Published on November 4, 2025 7:47 AM GMT

(I’m doing Inkhaven and have two actually important/potentially impactful posts coming up, and really want to polish both a bit more before publishing, so I wrote this quickly to have a thing to post instead. Apologies.)

Ever since I tried meditation, I love food.

Around April 2022, a friend convinced me to try Sam Harris’ Waking Up app

Meditation made me actually pay attention to my experiences; and the experience of food suddenly became much higher-dimensional, once I started paying attention.

By October 2022, I’ve been to a dozen Michelin-star restaurants.^[1]

At some point, I participated in a chocolate tasting hosted by Duncan Sabien. I couldn’t stand dark chocolate prior. Afterwards, I could no longer eat milk chocolate: it’s just too bad compared to good dark chocolate.

The really surprising thing about the tasting wasn’t that if I focus on the taste of chocolate, I could map it–maybe internally combined with the difference with some central chocolate-taste–to some ideas, feelings, concepts, objects, moods.

The surprising thing was that people independently (without groupthink, in separate groups!) came up with the same comparisons.

Pure dark chocolate: cocoa and sugar.

That tastes like strawberries. According to multiple people who try it and come up with this comparison before sharing it with their groups, before sharing it with everyone.

Or pure dark chocolate: cocoa and sugar.

That tastes like a muddy Amazon river.

A muddy Amazon river.

It was, I think, my fifth chocolate of the tasting. I thought about what the taste is like. And came up with the imagery of a strong, muddy Amazon river.

Other people came up with the same imagery.

This was very surprising. Chocolate shouldn’t taste like muddy rivers, to different people, was my assumption a day earlier. I assumed all of the tasting notes on wines were bullshit; this is not a thing that can happen. Sommeliers are fake; I vaguely recall studies that show that they can’t distinguish red wine from white wine with red dye.

Strawberries are okay: fine, I can imagine that the difference in embeddings between this specific chocolate and some central chocolate is similar to the embeddings of the taste of strawberries.^[2]

But why would it taste to different people like a muddy Amazon river? This is insane. What’s going on.

My theory is that paying attention to good food is like experiencing embeddings. Like a spectrogram of different things. Or music.

Good food would often have an interesting spectrogram/embeddings: narrow and pointy in exact places that complement each other in interesting ways; or maybe roundy and still causing a feeling of comfort; or some combination of the two. There is a huge amount of dimensions of the experience, and there is maybe some amount of rotation that can be applied while preserving some properties of experience.

It’s what it feels like, to eat good food. You can enjoy it the way you can enjoy an orchestral piece, feeling and understanding it as all the separate instruments and groups of instruments, enjoying how they come together and play with each other, and on a separate level, understanding and feeling the resulting sound as a whole, in the moment, and separately again, tracking and understanding the entire piece, from start to finish.

Three of these define how I listen to my favourite music: the entire progression and emotion and growth of some pieces; the sounds as I perceive them; and then, simultaneously, all the small things I can identify and separate and understand as instruments, and enjoying them as individual components beautifully contributing to the whole.

Great food is somewhat like that: there are the individual components you can identify pay attention to, while enjoying the whole; and sometimes, though less centrally, there’s also a change to them as you progress, but you rarely think of this progression and mostly just enjoy experiencing the embeddings/the spectrogram moment by moment.

While I’m at it: I often tell people that yFood^[3] is better than the bottom 10% of meals I’ve tried in Michelin-starred restaurants. (And I select meals mainly by how much I expect to enjoy them, though occasionally that includes interestingness or exploration value.)

They don’t normally believe me.

See, ready-to-drink Huel is a random mess of embeddings. It’s fine, it’s convenient, it’s not that terrible, but it’s not actively good and its taste is not particularly uniquely enjoyable.

yFood is actively good. It tastes like an incredibly good milkshake. Somehow, it has a nice, soft and somewhat gathered taste. I don’t know how they did it, but I like it, a lot, and so most of my breakfasts are yFood.

But what’s going on with chocolate?

It doesn’t have a variety of components^[4]. I guess the chemicals are complicated; I don’t have time to do a deep research query and write down myself having been deconfused; I can only share my confusions and experiences.

If you have ideas or independent theories of qualia that would’ve predicted this, please share.

(Thinking about it does somewhat update me that at least some people experience red the same way, or at least a have embeddings for things equialent if you adjust for rotational/other symmetries.)

1.  I was lucky to have had a lot of money at some point in my life, though feel somewhat guilty I’ve spending some of it on fancy food instead of donating, even though I’ve donated significantly more to MIRI than I spent on fancy restaraunts.

2. Even though this still makes no sense: what would that correspond to, in reality? Pure chocolate, chemically, is nothing like pure chocolate + strawberries.

   And why asking yourself what kind of ship or a mood of a sea or a forest a piece of chocolate is like works at all?

3. It’s like Huel–nutritionally complete–but contains milk. If you’re around the Bay, you can try some from me while I’m at Lighthaven, but you can’t easily buy it outside the UK/EU. Apologies to people in the US.

4. (Some good dark chocolate can have other stuff added to it.)