and data science.

A powerful tool when used properly. Complete nonsense in every other case.

Cautionary tale against blackbox models: #ChatGPT edition Try before you buy.

#ChatGPT is literally making stuff up.

First question, "Who is Harry Crane, professor at Rutgers?"

("Who is Harry Crane?" gives an answer about the character in Mad Men.)

Answer: