indeed he shared it!

here: yourfutureadventures.blogspot.com?m=1

yeah I feel twitter is better for silly thoughts somewhat, but... be silly! Whimsy is the best

you don't have to delete it, but you also don't need to send it/publish it

name them

1 in 3? I like those odds

Friends are there for you Manki! Remember that

A fun one is: dS>0

Maybe people are doing it already and we don't know! Would love to see if we can get some of the best known string theorists access to the cutting edge LLMs and see what they'd do

best thing to do in a new city

Anyways, I do hope to see in my lifetime that people find 1 CY compactification that gives the Standard Model and everyone is happy, and then people find a completely different one that also gives the Standard Model and people are a bit unhappy again, and that ML will be used for this

I think there's some cool stuff physicists can do to better understand ML itself (I loved that one paper of Vasudev and friends) as ML has lots of empirical stuff and not enough theory (was definitely the case 5 years ago :p)

Sending a legion of LLMs to do hep-th research... I think not yet

This might be the case, and that paper of Matt Schwartz is a nice example of this. But if that paper took Matt 2 weeks like he claims, I'd rather not have Matt spam the arXiv with ~24 papers a year either

I think it really sucks, and if having something as simple as "a place I know I will still be at 5 years from now" from the hep-th "industry" is so hard to get, then it's on them :)

of course a lot can be said about *that* director search :p but that's something better done off the record

I actually like that many string theorists have switched gears to ML methods, because there's been some really good successes. Also ML itself is an interesting thing :)

There are some symptoms, like the Potsdam Albert Einstein Institute not being able to find a director after Hermann Nicolai's retirement. But also greener pastures are everywhere!

What are you learning now, Manki?

Really sad...
haruhisa-enomoto.github.io/acadexit/

Happy New Year, Manki!

Hope you have a good 2026, Manki!

AI is also exciting because so many things work empirically without that much of good thery behind it. It's ripe for doing some fun experiments, and that's what physicists like are in their heart

Anyways, here's possible interpretations I agree with somewhat:
An academic life is selfish:
-Because it's something you really want
-Because you sacrifice family and friends
-Because you use your skills for science (and mentoring) rather than helping other companies profit $$$$

Do say more, Manki! Pic potentially related

It's a terrible feeling, going back to one's country and feeling like a stranger. Where is one supposed to belong, then?

There are also so many string people in the (ML)∩(Physics) community! I think physicists can really thrive there (not only do good work, but also feel they're in good community and using the skills they've learned)

I once helped supervise a master's student and it was very rewarding

another failure mode is: trying to please users

In this specific case, Midjourney is an example...

It's also really really fun to supervise people too, I have to admit! :D

This might be true! I wouldn't want that much attention and responsibility though. I'm very content with being able to enjoy the work of other people (maybe in a niche field) and have them enjoy mine

for a brief moment I thought they were rocks, looks delicious!

Very pretty! Circle inversions are good way to make fractals when the circles touch :) here's a cool video of these by Kento Nakamura www.shadertoy.com/view/XslyzH which he did using a package he developed, SchottkyLink