The h Bar

PM 2Ring

00:26

@RyderRude Stephen Wolfram proposed somehow combining a LLM with a more traditional symbolic AI like Wolfram | Alpha, back when ChatGPT. And he actually did it.

in Tavern on the Meta on Meta Stack Exchange Chat, 2 days ago, by PM 2Ring

@ꓢPArcheon See Wolfram|Alpha as the Way to Bring Computational Knowledge Superpowers to ChatGPT & https://gpt.wolfram.com/

in Tavern on the Meta on Meta Stack Exchange Chat, 2 days ago, by PM 2Ring

Disclaimer: I haven't used it myself. And I read recently that it's still of limited usefulness because even the latest incarnations of ChatGPT have difficulties in creating the appropriate commands for Wolfram | alpha. Which shouldn't be too surprising, since ChatGPT operates on the level of tokens, not concepts.

Actually integrating a LLM into a symbolic system is difficult because they operate in such different ways.

But we do know that you can't simply tweak a LLM to overcome its intrinsic flaws.

As I said over two years ago, "Perhaps some future AI program will be able to reliably answer SO questions, but ChatGPT certainly cannot do it. And even if you gave it ten times as much training data, it still couldn't do it, no matter how good that training data was". — PM 2Ring Feb 5 at 11:32

(cont) The GPT architecture is impressive at transforming language structures, but it has intrinsic flaws. It is not sufficient to create a reliable AI, no matter how large the model is, or how well it's trained. Reinforcement learning from human feedback can mask some of those flaws, but it can't repair them. — PM 2Ring Feb 5 at 11:32

Some writers discourage use of the term "hallucination" because it implies that it's an aberration caused by a mistake in the LLM's worldview. But a LLM doesn't have a worldview, and a so-called hallucination is simply the product of the LLM performing its usual next token prediction algorithm. See ChatGPT is Bullshit linked in meta.stackexchange.com/a/406431 by Jan Murphy. Briefly, a LLM just churns out plausible bullshit, oblivious to its truth value. Sometimes the BS is correct, sometimes it's not. — PM 2Ring yesterday

Interacting with ChatGPT resembles a conversation with a human, but it really isn't a conversation. ChatGPT has no workdview or Theory of Mind. It just has a "theory of language", which enables it to produce token sequences that appear to be appropriate.

Here's a simple analogy. If you hold a seashell to your ear it can sound like the ocean. But you aren't really hearing the ocean. You're just hearing sounds that seem similar to ocean sounds.

Oops. Wolfram proposed somehow combining a LLM with a more traditional symbolic AI like Wolfram | Alpha, back when ChatGPT first appeared.

Typing on Chat is hard when you can only see two lines, and your vision's getting blurry...

naturallyInconsistent

01:33

@TobiasFünke I wouldnt be as cavalier as you are. I have a colleague, right now, who continues to ask chatgpt for maths even though I made him confront chatgpt with a desk calculator in his hand. People like that are sitting in HR and making decisions over whether to hire mathematicians and physicists. Or deciding to decrease the number of students at universities.

And that person ostensibly had a technical background (in computer engineering).

Allie

hi

im sad and discouraged

naturallyInconsistent

dont be; hoomanity had been through worse

Allie

not even about that

just my own personal issues

i feel dumb and uncapable

naturallyInconsistent

that's not true. How many people can get to the level of understanding DFT?

naturallyInconsistent

05:42

@ACuriousMind finally finished this wonderful journalism. Nuance increased, but I think I still stand uncorrected. I simply do not see NA and EA LW extropians as separate enough cohorts to be infected by neoreactionaries separately. They are overlapping way too much to make such separation meaningful to meow. Including in their unsavouriness.

Ryder Rude

07:25

@PM2Ring nice answer

@PM2Ring exactly

so chatgpt works on the level of tokens and doesn't understand ideas

i was thinking that, if we make an AI that is good with language, another AI that is good with logic and truth another that is good with something else. and we assemble these together, can the assembled AI be called equivalent to a human?

it reminds me of the Chinese room experiment en.m.wikipedia.org/wiki/Chinese_room . when i first read about this experiment, i thought the answer was that "the room, as a whole, understands Chinese"

as in, the combined system "person + instruction set" understands Chinese

but now I think it is incorrect. to "understand" something means there must be some experiencer doing the understanding. but the "person + instruction set" system has no indivisible experience

in Chalmers' terms : a system is conscious if there is something it is like to be that system

Tobias Fünke

08:24

@Allie :( of course you are not. Though I would say the very most people in academia I spoke to have similar feelings from time to time...

Madder

ideas? chat.stackexchange.com/transcript/message/67186428#67186428

John Rennie

08:40

@Madder Hi :-)

The problem is that the terms matter and energy predated quantum field theory so their meanings are rather vaguely defined.

Ryder Rude

@Allie u r doing great

@Allie maybe also take a break

qwerty

@TobiasFünke you'd think it'd stop after leaving :')

Ryder Rude

@JohnRennie matter specifically seems to be ill defined. maybe it can be used to refer to particles in QFT

even the materialists these days prefer to say "everything is information" instead of "everything is matter"

qwerty

@Madder saying something like spacetime is a "thing" that "makes up" the universe is called reification en.wikipedia.org/wiki/…

Therefore I would not word it like that. But you can state something a bit more carefully, like Ellis does in the intro of his book "relativistic cosmology"

naturallyInconsistent

09:28

@Madder the premise is wrong and thus it is not sensible to chase the logical inference

Ryder Rude

09:52

this is such a bizarre opinion from Feynman physics.stackexchange.com/q/840939/156987

skullpatrol

10:12

Surely, he must've been joking.

naturallyInconsistent

10:29

Have you not read ACM's comments on that post? Feynman's opinion is not at all bizarre. It is somewhat trivial.

And it had already been shown true in some applications

not least because, again, there are trivial cases

Ryder Rude

it seems to be a different idea from simulating classical mech via classical computers

which can always be done

to realise Feynman's idea, one would have to store the wavefunction of an arbitrary quantum system into the quantum state of $n$ qubits. how would anyone do this

after storing it, one would need to "implement" the Hamiltonian of the system one wants to simulate

it is a bizarre idea

there is a key difference from classical mechanics. when a classical computer simulates classical physics (say, gravity), it does not actually use the force of gravity to imitate gravity. it just uses instructions

while Feynman's idea seems to be about using actual laws of physics (like a qubit evolution) to "simulate" other laws of physcs

it is bizarre

naturallyInconsistent

10:52

It is exceedingly obvious that you have not read any of the literature on quantum computing. There is nothing bizarre about the issues you have raised.

Ryder Rude

11:02

i have read that quantum computing is good for reducing the complexity of some computations. it is not good for Feynman's idea

one can maybe simulate Bell's inequality on such a computer. but then u r doing exactly the same experiment and calling it a simulation

skullpatrol

Full disclosure: my comment was referring to the OP.

Sarcastically mimicking Feynman's book :P

Ryder Rude

i thought u meant it was Feynman who was joking :P

OP says "speed-ups in quantum many body problems r expected". OP should give a reference for this claim

skullpatrol

Yeah, thus the clarification was necessary.

Sorry about that.

qwerty

11:36

TIL that mica is technically "edible" o.o

Tobias Fünke

RR what do you find bizarre? There is active research in "quantum simulation"

Ryder Rude

@TobiasFünke i think it is ridiculous to expect this can work...

Tobias Fünke

it works

Ryder Rude

maybe for some limited systems, it does

like simulating qubits using qubits

Tobias Fünke

optical lattices etc.

of course it is non-trivial

but I don't see what is "bizarre" here (or say more bizarre than QM itself)

Ryder Rude

11:45

to simulate one quantum system using qubits, one would need an identification between their Hilbert spaces and observables. and on top of that, one would need a time evolution correspondence

so it should be extremely restricted. contrast this with classical mech simulations using computers where there r no restrictions

cuz we r not using physics to simulate physics when we simulate classical mech using classical computers

Tobias Fünke

I think you are too narrow minded here, with all due respect

as I said, it is ongoing research, and many things have been achieved already

Ryder Rude

can u give me an example? it might be that they have achieved a lot relative to the restrictions

i don't think they can simulate e.g. standard model using qubits

that should be bizarre

@TobiasFünke i will google it

Tobias Fünke

nature.com/articles/s41586-022-04940-6

I don't say that you should expect to do arbitrarily complex things. But it is not like "you can do barely anything with it" either.

Ryder Rude

it seems they are simulating quantum lattices using quantum lattices. so the restriction i gave does apply

i mean the optical lattice

they say that the optical lattice has highly adjustable parameters

Tobias Fünke

yes

Ryder Rude

11:54

im not sure what this means. but it might mean that they can implement relatively arbitrary Hamiltonians

it seems like at least the lattice versions of the Standard model might also be simulatable

@TobiasFünke i didn't expect flexibility

like, the ability to implement relatively arbitrary Hamiltonians

Tobias Fünke

nature.com/articles/s41586-019-1614-4

yes in optical lattices you have a high degree of control for the potential and hopping terms

these are well-studied and versatile systems

Ryder Rude

wow

it seems pretty good. i didn't expect flexibility

thanks for the links

Tobias Fünke

^^

welcome

Ryder Rude

really cool that they are simulating electrons using atoms.

Tobias Fünke

yes, it is very interesting and clever

Ryder Rude

12:07

i think the reason I didn't expect flexibility was because I thought they would use electron qubits for the simulation

not much flexibility for the Hamiltonian of that

but atom Hamiltonians should have more parameters

PM 2Ring

12:53

@RyderRude Maybe this will help:

May 15, 2024 at 12:54, by PM 2Ring

@Relativisticcucumber Because a qubit captures the "central mystery" of quantum mechanics. From The One Mystery of Quantum Mechanics

May 15, 2024 at 12:54, by PM 2Ring

> We choose to examine a phenomenon which is impossible, absolutely impossible, to explain in any classical way, and which has in it the heart of quantum mechanics. In reality, it contains the only mystery. We cannot make the mystery go away by “explaining” how it works. We will just tell you how it works. In telling you how it works we will have told you about the basic peculiarities of all quantum mechanics.

May 15, 2024 at 12:57, by PM 2Ring

> I will take just this one experiment, which has been designed to contain all of the mystery of quantum mechanics, to put you up against the paradoxes and mysteries and peculiarities of nature one hundred per cent. Any other situation in quantum mechanics, it turns out, can always be explained by saying, 'You remember the case of the experiment with the two holes? It's the same thing'.

Also see

Bloch sphere

In quantum mechanics and computing, the Bloch sphere is a geometrical representation of the pure state space of a two-level quantum mechanical system (qubit), named after the physicist Felix Bloch. Mathematically each quantum mechanical system is associated with a separable complex Hilbert space H {\displaystyle H} . A pure state of a quantum system is represented by a non-zero vector ψ {\displaystyle \psi } in H {\displaystyle H} . As the vectors...

@RyderRude But we don't know how to assemble such radically things together. And if you don't understand what you're doing, and how those components actually work, you may create a Frankenstein-style monster.

Consider: it can be difficult for humans who are experts in literature & language, etc, to communicate effectively with humans who are experts in maths and science. And these are people who do (presumably) know what they're talking about, and who have sophisticated fully-developed theories of mind.

"we don't know how to assemble such radically different things together"

Ryder Rude

@PM2Ring the qubit does contain most of the fundamental mystery of QM, but I think it is still not sufficient to imitate all other quantum systems cuz the others may be infinite dimensional Hilbert spaces

@PM2Ring yes. the assembled entity is not a single entity even behaviorally

but in cases like the Chinese room experiment, it may behave like a single entity, but it does not have a unified experience

this means it is not a single entity in the experiential sense

if we just put things together in a box, it is not a single entity. for it to 'understand' things, the entity needs to be interconnected in an intricate way @PM2Ring

Integrated information theory says that the more the behavior of a system can decomposed into the behavior of its parts, the less "integrated information" the system has

ACuriousMind

14:01

@qwerty I mean in the end this is just nitpicking over whether we conceive of choosing coordinates as a single-step process where we fix a chart $U\to\mathbb{R}^n$ or as a two-step process where we fix a chart $U\to V$ with $V$ some abstract vector space of dimension n and then fixing a basis on $V$, i.e. a linear isomorphism $V\cong \mathbb{R}^n$ (which you can phrase in terms of picking a basis of $V$ or a basis of $V^\ast$ or whatever, it's linear algebra, it's all equivalent).

I personally don't see any advantage in splitting this into two steps.

Ryder Rude

14:50

a word which doesn't describe itself is heterological. is heterological a heterological word?

if u assume that this is heterological, then it isn't heterological. but if u assume it is not heterological, then it is heterological

User198

15:08

In the Hamilton-Jacobi equation $-\frac{\partial S}{\partial t} = H$, there is also a Hamiltonian governing time evolution.

That is a recurring theme just like: $\frac{df}{dt}={f,H}$ or the Schrodinger equation, that's all ok.
But what kind of a function (mathematical object?) is the $S$ actually. Hamilton's principal function, a generating function ok, but

Does the $-$ sign in the HJ equation indicate something similar that is going on with the Liouville equation.
There is the thing that:
$\frac{d\rho}{dt}={\rho,H}$ and not $\frac{df}{dt}={f,H}$ because states evolve with $-$ in referenc…

Would than $S$ be something more simmilar to an observable or a state?

Or neither?

Silly Goose

15:55

LSZ formula can be used to compute expectation values, right?

ACuriousMind

@User198 It's not really clear what you're asking. If you look at the definition, Hamilton's principal function is a function of the initial and final positions $q_0,q$ and the initial and final times $t_0,t$. What about the mathematical nature of this function is unclear?

@SillyGoose This question sounds as if you really want to know something else :P Let's say the answer is "Yes", what's the follow-up question?

Silly Goose

if fermi liquid theory can just be derived using (essentially) LSZ formula (in spirit)

because the cartoon of fermi liquid theory invokes an adibatic turning on of the interaction in order to study the interacting theory at the end of this adiabatic evolution. but i feel that this adiabatic turning on of an interaction can instead be framed as a (more general) problem of computing interacting field theory quantities in terms of the free field quantities, which is what LSZ formula generically does.

and just like how thinking of the derivation of LSZ formula in terms of turning on an interaction is not necessarily the "right" way to think of it (though it is physical), perhaps it is also not the "right" way to think of Fermi liquid theory in terms of this adiabatic turning on of an interaction

ACuriousMind

The LSZ formula is not intrinsically married to any "turning on" of the interaction (but the adiabatic switching is involved in some derivations)

I'm afraid I don't know anything about Fermi liquid theory beyond its existence

Silly Goose

well then let me pose the question this way

i want to study an interacting Fermi gas, interaction is generic for now

i want to compute thermodynamic quantities of this interacting Fermi gas in terms of the quantities of the free Fermi gas. The free Fermi gas system is a many-body fermionic system. That is, we mathematically can describe it using well understood c/a operators $c_{\vec{k}\sigma}$ and free Hamiltonian $H_0$

okay so it is a "QFT" that is not Lorentz invariant nor relativistic, but is rotationally and translationally invariant (at least).

Now, I want to study some $H := H_0 + \lambda V$.

In principle, can I just do what is done (sans maintaining Lorentz covariance/invariance) to derive the LSZ formula to get an analogous theory of the "interaction QFT" out of the "free QFT"?

ACuriousMind

16:14

I've never seen anyone use "LSZ" in non-relativistic contexts

Silly Goose

er well forget i said that then. i just mean: can I (as is done in QFT) construct interacting theory quantities in terms of the free quantities? for instance, i can turn interacting ground state expectation values into expectation values of the free ground state (using Wick's theorem and blah blah blah).

ACuriousMind

but, well, in principle the same kind of derivation should be possible, i.e. relating the S-matrix elements to the vacuum expectation values of field operators times the propagators

@SillyGoose but that's not what LSZ does - it does not relate the interacting VEVs and the free VEVs. LSZ relates the interacting VEVs and the S-matrix.

Silly Goose

Maybe a precise way to state my question is: The general computational machinery of QFT has nothing to do with what irreducible representations the fields transform under. In other words, the abstract computational machinery of QFT is representation independent.

It is perhaps algebra dependent, in that we need to know we are ultimately dealing with some fermionic or bosonic fields.

ACuriousMind

the relationship between the "interacting VEVs" and the "free VEVs" you're talking about is the interaction picture derivation we usually do after LSZ to get the expression for the interacting VEVs in terms of Feynman diagrams

Silly Goose

@ACuriousMind oh okay yes i did not clearly understand what precisely LSZ formula refers to

i do mean to talk about relating IVEVs to FVEVs (free VEVs)

ACuriousMind

16:22

i.e. what you mean is the Gell-Mann-Low theorem and variants thereof

This is actually a generic theorem in QM (but technically not in QFT due to Haag's theorem-adjacent issues)

Silly Goose

ah okay

ACuriousMind

so certainly you can apply it to whatever free/interacting system you have regardless of whether it's relativistic QFT or not

Silly Goose

@ACuriousMind hm by QFT here do you mean involving infinite CCR/CARs?

ACuriousMind

anything to which Haag's theorem or specifically the non-existence of the interaction picture applies

Silly Goose

oh that is not equivalent to involving infinite CCR/CARs?

ACuriousMind

16:26

Not necessarily, since you do not need to even talk about those operators to state (some versions of) Haag's theorem. It's correct that Haag originally showed it in the context of inifnitely many c/a operators

Silly Goose

oh

ACuriousMind

but e.g. the Wightman axiomatization of QFT in which Haag's theorem is a proper theorem is merely in terms of the field operators, not the c/a operators

of course, for a free field you may construct the infinitely many c/a operators as usual but there is somewhat significant debate over what the "real" issue is that produces Haag's theorem, since one does not need (at least not in an obvious way) to invoke the CCR between the field operator and its conjugate to prove it

Silly Goose

i see

also why does the wiki for GML talk about relying on the Dyson series as being a bad thing?

ACuriousMind

I don't think it does

it just points out there are recent proofs not relying on the Dyson expansion; it's perfectly normal for mathematicians to be interested in alternative proofs of a result, this does not imply the previous proofs are "bad"

Silly Goose

"As in the original paper, the theorem is typically proved making use of Dyson's expansion of the evolution operator. Its validity however extends beyond the scope of perturbation theory as has been demonstrated by Molinari."

I guess the verbiage here suggests to me that for some reason it is more general to not rely on perturbation theory and rather rely on adiabaticity

ACuriousMind

16:34

@SillyGoose That's a much more specific statement than the Dyson series being "bad" and to me the meaning of the sentence is perfectly clear: If you use the Dyson series, you need to prove convergence, i.e. put some kind of assumptions on your Hamiltonians so that the series converges. If you can avoid the series, you can avoid those "perturbative" assumptions.

Silly Goose

hmm but adiabaticity is also a strong assumption, or am I mistaken in thinking this

ACuriousMind

The article does not say that Molinari's proof relies on adiabaticity

but in any case it's certainly a different assumption than the assumption of convergence

Silly Goose

oh no i have conflated two alternate proofs mentioned

okay i see

ACuriousMind

I don't know why you think this is important given that you want to use it in a QFT context where the interaction picture doesn't exist anyway if we want to be rigorous :P

Silly Goose

insert professionals have standards meme

Well the reliance in adiabatic theorem in fermi liquid theory doesn’t make sense to me is why i am trying to understand it an alternate way

User198

16:52

@ACuriousMind Hm. Maybe I am more confused about the physical interpretation of $S$ than. According to the HJ, the Hamiltonian governs the time evolution of $S$.

But if it is not an observable or a state, I don't know what is than evolving in time?

ACuriousMind

$S$ is a function of time $t$. So it has a time derivative and hence some differential equation it obeys

there is no direct physical interpretation of Hamilton's principal function

bolbteppa

I was asking SkynetGPT about how the metric in LQG, and it just kept repeating $g_{ij}(x) = \frac{1}{2} \epsilon_{abc} E^a_i(x) E^b_j(x)$ as the definition so many times, just crazy

Silly Goose

I was asking my physics professor about the action and it just kept repeating $S = \int dt L$, just crazy ;)

User198

@ACuriousMind Alright. Thank you.

bolbteppa

@User198 For a function $f = f(p,q,t)$, you have this $\frac{df}{dt} = \{f,H\} + \frac{\partial f}{\partial t}$ formula, but $S = S(q)$ is just a function of $q$ and $t$, which you can see from $S = \int dS = \int p dq - H dt$

User198

16:59

@ACuriousMind And $S=S(q,t)$ . It is not a function of $\dot q$ or $p$ like the lagrangian and hemiltonian?

Hamiltonian*

@bolbteppa Ah ok. I see

Thank you

ACuriousMind

@User198 Indeed, the principal function is not a function of velocities or momenta.

bolbteppa

@SillyGoose At least that is consistent and doesn't have extra "c"'s floating around!

Tobias Fünke

17:21

Goose: Take a look in any (modern) condensed matter QFT/many-body book

of course they use Gell-Mann & Low, Wick and so on

Ryder Rude

nvm

Tobias Fünke

you in principle even compute the same (qualitatively) data (e.g. cross-sections, spectral densities and so on)

Silly Goose

17:41

hm maybe i'm too stuck in the past

Tobias Fünke

the future is now

Silly Goose

there is always the nag that one is using too fancy machinery for the problem at hand

Signor Feynman

18:29

Miao

bolbteppa

1

A: Is there an analogue of the LSZ reduction formula in quantum mechanics?

By Huygen's principle we can write the wave function at some later time $t_f > t_i$ at a point $\mathbf{x}_f$ in terms of the wave function at some initial time $t_i$ at a point $\mathbf{x}_i$ as $$\theta(t_f -t_i)\Psi(\mathbf{x}_f,t_f) = i \int d^3 \mathbf{x}_i G(\mathbf{x}_f,t_f;\mathbf{x}_i,t_...

The machinery is always there even in simple cases

Tobias Fünke

18:56

@SignorFeynman Do you want to report? :d

Slereah

Even works for integrals

Silly Goose

19:11

how do you formalize "degeneracy" for a gapless, continuous spectrum?

say the free particle

Tobias Fünke

19:31

good question. but why did you add the "gapless" condition?

Signor Feynman

@TobiasFünke I'm actually a little dismayed as it ruined my perfect score streak

But even more dismayed because I was contradicted after saying something that was both on the slides of the course and on FLP III, 21 (on which almost every introduction to Josephson effect is based on)

>When they are included, $ρ_1$
and $ρ_2$ do not in fact change, but the current across the junction is still given by Eq. (21.44).

Which cost me the charge of "not having a solid understanding of Josephson effect"

I don't think I was outstanding but I'm rather annoyed by someone punishing me out of self-contradiction

Tobias Fünke

19:57

@SignorFeynman oh no :(

FLP?

@SignorFeynman I see. :( sorry to hear that. Such experiences are ... indeed

Signor Feynman

@TobiasFünke Feynman Lectures on Physics

Tobias Fünke

ah

I hope you can forget this incident soon an move on :) I know it is not easy, though...

Signor Feynman

The very last lecture was a seminar about superconductivity including a brief simple account of the Josephson effect

Which inspired many other introductory explanations

Tobias Fünke

did you mentioned that you've found the statement there?

Signor Feynman

I mentioned that and the slides but it was no use

Tobias Fünke

20:03

... damn

Signor Feynman

As my thesis advisor says "sometimes you must be sympathetic" :P

Subtle and savage as always

My gauge era now begins

Secret

Good news for faster-than-light travel. Bad news for time travel.

►

Tobias Fünke

@SignorFeynman nooooooooo

@SignorFeynman lol

imbAF

When calculating the differential decay rate of the muon decay to electron, muon neutrino and electron anti neutrion, we do it w.r.t electron energy. And we say we can't do it w.r.t to the neutrino energy. Is that because we don't have detectors that can detect neutrinos in LHC?

qwerty

21:08

@ACuriousMind yeah that's correct. I'm still in the process of untangling my thoughts and rewriting some of my notes though. I guess personally, I was just trying to really understand how the co-ordinate system "naturally induces" a basis for each tangent space, and it felt like this was hinting that it's because we implicitly make a choice earlier on. but I need to revise my understanding of some things maybe

qwerty

21:19

(@Tobias I didn't forget you were asking but I was still editing my notes...)

@SignorFeynman so was your examiner incorrect? :(

Tobias Fünke

Hi

no problem, from what ACM said I understood

Signor Feynman

@qwerty well, she said that the passage from FLP quoted above (and elsewhere in the literature and in her slides too) is wrong, but anyways

Tobias Fünke

and my comment remains: I've never encountered this way before, and from my naive perspective/understanding it does not make anything easier

Signor Feynman

I'm over it. I'll start transitioning back to gauge theory tomorrow

If everything goes according to the plan, I should graduate by July

Tobias Fünke

@qwerty but you understand that a choice of coordinates induces a set of partial derivatives, roughly speaking, which constitute the basis of the tangent space? is there anything "more natural" than this?

@SignorFeynman wow, nice!

qwerty

21:34

@TobiasFünke my point wasn't that it makes anything "easier" though :/

Tobias Fünke

oh

Signor Feynman

@TobiasFünke and then I'll be free and poor

qwerty

@TobiasFünke yes. it's not "more natural". I'm trying to poke at the "why" this "naturally induces" happens

Tobias Fünke

yes haha

@qwerty hmmmh

Signor Feynman

Oh man, I wish I could take a sabbatical to consider PhD

Tobias Fünke

21:36

@SignorFeynman just get rich

and then you can take as many sabbaticals as you want. it is sooo easy :)

qwerty

@SignorFeynman 🍀🎉

Signor Feynman

But I already "wasted" this year (rushing I could have completed by last September and start a PhD)

Tobias Fünke

In Germany we say: Gut Ding will Weile haben.

Signor Feynman

@TobiasFünke oh man, you should link your online scam course based on nested Ponzi schemes after saying that

Tobias Fünke

@SignorFeynman hehehe that would be a great move now :d unfortunately I am not a high performer myself and I don't have any course to offer

perhaps at some later point in my life lol

Signor Feynman

21:39

Condensed matter is already a Ponzi scheme :P

Tobias Fünke

@qwerty if you want you can explain what exactly you mean to me, and then I can think about it (I don't know if I can help you, but I could try). Of course, only if you want to

lol

Signor Feynman

But those guys are the richest in my department, they always set up conferences with food

Tobias Fünke

haha

Signor Feynman

Which caused the joke: Physics and Astronomy---->Physics and Gastronomy

2

Tobias Fünke

well, IIRC, condensed matter is by far the largest field/community in physics

Signor Feynman

21:41

@TobiasFünke the bulk of physics

Tobias Fünke

and many subfields have direct applications and are relevant for industry...hence more funding (?)

qwerty

@TobiasFünke no it's okay haha. but thanks. I'm in the process of writing.

Tobias Fünke

OK, qwerty.

@SignorFeynman hehe

Signor Feynman

Qwerty, have you changed your opinion about cats?

Would you let a cat enslave you and be its rightful human? :P

qwerty

hahaha how come xD

not really :p

Tobias Fünke

21:43

oO

I gotta go to sleep

See you, folks. Have a nice evening/day :)

qwerty

21:55

@SignorFeynman instead of studying quark particles, you study quark cheese

Signor Feynman

@qwerty I'll crack this joke forever now. THANKS

@qwerty disappawinting

ACuriousMind

22:14

@qwerty Think about the definition of tangent vectors: Regardless of which definition you choose, they are all essentially trying to "transfer" the notion of directional derivatives from the coordinate chart to the manifold. And a choice of coordinates gives you a choice of directional derivatives $\partial_\mu$.

ebenezer

hi all; here I am with some gentle spam from the site, has anyone any idea about my question here physics.stackexchange.com/questions/842243/… thanks in advance and sorry to bother!

Transcript for