Mrs Beanbag

Creator of Amiga game Mr Beanbag. Demoscener.

Pronouns: she/her
Soundcloud: https://soundcloud.com/mrs-beanbag
Languages: en-GB(native) nl-NL(100% van de uil) de-DE(beginner) es-ES(beginnerer)

Public Key

npub19khyrrjtmw8u7n5dhtx2z0fr4l82dga074d82xm6qmvv45cxktqsfsvmq6

NIP-05 Address

[email protected]

Profile Code

nprofile1qqszmtjp3e9ahr70f6xm4n9p8536ln4x5whl2kn4rdaqdkx26vrt9sgpz3mhxue69uhhyetvv9ujuerpd46hxtnfduqs6amnwvaz7tmwdaejumr0dskp8m9u

Publishing to

nostr.land

Author Public Key

npub19khyrrjtmw8u7n5dhtx2z0fr4l82dga074d82xm6qmvv45cxktqsfsvmq6

Show more details

Last Notes

2026-08-29 11:52:28 GMT

- reply

i have been thinking about this very claim lately, and my conclusion is that it isn't *even* a tool. a tool is something you control, and with practice, master. you don't control an LLM. you just sort of, ask it, and it obeys, more or less, according to its interpretation of what you asked for. that's not a tool, it's a servant.

2026-08-28 17:47:02 GMT

- reply

you know, i actually went back and re-read my undergraduate general relativity textbook about a year ago, but black holes are such fundamentally weird objects, i just get more and more questions every time i think about them. One thing i remain sceptical of, though, is the idea of any physical singularity or otherwise infinite physical quantity. (yeah they're fine in maths, i'm not an ultrafinitist lol.) so every time i get convinced there must be some loophole somewhere, thinking this must be to do with a bad co-ordinate system or applying special relativity thinking to general relativity or something. then i have to go look at the nice picture of the Kruskal Szekeres co-ordinates and my plans to go black hole bungee jumping get shelved again.
anyway one strange thought that still occurs to me is that for an extended object such as a neutron star, the strength of gravity must reduce as one approaches the centre. thus i have an intuition that during gravitational collapse, the black hole actually forms *at the surface* and spreads inwards. so i have to wonder about what you would see or feel, if you were inside a low-density object many light years across, that just happened to become smaller than its Schwarzchild radius. one would not expect a singularity to suddenly appear at the centre.

2026-08-25 19:02:08 GMT

- reply

i gotta be honest i have a hard time untangling the apparent effects of the "co-ordinate singularity" from the actual physical effects. but i guess in other words, the thrust you need to "hover" at the event horizon is basically infinite, regardless of black hole size.

2026-08-25 16:53:38 GMT

- reply

i've often wondered about this stuff. does the concept of "event horizon" only apply to objects under unpowered projectile motion? because a black hole large enough would have a "surface gravity" that's low enough for a rocket to accelerate away from. so what exactly stops it going in and coming out again?

2026-07-29 13:41:10 GMT

- reply

thank you! this is an excellent review. it's validating to find out that the approach i've been taking isn't entirely beyond the bounds of acceptability in polite physics society, and also (i think) a relief that my own particular cake mix isn't on the list already (otherwise i have wasted a lot of thinking time when i could have been going to sleep). but the bit about gravi-weak unification caught my eye in particular, since that's where i've been putting most of my attention. it seems to me that tacking on an additional SO(6) afterwards isn't such a big deal after doing all that. Number 3 on the "to do" list remains, so i shall have to go and work out how to do it.
(i should know already that SO(1,9,C) is the same as SO(10,C) after reading that book about Twistor space lol. tbh i don't find such complex spaces very compelling.)

2026-07-28 21:41:21 GMT

- reply

so yes.. i think the theory i have in mind would come under the broad umbrella of "gravi-GUT" theories. it seems there are a few attempts out there. here is another based on SO(1,9, C) https://arxiv.org/pdf/2510.11674 (i shall read it tomorrow)

2026-07-28 20:12:10 GMT

- reply

ah! i think i stumbled upon this one the other day, but i didn't read it. i think i shall have to, now. oh it mentions SO(7,7) which i was looking at previously!

2026-07-28 18:47:56 GMT

- reply

i guess that explains why i never see it. i suppose my questions would be: does that actually matter if we are going to break it into compact subgroups? maybe that's the reason why it *does* break into those subgroups? and also, if we are to treat gravity, we are kind of forced to confront the unavoidably indefinite signature of the Lorentz group.

2026-07-28 16:21:09 GMT

- reply

i meant the SU(2) x SU(2) subgroup of G2.
I have skimmed through your paper about GUTs that you sent me. i am happy to see you prove that the GSM exists at the intersection of Georgi-Glashow and Patti-Salam, which is a thought that also occurred to me a couple of years ago. On which note, another thing that occurred to me, which i have not seen anywhere else, is the equal possibility of unification in Spin(4,6), which yields both Patti-Salam and SU(2,3). the intersection of any 2 of these 3 groups would seem to be the GSM. so i don't know if there is any significance to that.
I have encountered enough crank theories in my time tbh, i'm not surprised you get them in your inbox so i won't bother to add to it. I even got one myself as a PhD student. There was one circulating on mastodon a few months back that amounted to nothing more than setting all the physical constants to 1 and calling it job done!

2026-07-28 12:03:33 GMT

- reply

thank you for replying! i realised the GSM wasn't a subgroup of G2, but i couldn't find so much information about its complexification. i had concluded that it also wasn't a subgroup of that, although the SU(2) x SU(2) subgroup becomes SL(2,C) x SL(2,C) which i found very interesting, since i was looking for a group that would also contain the Lorentz symmetries. your talk about complexified Octonions made me temporarily doubt myself however, but i realised on looking at your paper again that you were in fact dealing with Octonion-valued matrices. I will get my head around your work sooner or later, no doubt, but having to learn all this stuff outside of academia continues to be a challenge.
On which note, i wondered if you could advise me on this. What should one do exactly, if one has a theory of physics, but is not connected to academia? It seems to me to be hard to get taken seriously (and for good reasons).

2026-07-25 18:15:59 GMT

- reply

i once met the guy who owned the company i used to work for, and one of the things he said has stuck with me ever since. he said he finds himself "trying to be careful what i think, in case it makes it real". the dude has had so much sheer good luck in his life that he's almost started to believe that he can control reality with his thoughts. to his credit he seemed self-aware enough not to consciously believe it. but if you could imagine the effect this kind of thing must have on someone who owns a significantly larger portfolio..

2026-07-25 17:34:09 GMT

- reply

i was having a think about it this morning. the thing i find daunting/uneappealing about Octonions is that they don't have a matrix representation (whereas of course the standard model gauge group does), which makes them very strange and unfamiliar things. however, there is G2. and i remembered that some time ago i was playing around with complexified G2 wondering if i could get the standard model gauge group out of it. this started from the discovery that it's a manifold of special holonomy (Ricci flat) as a subgroup of SO(7,7). I didn't know how to get anything useful out of it though so i went in a different direction instead. But i'm wondering now after reading about your work if it implies that the standard model gauge group actually *is* a subgroup of complexified G2 afterall?
The pseudo-Riemannian Berger list doesn't seem to get quite so much attention in the literature but to me it's very intriguing.

2026-07-23 18:15:52 GMT

- reply

ok i've tried to read the new paper but it's still a bit beyond me at this point. it's been a long time since i was in academia and i'm trying to get back into physics (i'm probably one of those cranks lol).
anyway i have seen octonions cropping up quite a lot lately. tbh they give me the heebie jeebies. (also every time i saw ℂ ⊗ 𝕆 i couldn't but read it as "Cheese and Onion".)

2026-07-14 11:28:30 GMT

- reply

is this related to the video on your website that i watched yesterday? interesting stuff

2025-12-05 16:56:21 GMT

- reply

oh yeah of course it does! obvious now you say it lol. but where you have, for instance, a special holonomy group that is, for instance Einstein or Ricci Flat, that property is a property of the restricted holonomy submanifold, or of the ambient SO(n) manifold, or both? or can the SO(n) manifold be basically anything?

2025-12-05 16:24:01 GMT

- reply

sorry, i haven't really studied this and i've been out of academia for well over 20 years and just trying to grasp some concepts that are to be honest far to advanced for me lol, but that's worked as a learning strategy before so..

2025-12-05 15:56:31 GMT

- reply

i mean, for a given restricted holonomy group of SO(n), Hol0, can that be considered a curved manifold embedded in an SO(n) manifold? and can the SO(n) ambient manifold in this case be flat?

2025-12-05 15:29:22 GMT

- reply

ohh, yeah ok 🤔 but in some special cases, those numbers can be smaller, right? for instance a de-sitter space (1,n) embedded in (1,n+1). in fact what i really want to understand at this point is the relationship between embedding and holonomy. i mean a restricted holonomy group of SO(n), which is a subgroup so can that also be an embedding?

2025-12-05 14:43:00 GMT

- reply

oh you might know the answer to this then, because i was looking for a similar result for pseudo-Riemannian manifolds. if you have a n+m manifold, does it have to embed in 2n+2m or are other signatures possible?

2025-12-05 13:20:21 GMT

- reply

ok i do know that the 3-sphere embeds in R^4 though (or in general, the n-sphere in R^(n+1)), i'm still trying to understand embeddings better.

2025-12-05 13:07:45 GMT

- reply

first time i ever met an ordinary mathematician. hello i see you are also in Edinburgh

2025-12-05 13:00:11 GMT

- reply

the logical basis is group theory. the quaternions are isomorphic to the group SU(2), the group of complex hermitian 2x2 matrices with determinant equal to 1. It turns out that every element of SO(3) can be mapped onto two elements of SU(2), which is how quaternions come to be useful for handling rotations. the "mixing up portions" isn't some arbitrary thing. it's what comes out of the maths. and you can have SU(3), SU(4), SU(anything).
SU(2) is also the same group as Spin(3) which pops directly out of the Clifford algebra. it's one of the "accidental isomorphisms".

2025-12-05 12:52:36 GMT

- reply

well yes, but a lot of maths proceeds without the motivation of working out the actual values of anything. it's all done in the abstract. and a whole lot of stuff goes on in there. anyway quaternions are very good for a lot of things of course even in computer graphics. multiplying two of them is very straightforward, for instance. but aside from any practical use, they are hugely important in abstract algebra, topology and geometry.

2025-12-05 12:48:52 GMT

- reply

precomputation? ehh.. mathematics doesn't exist purely for the sake of computer graphics programmers, believe it or not.

2025-12-05 12:43:52 GMT

- reply

oh quaternions are way less mushy than axis angle. axis angle is mushy. quaternions are solid, and very much more generalisable.

2025-12-05 12:42:29 GMT

- reply

right if you're using euler angles, which we're not. but regardless, any rotation can still be represented with at most 3 independent variables. the 4 variables in angle-axis representation are not independent if the axis is unit length (the axis itself is only 2 independent variables, in that case, up to sign, which can be transferred to the angle).
but if we are dealing with Lie groups, the actual rotations are obtained from infinitesimal generators through the exponential map.

2025-12-05 12:38:00 GMT

- reply

if it can be parametrised on the unit ball, surely it embeds in R^3.

2025-12-05 12:34:48 GMT

- reply

it's three. three axes of rotation. and it can be parametrised on the unit ball in R^3.
a rotation can also be represented by a unit quaternion, this is the lie algebra of SU(2) which is the double cover of SO(3). a quaternion has 4 independent variables in general. but for a unit quaternion only 3 are independent up to a sign flip (hence the double cover).

2025-12-05 11:33:24 GMT

- reply

SO(3) only has 3 free variables. its matrix representation may have 9 elements, but they are not independent. that would be GL(3).

2025-11-27 13:16:16 GMT

- reply

nicht lebe, oder nicht wohne?

2025-11-06 17:05:20 GMT

- reply

gotta side with ISO on this one. you can read the spine when the books are sat on a horizontal surface.

2025-09-24 11:19:19 GMT

- reply

1959, Minnesota: https://www.kyocera-avx.com/docs/techinfo/Connectors/The-History-of-Insulation-Displacement-Technology.pdf

2025-06-06 13:28:13 GMT

product idea: tape loader for PCs so you can load size-coded demos from cassette tape

2025-03-28 20:06:55 GMT

- reply

oh wait it's a door, i spent too long looking at that thinking it was a pool table lmao

2025-03-21 15:18:49 GMT

- reply

talking about bogosort levels of deliberately stupid

2025-03-21 15:18:04 GMT

- reply

oh well there's definitely some of it, but not *enough* of it

2025-03-21 15:16:42 GMT

- reply

it's tempting to just go and write a lot of code that doesn't work, or it works in theory but far too slow to be practical

2024-11-21 12:11:31 GMT

you know what, i don't even care how good AI is at coding. if the literal ghost of Donald fucking Knuth showed up at my desk and started typing my code for me, i'd tell him to fuck off back to his grave. It's *my* code and *I* want to do the coding.
[update] i have been informed that Donald Knuth is, in fact, still alive

2024-10-16 16:30:53 GMT

how is anyone ever able to afford to buy a car new? even a small hatchback seems to cost about a year's median salary