I learned science by learning how people do it wrong. From Ben Goldacre to Lukeprog to Elizabeth van Nostrand to Andrew Gelman to Richard McElreath to Rex Douglass to Ben Recht. (This informal education was probably more important than the stats degree I did.)

Does this general education in cynicism and stats give me insight into fields I have never trained in? I don’t know! But now you can judge for yourself as Spencer and I both make our guesses about psychology and other things.

When I was a young man I came up with this argument — extremely hand wavy — that says something like: what you like constitutes a decent chunk of who you are. Your preferences are adaptive. If you do stuff, sometimes you grow to like it. You can choose what you do; therefore, in some sense, you can sort of choose part of who you are, just if your preferences are, in fact, adaptive based on your behavior. You can choose your behavior.

a lot of the time with this, it’s like an identity problem. You view yourself as the sort of person who doesn’t do that thing. I didn’t exercise throughout my teens and 20s, and now I’m really into weightlifting. And part of this was just shifting the idea of myself as an intellectual who rises above the body (the merely material or something).

even if, on average, people’s personalities don’t change very much, it doesn’t mean that you couldn’t change a lot more if you’re really strategic about it.

the academic… incentives lie in the direction of picking a couple of things, and then publishing on them all the time and keeping up on these incredibly small, focused areas, and then you work on it till you die. [laughs] And this has always unnerved me a little bit because I can’t really say what my two or three research interests are. But we’ve been speaking about psychology for 15 minutes; I have no particular psychology background, it just interests me. My field is nominally machine learning and I’m quite interested in machine learning, but I definitely don’t think about it every day. I definitely don’t think there’s anything that I think about every day. So the pointer here is something like generalists or polymaths or dilettantes — as I’d prefer to call myself — and the academic incentives pull against such people.

throughout my teens and early 20s, I was like this classic science fan type — reading deGrasse Tyson and Carl Sagan and Dawkins and so on, and indeed, Kahneman and Daniel Gilbert and people — really quite devoted to pop science, and did base some of my life decisions on these kinds of things. So it was quite difficult. Now, realizing that all science — and social sciences maybe in particular — are just really difficult, and that the methods we’re using aren’t really up to the task, or we haven’t powered the experiments properly, or we’re literally misusing the null hypothesis significance testing paradigm, or these sorts of things, that’s a really crucial part of becoming a scientist. You need to understand just how insufficient the average standards are in your fields.

SPENCER: Let’s jump to our last topic before we wrap up. I want you to tell the audience, why were you hesitant about coming on this podcast? I think it’s a very interesting reason.

GAVIN: It’s the first long excerpt of my voice on the internet. You only need a minute or so of audio now to spoof. Using machine learning speech generations now, there are already horrifying stories about spear-phishing, about people’s accounts and identity being stolen, horrifying story of this woman gets a phone call and it’s from her daughter and she’s screaming, and this horrifying voice in the background saying that she’s been kidnapped. She’s getting ready to transfer ransom money to this person when her daughter, who’s away on a ski trip, finally gets through to her on the phone. This is already happening. Because it’s only a minute of audio, I feel that anyone who really, really wants to target me could already do it relatively easily. But yes, the future of privacy and this kind of thing was on my mind.

There’s also a human transcript if you are a man of taste.


Post a comment:

Comments appear after moderation. markdown. Page just refreshes after successful submit.

Tags: psychology, interview


[RSS] /
[podcast] /