Jonathan Libgober (same handle on other places)

6.2K posts

Jonathan Libgober (same handle on other places) banner
Jonathan Libgober (same handle on other places)

Jonathan Libgober (same handle on other places)

@JlibDoesEcon

AP in Econ at @USCDornsife. I do Information Economics & more. Best friend of Spinoza (Nosik) Petrov-Libgober.

Los Angeles, CA Katılım Eylül 2016
1.6K Takip Edilen2.4K Takipçiler
Jonathan Libgober (same handle on other places)
@JMenschEcon Sure, but I guess my point was that it’s not clear what this would be if the task is to come up with a new (positive) research program rather than to check that none of the previous potential (negative) counterexamples worked.
English
0
0
0
17
Jeffrey Mensch
Jeffrey Mensch@JMenschEcon·
@JlibDoesEcon I’ve done a back of the envelope calculation of power usage; it’s already *much* lower for AI queries. Sure, there is a lot of training that goes into the AI, but there is for humans, too, without the same economies of scale. And humans also don’t spend all their time on math.
English
1
0
0
29
Jonathan Libgober (same handle on other places)
I don’t think the question is whether AI will be able to do math better than humans—probably it would—but whether the amount of compute necessary to do this in a way that substitutes for humans is/will be feasible. I don’t think we’ve gotten much info on the answer to that.
English
2
0
2
363
Jonathan Libgober (same handle on other places)
In any case the most important thing I learned from my math undergrad is that humility is valuable. Not sure AI will be able to teach that more efficiently.
English
0
0
1
102
Daniel Litt
Daniel Litt@littmath·
@joshgans @JlibDoesEcon There’s a companion paper written by mathematicians who checked it (including me). I have no doubt the argument will be further checked by many more people, much more carefully than the median paper.
English
1
0
4
183
Jonathan Libgober (same handle on other places)
It also says this may be socially beneficial since the fact that we *assume* they are fishing gives stronger incentives to actually produce something impressive. Honestly, this sounds quite plausible to me. But the model doesn’t say I have to feel good about.
English
1
0
0
318
Jonathan Libgober (same handle on other places)
I’m a little bit bugged by the fact that AI is being used to solve hard problems but the details of how and what they’re doing is a bit mysterious and unclear. One of my papers (linked below) would suggest that we should just assume that the result is maximum fishing.
English
2
1
14
1.7K
Jonathan Libgober (same handle on other places)
The “even so” here seems underemphasized to me. It seems first-order important to know just how strong selection is here. It’s a very different story if OpenAI searched over all known problems and could solve one, versus it solved the first one they picked.
Daniel Litt@littmath

@yacineMTB I think it’s almost certain the cost for this one result was way lower than getting a human expert to produce a result of similar quality. Presumably they tried *many* problems but even so…

English
6
1
31
5.4K
Jonathan Libgober (same handle on other places) retweetledi
Andrey Fradkin
Andrey Fradkin@AndreyFradkin·
Prediction markets are our best tool for aggregating dispersed information. In a new piece with @brian_jabarian and Andrew Koh, we explain how to use them to learn about AI's near-term labor market impacts. For these markets to work, donors (including large labs) need to seed capital to break the chicken-and-egg problem. Link below for the article.
English
2
18
48
6.2K
Jonathan Libgober (same handle on other places) retweetledi
alberto bisin
alberto bisin@albertobisin·
"Social-science genomics" might sound scary ... but this is real serious work not to be missed - which hopefully will change your mind regarding the new advances in the field ... not to be missed.
AEA Journals@AEAjournals

Forthcoming in the JEL: "Social-Science Genomics: Progress, Challenges, and Future Directions" by Daniel J. Benjamin, David Cesarini, Patrick Turley, and Alexander Strudwick Young. aeaweb.org/articles?id=10…

English
0
2
30
6.6K
Jonathan Libgober (same handle on other places)
(3) even if you’re techmaxxing….so are lots of people these days. Generic applications can be a problem too. Having something interesting or unusual can help you stand out. Doesn’t need to be anything crazy, but just doing “what you’re supposed to” isn’t enough to get attention.
English
1
0
15
1.1K
Jonathan Libgober (same handle on other places)
(2) As others have said the path is long and arduous. As with many things in research, motivation is a huge variable worth paying attention to. If you can make a successful blog people will read, why not? Do what you can be productive doing. That pays off in the long run.
English
1
0
15
1.2K
Jonathan Libgober (same handle on other places)
As someone who does admissions regularly, I disagree with this take. (1) There is no one path to a PhD. And the best path depends highly on what you want to do. Theory? Sign up for grad analysis. Not necessarily a great move if you want to do dev (even if it wouldn’t *hurt*).
Luke Heeney@heeney_luke

Zane is correct. To get in to an econ PhD, anything other than math classes or econ research is a waste of time. However I think it is sad for our profession that this is the case. Economists who write and spend time with policy makers or industry are much more influential.

English
1
2
53
13K
daniel (michelle steel hate account)
how long is a reasonable amount of time for a paper to be on the editor's desk after all referee reports are in?
English
1
1
10
3.9K
Jonathan Libgober (same handle on other places)
@reillysteel For instance if students overvalue stipends then a small increase in stipends might lead to students ignoring horizontal differentiation, leading to departments “winning” more but with worse outcomes for students. Not saying I believe this, just that it’s unclear.
English
1
0
0
523
Jonathan Libgober (same handle on other places)
@reillysteel Normative criterion is that PhD students should be paid an amount equal to the marginal value of labor. Which can include prestige. You might be right about 2, but this seems unclear and could go in the other direction.
English
1
0
0
618
Reilly Steel
Reilly Steel@reillysteel·
I’m a former Princeton PhD student. Based on my experience, this is missing an important part of the picture. PhD students have to do research, which is valuable to the university because it enhances the prestige of the university. 1/n
Ashvin Gandhi@ashdgandhi

I'm a former Harvard PhD student. Based on my experience, current social science students probably make a bit over $250k + healthcare over 5 years, with just 784 hours of required TA work. That's almost $320/hour for the "work" and the rest is classes and your own research.

English
12
13
514
123.5K