Fanfan 🐻 🌲 🏔

https://social.sciences.re/@jaztrophysicist

Research #Ecology #Theory #Physics #Plasmas #Fluids #Astrophysics #HPC #CNRS

#Nature #garden #photography #MountainBiking 🚵‍♂️

Influenceur #ErmiteDescolllines et #lariegecestmoche

Upright #Bass #Jazz #Musician

"Musico-scientifique chelou", "paparoizzo", "chieur atomique" (Qi2Navet, bourrin malicieux☝️ (npub1juv…xv5g))

"Fait autre chose avec des équations mais on sait pas quoi" (Madame Mollette (npub1gxy…h566))

PP attaché à un grillage de poulailler par Claire WXYZ (npub1vmr…lpw8)

Banner: upright bass leaning against blackboard with equations and sketches

Website: https://www.francois-rincon.org
Ecology research blog musings: https://lookingup.francois-rincon.org
Orcid: https://orcid.org/0000-0003-4326-501X
Learning shit: since 1978

Public Key

npub13lvnxh64a6gg557ev2tvgghpad0fwdkt2gefamlalgx6500zs4rs65sne5

NIP-05 Address

jaztrophysicist_at_social.sciences.re@momostr.pink

Profile Code

nprofile1qqsglkfnta27ayy220vk99kyyts7kh5hxm94yv57al7l5rd28h3g23cpz3mhxue69uhhyetvv9ujuerpd46hxtnfduqs6amnwvaz7tmwdaejumr0ds2jq89h

Publishing to

public.relaying.io

nos.lol

relay.damus.io

Author Public Key

npub13lvnxh64a6gg557ev2tvgghpad0fwdkt2gefamlalgx6500zs4rs65sne5

Show more details

Published at

2026-04-11T23:54:12Z

Kind type

0 Metadata

Event JSON

{ "id": "c0eb92e816bf7eaccf9b388bef01083df239a938a2293098cf8dde1cf182c1ec", "pubkey": "8fd9335f55ee908a53d96296c422e1eb5e9736cb52329eeffdfa0daa3de28547", "created_at": 1775951652, "kind": 0, "tags": [ [ "t", "Astrophysics" ], [ "t", "photography" ], [ "t", "ecology" ], [ "t", "nature" ], [ "t", "jazz" ], [ "t", "physics" ], [ "t", "theory" ], [ "t", "garden" ], [ "t", "musician" ], [ "t", "hpc" ], [ "t", "bass" ], [ "t", "cnrs" ], [ "t", "mountainbiking" ], [ "t", "plasmas" ], [ "t", "fluids" ], [ "t", "lariegecestmoche" ], [ "t", "ermitedescolllines" ], [ "proxy", "https://social.sciences.re/users/jaztrophysicist", "activitypub" ], [ "L", "pink.momostr" ], [ "l", "pink.momostr.activitypub:https://social.sciences.re/users/jaztrophysicist", "pink.momostr" ], [ "-" ] ], "content": "{\"name\":\"Fanfan 🐻 🌲 🏔\",\"about\":\"Research #Ecology #Theory #Physics #Plasmas #Fluids #Astrophysics #HPC #CNRS\\n\\n#Nature #garden #photography #MountainBiking 🚵‍♂️ \\n\\nInfluenceur #ErmiteDescolllines et #lariegecestmoche \\n\\nUpright #Bass #Jazz #Musician\\n\\n\\\"Musico-scientifique chelou\\\", \\\"paparoizzo\\\", \\\"chieur atomique\\\" (nostr:npub1juvs7p8jfx5la5m7yj64n8yft707qec8xpdc4zzcc3tmk0xw96kqypxv5g)\\n\\n\\\"Fait autre chose avec des équations mais on sait pas quoi\\\" (nostr:npub1gxyktp6y4ngudyja5gydqvhjpzr2nuc4m2kr4pa58h5uuz3tn08qv7h566)\\n\\nPP attaché à un grillage de poulailler par nostr:npub1vmrrc93dmayg80vrtps7vsnhwpcx5mf43uqkx3fe8suqcwrt95jqaelpw8\\n\\nBanner: upright bass leaning against blackboard with equations and sketches\\n\\nWebsite: https://www.francois-rincon.org\\nEcology research blog musings: https://lookingup.francois-rincon.org\\nOrcid: https://orcid.org/0000-0003-4326-501X\\nLearning shit: since 1978\\n\",\"website\":\"https://social.sciences.re/@jaztrophysicist\",\"picture\":\"https://social.sciences.re/system/accounts/avatars/112/145/032/078/806/989/original/d9147e8ea08075d8.png\",\"banner\":\"https://social.sciences.re/system/accounts/headers/112/145/032/078/806/989/original/d16f535dade9f9e3.png\",\"nip05\":\"jaztrophysicist_at_social.sciences.re@momostr.pink\"}", "sig": "5f02c8592be044556d679c72583b627a379559c5f72c77c52f0ca677a358fddc09166490f68e909aba7b0f3104926f28a304e13b181cca18e37dea00e3c8d346" }

Last Notes

2025-10-03 05:46:51 UTC

#Physics is such an old field it's a real bitch.
1. Have a nice promising idea all by yourself
2. Find out a well-known guy had it 30 years ago
3. Refine the question so as to go further than the 1st dude.
4. Find out another even better-known guy who looked at this particular question 50 years ago.
5. Look at references in the paper.
6. Find out that Eddington first had a look at this in the early 1930s.
7. Give up and go back to your usual infinitesimally incremental shit or persevere by reading the 200 papers that you discovered along the trail of references, at the risk of facing an even more depressing surprise.
So, after thinking I had an original thought on Monday, here's me studying papers from a century ago to figure out if idea can be salvaged somehow.
#InfiniteLoop #ResearchIsHard #VisMaVieDeScientifique

2025-09-03 12:15:12 UTC

- reply

Thank you, I guess it means I'm a considerate person in posting stuff that is not going to be a waste of people's time 😉

2025-08-15 07:23:31 UTC

- reply

ALERTE MARTIN-PÊCHEUR 🚨 #birding

2025-08-15 07:05:44 UTC

#LeTarnCestMoche
https://social.sciences.re/system/media_attachments/files/115/031/502/624/835/739/original/1ed15869e0f6e4f3.jpg
https://social.sciences.re/system/media_attachments/files/115/031/502/976/797/913/original/a63ef7c8a78fcd72.jpg

2025-07-22 07:18:22 UTC

Bonjour, aujourd'hui c'est le jour tantadandu depuis 3 mois, safari photo-vtt avec junior sur les hauts plateaux 🤩 🥰 #levercorscestmoche

2025-07-18 18:53:15 UTC

Aujourd'hui une toute autre ambiance avec la visite de Valchevrière. J'ai fait des belles photos, dignes du lieu j'espère, que je partagerai plus tard.

2025-07-14 16:47:59 UTC

Mastoterelle #MacroPhotography #naturephotography
https://social.sciences.re/system/media_attachments/files/114/852/595/697/184/146/original/e61ced10d1168c28.jpeg

2025-07-10 09:51:39 UTC

Could someone refresh my understanding of why, oh f* why, we need to wait between 6 and 12 months to learn whether our 10% success rate grant proposals are getting funded ? What is the f* bureaucracy doing with them all that time ? Why is it so hard to ask reviewers to send a review within a month and then have a committee that figures out the few outstanding proposals and then randomly draw a few among the remaining good ones, and that's it ?

2025-06-16 16:53:54 UTC

So, there was quite a lot of talk, in the workshop I just returned from, about the use of machine-learning for nonlinear physics (like astrophysics, atmospheric, or ocean turbulence).
In the end, my take on this problem has not changed much and can be summed up as follows.
- It can do some rather good things (like accelerating things or making things a bit more precise) on *interpolation* problems (i.e. to do inference on problems it was explicitly trained for, in physical regimes it was trained for based on full-physics simulations), and may even help to limit the use of full-scale HPC/numerical resoures in this context IF USED SMARTLY AND REASONABLY (a big if when I look at the way many astronomers are using it).
- It essentially remains untrustable and uncontrolled garbage for *extrapolation* problems, i.e. to provide new results in physical regimes well-beyond those it was trained on, and beyond those that we can explicitly simulate with physics codes. It notably remains utterly terrible when there is no proper separation of scale in the problem, that is for, well, most of the hard key problems we have to deal with in this field.
Most of the hype is about the latter, though. The progress with interpolation/fitting is real, but is much more incremental and much less flamboyant than the hypothetical conceptual breakthroughs promised by the proponents of AI. That ML performs well on interpolation problems is not particularly a surprise or controversial in itself either, as these things are essentially giant fitting factories.
All of this makes it very clear that it is critical to *define* what AI means, and the kind of things one wants to apply ML or other techniques on , before any useful and reasonable conversation on their merits can be had. #AI #physics #machinelearning

2025-05-25 10:41:39 UTC

Paper-editing session POV #SilentSunday #Science #vismaviedescientifique
https://social.sciences.re/system/media_attachments/files/114/568/043/664/982/433/original/0a8d4b00cb16dcbe.jpg

2025-05-11 10:21:45 UTC

La macro manuelle >> Emmanuel Macron
#insectodon #nature #GardenPhotography #macrophotography #FleurisTonFil #BloomScrolling
https://social.sciences.re/system/media_attachments/files/114/488/695/210/859/001/original/e4d92394378379d6.jpeg

2025-04-09 21:34:09 UTC

Philippe Poutou qui devient libraire, beh je suis sincèrement content pour lui, il va surement pouvoir vivre une meilleure vie et c'est tant mieux comme ça.

2025-03-23 20:04:21 UTC

- reply

Singing in the rain #Birding #BirdPhotography
https://social.sciences.re/system/media_attachments/files/114/213/532/242/516/182/original/55981d237160d6f2.jpeg

2025-03-23 18:45:45 UTC

🧵 #Zanimo après manger 😏 #Birding #BirdPhotography
https://social.sciences.re/system/media_attachments/files/114/213/242/881/715/946/original/98d11c86e15721e5.jpeg

2025-02-25 22:09:13 UTC

C'est nul, avec mon filtre anti Musque, tous les pouets qui parlent de trucs au mois de mars sont maintenant masqués. C'est embêtant vu le calendrier.

2025-02-19 18:28:27 UTC

- reply

A truly wonderful flower #macrophotography #Bloomscrolling #FleurisTonFil
https://social.sciences.re/system/media_attachments/files/114/031/962/751/607/016/original/94e1dfc25befcb9c.jpeg

2025-02-19 18:24:53 UTC

- reply

#Bloomscrolling 💚 #FleurisTonFil #macrophotography
https://social.sciences.re/system/media_attachments/files/114/031/945/753/361/201/original/893438ddd9a435d0.jpeg

2025-01-22 21:51:48 UTC

- reply

(and it appears marking all media as sensitive by default is inefficient against that too).

2025-01-22 21:49:38 UTC

- reply

But we may want to read the contents of the toot itself, without seing a brutal visual. So hashtag filtering is not the solution here.

2025-01-22 21:41:07 UTC

#Mastodon we have a problem with link embedding bypassing control mechanisms for exposure to shit, and showing unsollicited pictures (say, of a real bastard making a nazi gesture), without the possibility to filter that out a priori (unless people who post these links realize by themselves this displays the picture and use a CW -- but they may not be aware of this when they post, as it takes time for the embedding to show up after a post).
Yesterday I had to remove a link from one of my toots after realizing a posteriori what the embedding was showing.

2025-01-18 18:14:36 UTC

- reply

Un couple plus classique, mais néanmoins choupi
https://social.sciences.re/system/media_attachments/files/113/850/714/246/115/946/original/9355f8e5c2449ae9.jpeg

2025-01-15 22:19:52 UTC

- reply

Admittedly my formulation was not very clear either, I kind of fumbled in trying to question Santiago's argument without sounding harsh. Peace.

2025-01-15 22:07:55 UTC

- reply

It looks like a very weak effect if the angular resolution limit due to atmospheric turbulence is
λ / r0
and r0 goes like \lambda^{6/5} !

2025-01-15 22:04:43 UTC

- reply

Yes thank you I know that (I'm a physicist). My point is that the explanation can't be it's because of "molecular scattering" being different in different wavelengths if the point is to limit turbulent refraction effects.
The explanation appears to be that refraction effects are also smaller in IR
https://www.astronomik.com/en/infrarot-passfilter-infrared-pass-filters/proplanet-642-bp-ir-passfilter.html
"Why Infrared? Every observer knows the flickering stars and planetary disks when looking through an eyepiece or at your monitor Slight differences in temperature and preasure result in a different index of refraction for "bubbles" of air. As these bubbles move, the image moves too! This effect is called "Seeing".
This change in the index of refraction becomes smaller the further you go to the infrared (longer wavlengths), and due to that, the Seeing becomes better too!"

2025-01-15 21:49:36 UTC

- reply

But turbulent refractive scattering is distinct from Rayleigh scattering ? (I've been wondering too).

2025-01-15 13:23:37 UTC

*quote* Naturally *quote*
https://www.theguardian.com/us-news/2025/jan/15/polluter-pay-bill-climate-disasters

2024-11-17 16:29:23 UTC

- reply

they are categorizing in a thematic way, not in an account/person specific way. People are multidimensional and i don't like the idea of cataloging them in a one-dimensional way as if they were a product serving an (information) function to be consumed. That's not my vision of social media at all.

2024-11-17 15:58:03 UTC

- reply

i guess i just dislike categorizing(and mimicking other stuff)

2024-11-17 15:12:30 UTC

Why do people need followpacks, and why imitate other platforms sheepishly ? Mastodon has its own discovery mechanism, it's called hashtags and you can follow them.