Albert Einstein's theory of relativity in words of four letters or less (1999) | Hacker News Reader

Albert Einstein's theory of relativity in words of four letters or less (1999)

muppetlabs.com

240 pointssigna11130 comments

Reminds me of Yasha Berchenko-Kogan’s excellent answer to the question “What do grad students in math do all day?”

https://www.quora.com/Mathematics/What-do-grad-students-in-m...

> a bit like trying to explain a vacuum cleaner to someone who has never seen one, except you're only allowed to use words that are four letters long or shorter.

> What can you say?

> "It is a tool that does suck up dust to make what you walk on in a home tidy."

Somehow the sequences of small words and ample syntax make this sentence quite difficult to parse.

Maybe just go full pidgin:

“Tool to suck dust, make tidy for walk in home.”

Oh come on, this shit is easy. Why did they say "it is" and not "it's", by the way? To put it that way can't help. So yeah, it's a pipe that can suck, and you push it all over your room, to suck the dust and dirt up off the rugs and such, and in fact off of any low down flat part. One kind can even move on its own! But what I want to say here, in the main, is that you math guys have all lost your grip on how to say any idea in an easy form. You are not able to do it any more, 'cos too much math has made you sick in the head.

Nice one but "'cos" does not go by the rule, I feel. You can use "for" at the same spot but, well, it has a tone you did not go for in your text.

One can also use "as" in that spot, no?

  s/, 'cos/;/

I feel like there's still room to avoid pidgin while making it less awkward, e.g.: "It's a tool that can suck up dust or dirt to make your home more tidy."

This version reminds me of Poetry for Neatherthals (board game).

You have to get others to guess a (typically multi-syllabic) word or phrase, but use only one syllable words to get them there.

"Tool suck dust, make not dirt for walk in home."

You don't need the awkward "does". I'd go with:

It is a tool to suck up dust and dirt from rugs, wood or even tile.

I liked that the original explained the value of the vacuum cleaner. It's not that it removes dirt and dust per se, it's that it makes spaces you walk on tidier.

A tool to take away dust and dirt in the home.

It suck shit up

As someone who's encountered the "pet-shit-on-the-floor" problem, I can assure you that this description is inaccurate.

Why use many word when few word do trick

job*

This essay is fantastic at demonstrating that putting a word length limit actually makes explaining things more complicated. I got lost at around chapter 5 because the author couldn't use words like "gravity" and "acceleration" and I got confused by which one is "new pull" and which one is "old pull". It's too bad, as it was interesting up to that point.

Of course you find it hard to distinguish the two! You don't have equipment for measuring tidal forces, and they are locally indistinguishable.

Of sure you find it hard to tell the two away! You lack the gear for tide pull test, and they feel the same here and local.

I hate this.

The new pull and the old pull both just feel like a pull, if you can only feel the pull at one spot. To see how the old pull is not like the new pull, you have to test the pull at a spot near you (but not the same spot), too. The new pull will be the same at each spot, but the old pull may not be the same (we call this the tide), and you test the sum of the new and old pull.

(This is hard.)

It's fair that it's hard to keep the two from becoming the same in your head, you need fancy stuff to test for the force of the tide, and they are more or less the same from a close-up (any which is much closer than, say, the moon) view!

(Verbosity is your friend)

> It's too bad

I think that's the whole point. It was never meant as being easier to grok

What was it meant as?

A statement on the value of vocabulary, perhaps.

There’s a reason why vocabulary exists. It isn’t to make things harder to understand. Sometimes the best way to explain something to someone with a limited vocabulary is to expand their vocabulary in the process.

To expand the vocabulary, you'd still need to break down 'gravity' and 'acceleration' into simpler words. Though still easier without an arbitrary length limit. You could say, "change in the speed a thing moves" but that needs five or six letters.

It's an exercise. I would have much preferred using the 20k most common words or something like that. The first thing that came to mind is "elevator" which is where the equivalence eureka comes from. It can be done in British English as "lift" but difficult otherwise.

Elevators are cool like telephone booths. I've wondered what a dog thinks using them for the first time, then accepting what they do and how much they understand its geometries.

Reminds me of Guy Steele making the point about big languages and small ones in his talk about Scheme. Started the whole lecture using only one-syllable words then gradually defined two-syllable words using only single syllables and so on.

Ah, I take it that's "Growing a Language"?

There was a talk at a university, where the presenter used only words of two or less SYLABLES , but he allowed himself to use more complicated words after explaining them (but kept that to a minimum).

I can't find either the author or the talk. I think it was some 5 years ago.

At first, I thought it was Randall Munroe, but I might be remembering this: https://xkcd.com/thing-explainer/

I've also tried with Paul Graham, who has some articles trying to convey something similar, but no luck there.

Edited to add : I think the original proponent of a similar idea was Richard Feynman : https://www.hpcdan.org/reeds_ruminations/2022/03/understandi...

Reminds me of "Gadsby", a 50.000 word novel without the letter "e":

https://en.m.wikipedia.org/wiki/Gadsby_(novel)

I imagine LLMs would excel in this kind of writing these days.

But really impressive for the time.

They’d probably sucks at a challenge like that because they work on tokens and don’t really see individual letters.

There was a post here a little while back asking AI models to count the number of Rs in the word raspberry and most failed.

I think it's the exact opposite, as they operate on a token-level, not a character level, which makes tasks like these harder for them. So they would generate a sentence with multiple es in it and just proclaim that they didn't.

(Just tried it, "write a short story of 12 sentences without one occurence of the letter e" - it had 5 es.)

You're assuming all you can do is prompt it. Surely you could also constrain its output to tokens that genuinely contain no e’s (or make only max 4 letters per word). LLMs actually output a probability distribution of next tokens; ChatGPT just always picks the top one, but you could totally just always filter that list by any constraint you want.

I was going to point that out.

What I will add is that constrained generation is supported by the major inference engine like llama.cpp, vllm and the likes, so what you are describing is actually trivial on locally hosted models, you just have to provide a regex that prevent them to use the letter 'e' in the output.

You can do this more properly with the antislop sampler and we are working on a follow up paper to our previous work on this exact problem.

https://github.com/sam-paech/antislop-sampler

https://arxiv.org/abs/2306.15926

But the problem is that the tokens are subwords, which means that if you simply disallowed tokens with es, you'd make it hard to complete a word given a prefix.

For example, it may start like this "This is a way to solv-", or "This is th-"

If I understand it correctly, that's a valid concern but the way structured generation library like outlines[1] work is that they can generate multiple variants of the inference (which they call beam search).

One beam could be "This is a way to solv-". With no obvious "good" next token. Another beam could be "This way is solv-". With "ing" as the obvious next token.

It will select the best beam for the output.

[1]:https://github.com/dottxt-ai/outlines

... What if you retrained it from scratch, on an e-less corpus?

Yes, that would probably work quite well, given enough training data. However, I interpreted the question/claim as a task that LLMs excell at, meaning that writing text while avoiding a certain character is a task for a general purpose LLM.

I tried something like that some time ago. The problem with that strategy is the lack of backtracking.

Let's say I prompt my LLM to exclusively use the letters 'aefghilmnoprst' and the LLM generates "that's one small step for a man, one giant leap for man-"[1]. Since the next token with the highest probability ("-kind") isn't allowed, it may very well be that the next appropriate word is something really generic or, if your grammar is really restrictive, straight up nonsense because nothing fits. And then there's pathological stuff like "... one giant leap for man, one small step for a man, one giant leap for man- ...".

[1] Toy example - I'm sure these specific rules are not super restrictive and "management" is right there.

The next token is obviously "goes". (Any language model that disagrees is simply wrong.)

I'm not sure if my chain's bein' yanked right now, but surely you mean "gos"‽

The plural of mangoe is mangoes. https://en.wiktionary.org/wiki/mangoe

All the training data contains 'e's.

That is not a counter point! The output has a probability distribution so you can assing zero to any e-containing token and scale everything else up accordingly.

I think an LLM would do well on this if you gave it a function that located words with an e so it could change them.

You don't need to go all the way to LLMs when a simpler approach may do.

Here's a "What if?" on a very similar issue that uses Markov chains: https://what-if.xkcd.com/75/

I wrote the relevant paper about this:

https://arxiv.org/abs/2306.15926

https://github.com/Hellisotherpeople/Constrained-Text-Genera...

LLMs are usually shit at this kind of wordplay, they don’t understand the rules - words that begin or end or include particular letters, words that rhyme, words with particular numbers is syllables - they’ll get it right more often than wrong, maybe, but in my experience they just aren’t capable catching wrong answers before returning them to the reader, even if they’re told to check their work.

I'd be curious to know if it was easier or harder (or perhaps just as difficult) to write than the French equivalent. [0]

The Wikipedia article goes on to discuss interesting aspects of how the book was translated in different languages, with different self-imposed constraints.

[0] https://en.wikipedia.org/wiki/A_Void

I can’t say for certain, but I’d guess that writing without the letter “e” is slightly more difficult in French than in English. For one, “e” is a bit more common in French (around 15% of all letters, versus about 12% in English). But more importantly, French grammar adds extra challenges—like gender agreement, where feminine forms often require an “e”, and the frequent use of articles like le and les, which become unusable.

That said, I think the most impressive achievement is the English translation of the French novel. Writing an original constrained novel is hard enough, but translating one means you can’t just steer the story wherever you like. You have to preserve the plot, tone, and themes of the original, all while respecting a completely different set of linguistic limitations. That’s a remarkable balancing act.

Georges Perec did the same with his novel "La Disparition".

What is almost as impressive is that these novels (at least Perec's) have been translated to other languages.

8 of them on the cover!

It was interesting to notice that not all short words are necessarily simple. Words like "void", "iota", "mass", or "veer".

Thanks to Javascript, I know void.

Thanks to Go, I know iota.

The question is: Do you know the words or do you know the meaning in the context of the article?

They demonstrate they know 'veer' (off topic) without saying 'veer'.

Now I feel old. I know void from K&R C and iota from APL

Thanks to Java, I know pain.

This is kinda confusing at it's more for people who already know the meaning. Take the bus exemple, it's so short that it skip explaining why someone on a moving bus will see different timing for the asteroid landing. You can decipher it if you know it, you'll not gonna learn from the story line.

"It's tempting, if the only tool you have is a hammer, to treat everything as if it were a nail." Maslow 1966. The essay is about physics, but all comments are about formatting and LLMs.

The conversationalist tone of the essay is misleading too. Hilbert, Minkowski, & Poincare, had done all the heavy lifting math and had held Einstein's hand all through 1915. As mathematicians they wouldn't qualify for Noble prize so made no claim to the discovery of GR.

I'm not sure if this is physically accurate, but the best description I've encountered for relativity is:

You are always traveling at the same speed. That speed is 'c', the speed of light.

If you are sitting still, you are 'falling' through the time dimension at 'c'. If you move in the X,Y,Z dimensions, you must move slower in the 't' dimension so that your velocity vector still sums to 'c'.

An immediate follow-up is: why do we always travel at c?

Maybe it's just a cosmic constant, or a result of other parameters and how they interact (a second order constant). Why does the electron have the mass that it does? Why do protons have the charge that they do? Etc.

For reference, Poul Anderson's 'Uncleftish Beholding' -- an essay on atomic theory written in modernized anglo-saxon.

https://en.wikipedia.org/wiki/Uncleftish_Beholding

Up Goer Five; rocket science explained using only the one thousand most common english words.

https://www.explainxkcd.com/wiki/index.php/1133:_Up_Goer_Fiv...

https://www.explainxkcd.com/wiki/index.php/Thing_Explainer

I love “Uncleftish Beholding”, which someone said is written in “Anders-Saxon”. I think it would be fun to do it live as a Power-Point presentation.

I was thinking this morning how weird it is that everyone knows who Einstein was

But much smaller percent Niels Bohr

https://en.wikipedia.org/wiki/Bohr%E2%80%93Einstein_debates

I appreciate this, though the hard rule seems to be doing more harm than good. For example, one 5-letter word became 6 words, because 5-letter words aren't allowed!

So while the vocabulary is kept low, the writing style becomes harder to process, at least for me. I wonder if there's a way to win on both fronts, to make it maximally comprehensible for all involved.

I'd argue "use normal words that everyone knows" (even if they are 5 letters!) would be included in such a strategy.

Edit: Okay now I made it further in and I'm being asked to keep several different perspectives in my head simultaneously, perceiving different events at different rates of time... I think I need a diagram... or a microdose...

Several variants of simplified English have been designed for the purpose of being understood by learners or people with only basic command of English as a foreign language. Wikipedia has a version in Simple English for instance: https://simple.wikipedia.org/wiki/Simple_English_Wikipedia.

OK, I'll give it a go...

"Any glow from an item will move in a void at the same rate. Each item has mass. Mass is the same as a glow. Mass will bend the area near it. Each item will move on a bent path in that area. If you move at a fast rate, time will seem to slow. An item will feel a pull down if the item goes up and does not stop."

That's way more work than it seems! Not being able to use -s or -er words is a real problem.

The explanation still kinda sucks. I like this one:

The easiest way to understand the relationship between time and space is repeat the thought experiment with the void, but assume that there is no consciousness there (i.e nothing running that can sense time passing).

Now imagine the only action you can take is to fire particles (say photons) in a given direction. In a void, that action is meaningless - the particle fires and never comes back. No information exists.

Now imagine there is a mirror somewhere in space. A particle fires, and then comes back. And maybe interacts with another particle. But still, this is generally meaningless and you cant derive any measurable thing from it, but you have a piece of information - particle comes back.

Imagine there are 2 mirrors in different directions. What you do is you set up 2 identical devices. Each one fires a particle, and when the particle comes back, it triggers a certain color ball to fall down a common shared tube, and then the particle gets fired again.

So with 2 mirrors, you get a sequence in the tube that looks something like blue, blue, blue, green, blue, blue, blue, green. Now you can make a measure of distance. You take the "blue" mirror as your unit, and say green mirror is 2 away.

You have also in fact created a clock. The tube contains information on how many cycles have passed - i.e in order to say that mirror is x away, you need to have counted x blue balls before that respective ball shows up. So you can see how distance and time is intimately intertwined. To measure distance, you have to necessarily have something that measures time.

Now lets say that the "green" mirror starts moving away from you, at a slow speed (i.e your particles are much faster. You start to see 3 balls in sequence, then 4, then 5, and so on. By comparing the difference in the subsequent position of the green balls, you can measure speed.

What happens if the speed of the mirror is 99% of the particle speed? The particle takes its sweet time getting there, and sweet time coming back. Even if you fire the particle as the green mirror is close to the particle emitter, its going to result in a measurement of a very large distance.

This is the relativistic effect where the space behind something moving fast increases.

This whole experiment demonstrates that what we consider space is precisely defined by measurements, and relativistic effects alter these measurements, which alters our perception of space.

You can do similar thought experiments to understand why space in front of you seems to shrink, why time dilation becomes a thing, and so on.

That explanation seems like it would not line up with the mathematical reality of the situation. It seems like one of those handwave-y things that always confused me as a child. “Gravity is just massive objects deforming space like a weight deforming a sheet, and things fall into the well they make.” Ok but what would make something fall into the well, there is no gravity.

That is a very good point! Gravity is just such an ingrained intuition that people tend to be ok with saying things go into the spacetime wells, but it is a little tautological.

My understanding is that a more correct intuition is thinking of straight paths on the curved sheet. Say it's like a loosely woven tablecloth - objects in free fall will go along the threads of the weave, so if you stretch the fabric by placing a heavy object on it, the paths of smaller objects on the fabric will be stretched towards the heavy thing.

This metaphor falls apart for orbits though, as it requires "stretching" the fabric so much that the threads now somehow go in a circle around the mass heh. But the underlying principle is the same - an object in orbit is in free fall along a straight path in curved spacetime.

In my explanation, if the particles trajectory can change, you can see how gravity affects distortion of space. Its not really that the space is some sort of entity that is being distorted, its the concept of measuring what you define as distance gets influenced by how these particles fly about.

No. What you described is still 100% Galilean relativity. Special relativity cannot be explained with Galilean relativity.

I guess I forgot to mention one fact - the particle speed is set and constant, and does not change.

The point overall is to demonstrate that ideas of time, space, and speed are all just sequences of events from which you derive those metrics, and relativity is the fact that those sequences change with all three.

>I guess I forgot to mention one fact - the particle speed is set and constant, and does not change.

Constant in all inertial frames of reference. That's the important bit.

I think I get it … kinda. Thank you.

Hi, as a person who can only read words with 4 or less characters your explanation is really confusing

I personally don't find metaphorical explanations helpful, especially considering this is not the only time I have heard or will be hearing about relativity, so if I get another explanation I will have to either map the concept of balls to whatever metaphor another teacher uses, which is just more work. I'm fine with using generic words like 'information', which I can map more naturally to other explanation wordings like 'signal'.

The same applies for explanations of bitcoin, or Machine Learning, or stock markets, just use the proper wording, difficulty, weights, secondary market. Metaphors are not teaching.

Reminds me also of the "Up Goer Five". An xkcd poster which roughly explains Saturn V with only the top 1000 used words in English[0]. Even better IMO is the collab video with MinutePhysics[1].

[0]: https://xkcd.com/1133/

[1]: https://www.youtube.com/watch?v=2p_8gx-XHJo

Randall Munroe (of xkcd) went on to write a full book in that style: https://xkcd.com/thing-explainer/

People talk about the 'good old days' of the web, but boy, in a multi-tab environment it stucks to try and read something that doesn't put any effort at all into side margins.

Reader mode (FF) helps a lot here.

And yet, it is 1000 times more readable than any "modern" website.

Hell no. Typograhy class Lesson 1, never put more than a certain number of words per column (in any case never the whole screen width, on desktops)...

And yet, for some reason, most people want their tabs and toolbars on the top, and the taskbar on the bottom, taking away vertical space that is actually used, and leave swathes of space on the left and right completely unused. I will never understand the resistance against putting things on the side.

Probably because English is written left to right and doesn't work so well top to bottom. So vertical bits tend to not include text, which is quite limiting. I do remember older UIs that included text at 90 degrees on sidebars, but that seems to have gone completely out of fashion.

The real problem is our screens are the wrong shape. 16:9 is a stupid aspect ratio for a computer monitor if you work mostly with text. Square is probably the best. Using floating windows (like classic Mac OSX) is an option, but for some reason people like full screen windows these days. A tiling window manager is another option to effectively divide the screen up into better shaped areas. I tend to have 6 columns across 2 screens which works well.

> 16:9 is a stupid aspect ratio for a computer monitor if you work mostly with text.

Why? I also usually use it with three columns of text, and it works well at that aspect ratio.

Right, so you essentially use mini 16:27 screens which is a better aspect ratio than 16:9.

What you really have to ask yourself is why 16:9 is a better choice than anything else. It wasn't picked because 3 16:27 columns is ideal for text. The main reason we have them is because it's easier for the screen manufacturers to have just one aspect ratio to deal with. With 16:9 we're forced to have three columns per screen, but what if I only need two? What if I want more lines of text per screen? Square seems like a much more obvious middle ground.

That makes sense, thanks. I've seen people give similar praise to 4:3 and 3:2 laptops.

>The main reason we have them is because it's easier for the screen manufacturers to have just one aspect ratio to deal with.

Not just that, but it was an average of aspect ratios considered for the aesthetic of video data. Not text presentation.

> Probably because English is written left to right and doesn't work so well top to bottom.

My browser tabs are on the left side of the screen and they are all written left to right. There is ample space on the sides of websites to fit this and show the entire tab title. Meanwhile, other people have tabs along the top of the screen that all show one or two letters followed by “...”.

My taskbar is on the right and consists entirely of icons. The only text on it is the date/time and that is also written normally and not rotated.

This is the setup that I was saying I don't understand the resistance against — not 90° rotated text. When people say “vertical tabs”, they are not talking about rotating the text.

> I will never understand the resistance against putting things on the side.

It is a side effect of mobile devices -- Mobile-first design. When apps got popular, people decided that everything needed to be designed for mobile devices first, and adjusted for larger browsers screens later. So people started making everything tall and narrow. And said it was bad practice to do otherwise. Now people who learned UI design after those days often follow that idea. Personally, I am not a fan, but that is where it comes from.

That explains the layout of websites and web apps, but not the desktop taskbar and the desktop browser’s tab bar.

the lack of large video ads really is jarring too

I was hyper confused that I didn’t need to hide ANY HTML element.

Also, reader mode was not suppressed by some naughty trick.

Unreal.

And what is the deal with functional scrolling?

It also did not ask about my cookie policy, how am I supposed to tell the site my preferences? I just wrote them via email to make sure.

It depends on the purpose of the cookies: https://www.blackboxdesign.co.uk/gdpr-do-i-need-a-cookie-war... for in-depth answer.

Works great on mobile, fwiw

What does multi-tab have to do with it? You are in control of your computer aren't you? Just make the window narrower.

Or set the zoom to around 240-300%

It's annoying for sure, but at least you can resize the window.

Side note: Dan Luu claims[0][1] that there's no readability advantage to narrow line width. I haven't really looked into it, but in my experience it feels like he's very wrong.

[0]: https://danluu.com/slow-device/ [Appendix: this site vs. sites that don't work on slow devices or slow connections]

[1]: https://nitter.net/danluu/status/1115707741102727168

I'd suggest you not take readability advice from a guy who uses a red-green color scheme for their tables.

He makes that claim because he doesn't like the sources, not because the sources are wrong.

Reading the wikipedia references show me that this is well studied and sourced.

Academic sources for his argument either dont exist or are so rare even he can't find them.

IOW, it's a post hoc rationalisation from him.

Open the "developer's tools", find the '<body>', inject a 'margin' CSS - customize the page locally.

Folks, just for these kinds of websites I made an extension that trims the body of the text to 80 characters. I don't have a way to pay to get it on google's or firefox's extension marketplace, so you'd have to install it from source.

https://github.com/flysand7/breader

Well I can see why you can’t afford to put it on a marketplace.

Just a suggestion

You can write this feature as a userscript and then publish it to greasyfork

We did have ways to create margins, you know :/ Aside from simple CSS, you could still do it with pure HTML.

Screens were much narrower then so constraining the width of text was not necessary.

I have a bookmarklet, since forever, labelled "sane width", with the following code:

  javascript:(function(){var newSS, styles='body { width: 800px ! important; margin: auto !important } '; if(document.createStyleSheet) { document.createStyleSheet("javascript:'"+styles+"'"); } else { newSS=document.createElement('link'); newSS.rel='stylesheet'; newSS.href='data:text/css,'+escape(styles); document.getElementsByTagName("head")[0].appendChild(newSS); } })();

It forces the body width to 800px and centers it. Crude but it is enough for me.

I haven't checked and I don't know how it would render. But it is worth noting that since this was designed against an earlier version of css, it might render differently in older browsers.

For example, older monitors had less pixels, so it's likely that the wrapping was sensible in older monitor/browser configs.

To say nothing of browser defaults being different, if this was pre-css, then the margins might have been baked into the default browser interpretation. In other words, pre-margin property, a webpage without margin didn't mean "this has no margin", in the sense that a modern webpage without margin specified would mean "DO NOT ADD MARGIN TO THIS!".

Or, if you want to learn the core ideas and math involved, but in terms of computer programming, see:

Functional Differential Geometry by Gerald Sussman, same author behind Structure and Interpretation of Computer Programs

https://archive.org/details/FunctionalDifferentialGeometry/

> [R]ays move just the same if what puts them off is held fast or not. Izzy had no way to know that, back then, but it's true.

Isn't this historically inaccurate? Wasn't Einstein looking for an explanation for why Maxwell's equations mysteriously assigned a constant velocity to electromagnetic phenomena, apparently contradicting Galilean relativity?

Needs a (1999) tag

I turned this into a little audio book: https://www.pdftomp3.com/shared/67fcc7f933aa6c3115b114da

No, you didn't. This doesn't match the original text.

0:47 Added in text: "Okay, here's the text prepared for reading aloud."

0:58

Original: "Okay, yes, it's a dumb idea,"

Audio: "Okay, yes, it's a bit of a strange idea"

1:08

Original: "Or do you, say, list off to the left some? What I want to ask you is: Can you find out? Hell no. You can see that, sure."

Audio: "Or do you drift off to the left a bit? The question is, can you figure it out? No, you can't. You can see that."

---

It appears you are using "Variational Lossy Autoencoder (VLAE)" as the basis for your website[1], which might be good for simplifying more complex things but defeats the purpose here. It's using more than four letters in words, and censoring out "dumb" and "hell"?

Why don't you try pointing that another explanation of the theory of relativity without this limitation? Seems like that'd be a more interesting exercise.

[1a] https://www.pdftomp3.com/shared/67e178f428779824db2e06c6 [1b] https://pdf-reader-storage-f55b8c51173224-staging.s3.us-east...

Ah, I just want to clarify that I'm very unhappy about the censoring of "dumb" and "hell".

I allow the text to get slightly optimized for audio experiences, e.g. page numbers or mathematical notation gets replaced. But I have think about that again.

I wrote the OG paper about making LLMs do this task (before chatGPT came out too!!!)

https://arxiv.org/abs/2306.15926

Four letters is an interesting constraint, but it doesn't guarantee simplicity. I'd replace

> no one can say who's held fast

with "no one can what does move and what does not"

...and of course I missed a word. I meant to type:

"no one can say what does move and what does not"

On the theme of difficult ideas explained using a limited vocabulary, I suppose one has to mention xkcd's Thing Explainer as well as George Boolos's classic _Gödel's Second Incompleteness Theorem Explained in Words of One Syllable_:

https://en.wikipedia.org/wiki/Thing_Explainer

https://languagelog.ldc.upenn.edu/myl/BoolosOneSyllable.pdf

I'm patiently waiting for the translation of Misner Thorne and Wheeler to Toki Pona.

Reads like it could have been AI generated.

Not in 1999.

len("gravity") = 7 > 4

Fun fact, this is a great piece of art because it was written in 1999.

If you were to write this in 2025, it would be indistinguishable from trash.

So many doors are closing.

Reply on news.ycombinator.com