Just and fyi alain, I would advise editing your spelling here .......

So proud of my fellow nerds 🥲

Yet more proof of my, everything is basically a bunch of filaments, theory. Excellent work by all involved.

Three excellent people and a great project. You will learn mathematical modelling and how to interact with our experimental colleagues, proper interdisciplinary science. Plus we do excellent cookies at lunch time meetings 😎

I suspect words might be had if this particular copy was found on your bedside table...

Did you notice anything different with our approach to crosses/corners today ?

I sense a paper 2 with twist then !

Proteins get loads of their writhe from helical coiling, but proteins have lots of non local binding to hold it in place.

I'm looking forward to hearing about that. On this paper, my only quick thought is that the one clear way a curve can build up significant writhe, which doesn't have anything to do with knotting, is helical coiling. Is it just that's teens to be prohibited in these in these simulations?

Dragon content represents an excellent first post!

I can't access that link, is it the epple article ?

To be fair since I moved on here from X/twitter I've seen almost non of that and feel alot better about the direction of the team for it. You'd follow someone saying something interesting about tactics and the algorithm would throw ten people just mouthing off about the team. Not here so far...

That's going to be quite an impact case study !

I'm with the physics majors on that, "those? They're just some projects I have on, the go"

This got a bit programming specific, but I think there are equivalent issues in most subjects.

I'm sure there will be loads of issues encountered on the way, but I think it's best to start now and get it right, rather than fight the inevitable for years.

So you can focus more on fundamentals.

But then, some things, like say plotting scripts I don't really think are fundamentals, and god knows I've wasted so much time with them over the years, and which cost loads of time in class, can just be llm'd out.

In short I think you can still give them their fundamental principles, just the style needs to change.

In the past we could get away with the less labour intensive (for us) approach. But rather than bullishly stick to that and waste loads of time developing methods to stop students using llm's just accept we need to alter the approach.

With gpt they can nail that without thinking. But if instead you present the code, and ask them what it is calculating, they can't then "cheat" but they still get the same point.

Okay! Well the original notion was don't ban llm's. There are simple points like, e.g right now we would show a student code to add up 1 to 10 with a for loop, then as to modify to only do odd numbers or the square of them. Then turn it in.

So we should teach how to use it practically.

Yeah, but then I guess you need to arrange more complex problems where (you've checked) gpt wouldn't get the whole thing but can construct the constituent elements. I guess what I am saying is that is often likely how it will be used in practice in the "real world"

Basically force them to asses its output to engage with it, rather than keep the current structure where they can just hand in its output without thinking and try to ban it.

We're having this with our first year programming (python) gpt can basically nail all the assignments with little input. The thinking is we should be asking students to critique existing code (point out bugs, discuss efficiency).

I had you down as a botanist