Thom Scott-Phillips

We're All Trying To Find The Guy Who Did This. #AI
www.theguardian.com/technology/2...

Lot of sense here

For sure. The reasons/flaws are much clearer in some cases than others, those tend to be cases where they will get caught out sooner rather than later, and supporters in those cases are daft. But still, it's all a continuum I think

I sorta think this is true for all PMs. The job is v demanding and the evidence is there at the beginning about the way in which it will catch up with the PM in the end. Some do much better than others of course, but it gets them all

Photo of book. Sociopath, by Patric Gagne.

Incidentally, I am currently reading this. Very well written. And what I know of SBF very much fits what I am learning here. (Don’t know enough about Altman and others.)

I would not be surprised if Sam was a pretty popular boy’s name among affluent coastal families in the 1980s. Not that I know that world, but from the outside it feels like it fits

Honestly, the thing that is much more scary than people realise is the financial consequences of the AI sector going belly up. I’ve read enough about 2008 and what nearly happened then

John Lanchester · He-Said, They-Said: Crypto Corruption Crypto is an ideology, an anti-government, individualistic belief system, one that Sam Bankman-Fried didn’t really...

Did you read anything about SBF? He was guilty of the same dumb utilitarianism

www.lrb.co.uk/the-paper/v4...

It's funny to think that my preference for urban living in areas with dense, mixed-used, walkable neighborhoods, multi-family housing, small businesses, and diversity probably stems from Sesame Street. Like an early education in Jane Jacob's urbanist theories.

As a teenager growing up in inner city London in the 1990s, I was twice approached at bus stops by older boys carrying knives. I gave them what I had. I well believe that UK cities are much safer now

*Yet again* you attribute to me views I don’t hold, and insult me on the back of it! I didn’t say these issues were unimportant, I said I haven’t been following them

And that really is my last word here

It's not comment I objected to, it was the condescending tone

You misread because I did not make any comment on things I don't understand. On the contrary, I did the exact opposite! I flagged that I have not followed these issues and I wondered if I should update my priors

And in response, you threw condescension at me, twice!

No more replies from me here

No idea why you think it's ok to write something so condescending ("..you should..."), and why you choose to misread what I wrote. I *explicitly said* I wasn't following these issues closely

But again, this was an aside in my first reply. The big problem is power asymmetries

Honestly I think it’s hypocritical if/when people complain that there are abuses of power in academia (eg PI over postdocs) but also sit on job panels that ask for letters of recommendation

Without the letters it is easier to hold the decision maker(s) to account. They have a well defined task to acquire relevant info of their own and make the best choice. With the letters they can push responsibility elsewhere. “We made the best choice with the info we had, but the letter writer lied”

Are we misunderstanding each other? :)

I'm on a job panel. Candidate A has letter saying they're a great colleague. They turn out to be an awful colleague. Someone asks me why we chose him

Many people would rather say "What can you do if letter writers are gonna lie?" than "I messed up"

Sure, but that's then squarely on them. Letters allow panel members to push at some of the responsibility elsewhere. Or to put it another way, letters facilitate defensive decision making

(And just to add, I do not think this is their worst effect: that would be the creation of power asymmetries)

"We didn't realise he'd be a bad colleague. His reference letters were outstanding!"

I think the whole norm of letter writing is a sector wide moral failure. Aside from being inefficient, it actively creates imbalances of power and absolves decision makers of responsibility

In every other sector recruitment is based on institutions’ own assessment process

I think the whole norm of letter writing is a sector wide moral failure. Aside from being inefficient, it actively creates imbalances of power and absolves decision makers of responsibility

In every other sector recruitment is based on institutions’ own assessment process

Right. But again, all of those three needs are true of the 1-to-1 format also. I'm curious to know if the effort/costs are substantially higher for 1-to->1

Are my priors way out? Myself I need zero wrangling if someone wants to listen to what I have to say!

You think? Just looking at academia, most academics are always happy to talk about their areas of expertise and most do not get much public attention

Right. I see this and it makes sense. But In Our Time manages it well. So does Start The Week (admittedly that is closer to being three mini-interviews rather than a synthetic conversation). So I think it's possible even if hard

About podcasts. Many have 1-to-1 interviews with thinkers, writers, creators, etc. But why do they rarely have >1 such guest at the same time?

In Our Time does it, but that is a BBC production that started decades ago. Is the format much more expensive and time consuming than the typical podcast?

That is especially disappointing from Stock. I am not following the culture wars closely but she seemed, to me, to be someone who still has her head screwed on, even when I don't agree with what she says. What you present here suggests that I need to update my priors

“I’m not angry, just very disappointed”
- Every parent to their teenager, at least once

I am not the parent of modern institutions of science, but I feel very much the same. I thought they would be much more conscientious than they have turned out to be

"Instructed in assembly line procedures" is a good description that applies to all science (& philo).

Teaching research design as a creative act rather than a cookbook is hard.

Most grad students & ECRs want to be accepted into a club/identity. That means learning its norms, not free thinking.

ABOUT SCALARS:
Ira Noveck and I published in 2007 "The why and the how of experimental pragmatics: The case of scalar inferences" in _Advances in Pragmatics_ (N Burton-Roberts ed.), republished in D. Wilson & D. Sperber _Meaning and Relevance_ 2012, as ch 14.