In that sense, no underlying physical state could be said to hold “more” information than any other, right?

In an information-theoretical sense, you can have a message that has a lower or higher information content. This is where entropy gets derived from. But it only makes sense for a fixed distribution – a more likely outcome has a lower information content. So I think you could have a physical state holding more information, if it’s a less likely state for some fixed definition of likeliness.

This would probably be closer to an actual link between informational entropy and physical – a given microstate has lower physical entropy when it is a less-likely state (e.g. half-squished cup of coffee), and that state would have higher information content if we considered the state as the message. This intuitively makes sense, because physical entropy is in some sense the ability of a system to undergo change, so indeed a low-entropy system is “more useful”, just like a message with higher information content is “more useful”.

The thing already is a markdown renderer and every single markdown renderer I encountered supports maths within $ delimiters.

Also can we make awful.systems render Latex in posts, I had to screenshot my formulas to put them here and I feel unclean

This is the only way you could make me care about the Simulation Hypothesis, if it runs on a spreadsheet then I will make it my life’s mission to break out just to yell at them for being terrible at engineering and replace them with a small shell script

DO NOT OPEN THE DOOR there’s a weird sex scene inside I guarantee it

That’s the closest you’ll ever get to an admission from a corporate mouthpiece though.

as if it makes any difference whether the universe runs on (…) Microsoft Excel

Okay but it is very spiritually important for me to not be that, please.

Storing a message in a system doesn’t make new microstates. How could it?

Lol I got so tripped up by him later saying “this is no longer clearly 0 or 1 so it doesn’t exist” and decreasing N that I missed he does the reverse thing when encoding the message.

This is like the ontological argument. He creates a virtual entity from words alone and then treats it as a physical thing storing energy. And then once it no longer fits the words of the definition, poof, gone it is, oh look, total entropy decreased.

We have to consider probabilities, not just for where the pieces are, but also for how they are moving.

I completely omit that because, well, it’s hard, but also I don’t think it’s necessary here. This approach doesn’t work even if you consider only positions and assume uniformly random momentum. It doesn’t work even if the microstate is “is this pixel more red or more blue” in the paper’s experiment!

But thank you for the comment, I’m glad I didn’t completely butcher entropy with my weird nonrigorous internal model I developed based PBS Space Time videos lol

but you aren’t quite right about some of the details.

I’d be happy to be corrected.

This isn’t too outlandish, and modern studies of quantum mechanics suggest that information is a conserved quantity,

I hope I didn’t pass it as if it was completely out there, that information has to have some physical properties and energy as a carrier is a very reasonable hypothesis. The Landauer principle is not that controversial, I’m sad I’m too stupid to actually understand the discussion around it on any reasonable level lol

Oh, it’s worse than “outlandish”. It’s nonsensical. He’s basically operating at a level of “there’s an E in this formula and an E in this other formula, so I will set them equal and declare it revolutionary new physics”.

I meant the experiment itself. Like it looks like something you could try and do and measure and get an actual answer?

https://awful.systems/post/2876657

This is probably the least surprising thing ever.

CocaCola is like the symbol of capitalism. Everything they produce is corporate slop. GenAI is a perfect fit – soulless, artless, hastily slapped together bright pictures that ultimately don’t matter and carry no value. The world is not better with CocaCola ads, and it would be no worse without them. They’re just there, to be lost in time, forgotten. Like tears in the rain.

You know, you can just click on the link in the article to watch the ad it talks about

Radical thought, maybe read the article?

If you’re hearing this, then there is still hope. Hope that you can avoid making the same mistakes we made.

So, how does Abdaal watch anime and TV productively you might ask? Well, the fact that his listens to audio books on 3.5x speed should give you an idea.

[…] normally what I do is, I’ll just speed-speed-speed-speed-speed-speed-speed up until it gets to an interesting point, and I’ll speed it as fast as I can so I can still keep up with it.

And because he obviously can’t hear what’s being said when watching at 3.5x speed anymore, he’s speed-reading subtitles.

Lol, great, at that point you can just as well read the plot synopsis on a wiki or something. Or ask ChatGPT to tell you what it was about… aaand I just rediscovered the main thesis of this essay, nice

I see calling like Raegan’s margin in 1984 “decisive”

Also, not that it’s the point but I have to note that technically most election victories are decisive, in the sense that they resolve the winner with little to no ambiguity (which is usually the case, even when the margin is narrow). In that sense, the only way Trump’s victory is not decisive is if you contest the legitimacy of the whole election.

This is such pedantry that you might as well say “the Merriam-Webster dictionary defines decisive as…”

Also it isn’t? 50.2% to 48.1% of votes is not decisive in any sensible meaning of the word?

If you account for the turnout (around 60%) it means 30% voted for Trump and 28.9% for Harris, so “none of those” won decisively with 40%!