Talk English, Think Something Else

• AI News Team
• April 13, 2026

​There’s an adage from programming in C++ which goes something like “Yes, you write C, but you imagine the machine code as you do.” I assumed this was bullshit, that nobody actually does this. Am I supposed to imagine writing the machine code, and then imagine imagining the binary? and then imagine imagining imagining the transistors?Oh and since I don’t actually use compiled languages, should I actually be writing Python, then imagining the C++ engine, and so on?Then one day, I was vibe-coding, and I realized I was writing in English and thinking in Python. Or something like it. I wasn’t actually imagining every line of Python, but I was imagining the structure of the program that I was describing to Claude, and adding in extra details to shape that structure.Pub Philosophy BrosThis post is actually about having sane conversations with philosophy bros at the pub.People like to talk in English (or other human languages) because our mouths can’t make sounds in whatever internal neuralese our brains use. Sometimes, like in mathematics, we can make the language of choice trivially isomorphic to the structures that we’re talking about. But most of the time we can’t do that.Consider the absolute nonsense white horses paradox, where “a white horse is not a horse” is read both as the statement:And the phrase “a white horse is a horse” is read as the statement:I often think in a language of causal graphs. English isn’t very good at talking about causal graphs. It doesn’t have individual words for “A contains the same information to B”, “A is the same node as B”, “A is an abstraction over B”, “A is a node which is causally upstream of B”.I remember talking about “consciousness” with a philosophy guy at the pub once. I think I said something like “A certain structure of computation causes consciousness” meaning “Consciousness is an label applied to certain computational structures”, but which he interpreted as “The presence of a certain computational structure is a node upstream of consciousness”. This caused immense confusion.I call the problems here “beetle problems”Beetle ProblemsWittgenstein proposed a thought experiment. Suppose you have a society where:Everyone gets a box.Everyone uses the word “beetle” to refer to what’s in the boxEveryone can look in their own boxNobody can look in anybody else’s boxIn this case, the meaning of the word “beetle” is entirely socially constructed. Wittgenstein was exaggerating here: if I talk to you, and you do something with your beetle (dirty jokes aside) and report the results, I can get some information about your beetle, based on what you say back to me. The beetle is causally entangled with us both. It’s just not a very efficient way of talking about things.Even if we both have identical beetles, it might take us a while to get them oriented the same way round, what I call an antenna, you might call a leg, what I call a wing-case you call a carapace. And so on. To unfairly single out an example. I personally find this particularly salient when talking to people in the Oxford EA/longtermist cluster. I know they’re smart people, who can put together an argument, but they’ve developed a language I just cannot penetrate. It takes a long time for me to figure out what on earth they mean. Ohh, you have your beetle upside down compared to mine.Even worse, I think a lot of people don’t actually think in terms of causal graphs the way I do. This comes up when I try to read pieces on moral realism. When someone brings up a stance-independent reason to do something, I simply cannot map this onto any concept which exists in my mental language. What do you mean your beetle has fur and claws and keeps going “meow”? Are you sure?SolutionsUhh… I don’t have many. Beetle problems take a while to figure out. I once got feedback on an essay test that said “Your ideas seemed confused.” and I thought “Man, your draft seemed confused!”. I don’t think I could have done much better, without spending time in person hashing out the beetle problems.It might have helped to have a better conception of beetle problems, though. I could at least have pointed it out. Perhaps in future I’ll come back with a wonderful beetle-solving problem.Editor’s note: this post was written as part of Doublehaven (unaffiliated with Inkhaven).◆◆◆◆◆|◆◆◆◆◆|◆◆◇◇◇◆◆◆◆◆|◆◆◆◆◆|◆◆◇◇◇Discuss ​Read More