Sobolev_space

2.5K posts

Sobolev_space

Sobolev_space

@orthogonal_j

Postgraduate mathematics student

London Katılım Ağustos 2019
75 Takip Edilen23 Takipçiler
Sobolev_space
Sobolev_space@orthogonal_j·
@surplustakes @RenaudFoucart On reflection it does actually seem fine to leave long dissertations alone and to not really worry about LLM use. If the goal is to test real research ability then just mark harshly enough (e.g 75+ is publishable quality).
English
0
0
0
13
Sobolev_space
Sobolev_space@orthogonal_j·
@surplustakes @RenaudFoucart (Also FWIW in my university my friend studying a political theory masters informs me it has reverted to almost entirely in person exams except for the big summer dissertation.)
English
1
0
0
9
David Algonquin
David Algonquin@surplustakes·
The QTs here are an absolute masterclass in learned helplessness, where academics agree it is completely infeasible to spend a few hours of academic/postdoc time per undergrad per year grading exams at a cost in the low hundreds of dollars, when each student is paying thousands
Adam Zivo@AdamZivo

I find this discourse perplexing because the solution seems straigthforward: make university marks almost entirely dependent on lengthy in-person exams that combine handwritten essays with oral questioning. Why is this even a conversation? Are there implementation barriers or something?

English
4
1
34
2.5K
Sobolev_space
Sobolev_space@orthogonal_j·
@surplustakes @RenaudFoucart I mean also I think frontier LLM's would *easily* outperform expert human's on marking most undergraduate essays. Just have students sit in person taking exams and have an LLM mark it. If they complain who cares.
English
1
0
1
698
Sobolev_space
Sobolev_space@orthogonal_j·
@vit_tucek @littmath @gwh1tehawk I'm not sure I really agree with this. In whatever the IMO medallists field is (I presume you're hinting at combinatorics) the standards would just be raised, surely? It's not like is solving the biggest open problems... And is there really such a thing as a 'technically weak
English
2
0
0
45
Daniel Litt
Daniel Litt@littmath·
FWIW if your aim in doing mathematics was to achieve some kind of immortality, this was always (at best) an extreme longshot!
QC@QiaochuYuan

the most interesting pull quote from gowers' post, to me: > I sometimes get emails from people who are interested in doing mathematical research but are not sure whether that makes sense any more as an aspiration. I have a view on that question, but it may very well change in response to further developments. That view is that there is still a great deal of value in struggling with a mathematics problem, but that the era where you could enjoy the thrill of having your name forever associated with a particular theorem or definition may well be close to its end. So if your aim in doing mathematics is to achieve some kind of immortality, so to speak, then you should understand that that won’t necessarily be possible for much longer — not just for you, but for anybody.

English
27
12
325
24.2K
Daniel Litt
Daniel Litt@littmath·
@orthogonal_j @vit_tucek @gwh1tehawk Evaluation now relies (unfortunately) on publications. There’s no way the existing journal system survives the oncoming deluge without significant changes, for example.
English
2
0
2
127
Sobolev_space
Sobolev_space@orthogonal_j·
@littmath @aaswaminathan01 I think the most natural thing that will change in training is simply that graduate education will be increasingly taught and not research. If the frontier that is answerable by humans and not machines then more formal training will be needed to reach that frontier.
English
0
0
1
86
Daniel Litt
Daniel Litt@littmath·
@aaswaminathan01 I think advising, training, and education will have to change quite dramatically. I agree with you that this will be discouraging for some time. But it's coming to literally all professions!
English
3
0
13
601
Daniel Litt
Daniel Litt@littmath·
@orthogonal_j @vit_tucek @gwh1tehawk “Just hire the best” is of course the goal, but you need some evaluation mechanism to find the best people, which seems much more challenging when AI can one-shot some open problems.
English
2
0
2
139
Sobolev_space
Sobolev_space@orthogonal_j·
@littmath @vit_tucek @gwh1tehawk Why would it take any active effort to change how you hire? Just hire the best candidates, with the best research to date with the best future research promise. You would expect AI to raise the bar of everyone and maybe shuffle around capabilities, but why would that
English
2
0
0
90
Sobolev_space
Sobolev_space@orthogonal_j·
@TypeForVictory Yeah. A lot of the most visible stuff is cheap, the issue is no one knows what anything costs and that a library is probably the same cost as caring for a small handful of elderly people with complex needs.
English
0
0
0
11
James 🇬🇧 👑
James 🇬🇧 👑@TypeForVictory·
This obviously involves a major overhaul of local govt - centralise (or regionalise) care. Leave local govt to focus on public realm. Coupled with that, reforms to how we tax property etc.
English
1
0
4
410
James 🇬🇧 👑
James 🇬🇧 👑@TypeForVictory·
Tragically, I agree entirely. Only solution I can think of is to use a combined package of sweeping deregulatory reforms & changes to how we tax, rein in welfare, and invest hard in visible public realm - the stuff voters will see daily, so they *feel* like tax does something.
D@_Unknown_D_

Realistically I can’t see the UK cutting back on government spending until we are in IMF bailout territory, irrespective of which party gets in to power. Fiscal responsibility is unelectable at this point. People have become accustomed to existing state benefits and public services and won’t surrender them unless forced to.

English
6
4
61
5.8K
Sobolev_space
Sobolev_space@orthogonal_j·
@watling_samuel @danielluo_pi @JohnRuf6 inside it's radius of convergence. This follows from two very simple facts: polynomials are continuous, and the partial sums converge uniformly on every bounded interval.
English
0
0
0
21
Sobolev_space
Sobolev_space@orthogonal_j·
@watling_samuel @danielluo_pi @JohnRuf6 Also the example you gave is actually a really easy example. It's a standard result; would be considered 'seen' in an exam. It's unclear exactly what you mean, but it's typically packaged as a general theorem that says a power series defines a continuous function
English
1
0
0
14
Sobolev_space
Sobolev_space@orthogonal_j·
@SebMilbank FWIW I'm not actually sure social democrats do or should view 'social mobility' as a primary end goal
English
0
0
0
3
Sobolev_space
Sobolev_space@orthogonal_j·
@SebMilbank I mean this is basically what I believe haha. Small businesses are not remotely engines of social mobility. The actual driver of social mobility, which you do not like, is higher education.
English
1
0
0
3
Sebastian Milbank
Sebastian Milbank@SebMilbank·
Perhaps social democracy means Costa rather than independent cafes, supermarkets rather than grocers or butchers, and Waterstones rather than second hand bookshops, but I suspect that isn’t what 99% of social democrats want or intend
keir 📖🇵🇸@keir_92

people often complain that too many people/not everyone should be going to university, but the real problem is that, evidently, too many people have been encouraged to start businesses that simply shouldn’t exist.

English
7
6
96
7.6K
Sobolev_space
Sobolev_space@orthogonal_j·
@steamedhamms I.e the segment of the population that is 'left economically and right culturally' are actually not Labours voters and haven't been for a while. As Labour found out, people won't just swap around in your interest.
English
0
0
0
7
Sobolev_space
Sobolev_space@orthogonal_j·
@steamedhamms Yeah it's a really simplistic analysis. It's not the case that voters *uniformly* are in that part of the political compass: there is enormous variation. And voters behaviour is not determined efficiently before each election based on value alignment.
English
1
0
1
38
a rare photo of sean connery signed by roger moore
a rare photo of sean connery signed by roger moore@steamedhamms·
one reason nobody has successfully done this is the “right wing culturally” and “left wing economically” part intellectuals would build differs quite substantially from the “right wing culturally” and “left wing economically” that voters want
Tim Stanley@timothy_stanley

Today we discussed a focus group on the Daily T and it concluded that voters are right-wing culturally, left-wing economically. As they have been since I can remember. Yet we head into another election with no one having the sense to offer just that. Such a strange country.

English
9
6
211
22.9K
Sobolev_space
Sobolev_space@orthogonal_j·
@watling_samuel Out of interest do you know the reason why they don't just use harder tests? This issue doesn't arise in the UK because A Levels, and then especially entrance exams, are hard enough to discriminate between top candidates.
English
0
0
0
55
Samuel Watling
Samuel Watling@watling_samuel·
I mean he probably was, its just that Ivys are so competitive that most people are on the admission threshold and so it is quasi-random for most people.
Jeremiah Johnson 🌐@JeremiahDJohns

This guy is exactly what I was talking about the other day about how stupid the Ivy discourse is. This kid got into Duke, an absolutely elite institution, but goes on national TV to cry and moan he didn't get into Harvard and Yale Buddy, you weren't good enough.

English
3
0
24
3.3K

Keşfet

@surplustakes @RenaudFoucart @vit_tucek @littmath @gwh1tehawk @aaswaminathan01 @TypeForVictory @watling_samuel