"It isn't magic"

[-]AnthonyC5mo153

I don't keep count, but there are many, many times I've had conversations with people who think something is fundamentally impossible when it's actually already been done, sometimes often and at large scale.

[-]Ben (Berlin)5mo150

Aye.

The one I notice the most is people saying "cows will still have to be milked", without having realised automatic milking machines arrived in the early 90s.

[-]diogenes5mo50

care to share any examples?

[-]exmateriae5mo164

Related to AI, two years ago, image generation. It was not solved but it was much more advanced than most of my friends realized. (well, they did not think any kind of image generation existed)

Today, translation is a big one. I'm telling people that most of the advantages of learning languages will soon disappear and they often get angry. Most still think that "thinking" AIs almost exclusively work by text interfaces and that audio generation is a completely unrelated arcane. They are scared of deepfakes but do not realize that if this works, it means a lot of other things work too. Like live audio translation. They were shocked to see what Astra or Meta promise to deliver not in ten years but ten months.

Writing too, by which I mean not the ideas but the writing itself. People either see it as mostly rubbish but with useful info or they feel the LLM, which of course is a real thing but that's because too many users only copy paste. You're not writing life changing essays or deep meaningful texts (yet) but I feel you can get out of the obvious LLM vibe with little effort. If I give the ideas and structure, I would rather ask o3/4.5 to write it than most of the people I know (and that's not counting for the speed advantage).

There's also the case of people who write a ton and don't realize that most people are unable to write a paragraph about anything. I have known a few who pretty much thought AI couldn't write without realizing that their level of "can write" is barely attained by 0.01% of humans.

[-]Said Achmiz5mo3-4

I’m telling people that most of the advantages of learning languages will soon disappear and they often get angry.

It seems to me that “most of the advantages of learning languages” do not, and never have, consist in merely being able to know what is the translated-into-your-native-language version of some text.

[-]exmateriae5mo1-2

What are they then? I'd say there were two massive advantages : reading text and talking. The rest is extremely marginal. Sure, there are a few people with specific cases where they have other interests in learning languages but when internet people all started to learn english, that was because everything good on the web was in english. They wanted to understand and communicate with others and that's pretty much it.

But you're already able to do both with current technology? Text translation is solved already and in most cases better than a human knowing the other language. Granted, voice translation makes for a janky conversation but you can already understand anyone anywhere anytime as long as you have access to a device. And this won't be a problem for long with the speed of progress and the new types of AI first devices that are coming in.

[-]Said Achmiz5mo20

What are they then?

There are two:

Reading literature / poetry / etc. in the original. Translations are fine for getting the meaning across, but different languages are, in fact, different; structure, prosody, nuances of meaning, various aesthetic details, usually do not survive a translation. (Conversely, appreciating a good translation is itself a unique aesthetic experience.)
Benefiting from different perspectives imposed by different languages. The strong Sapir-Whorf hypothesis (a.k.a. strong linguistic determinism) is false, but there is a weaker version which is entirely true: that language influences thought. Yes, it is almost certainly possible to think any thought in one language that may be thought in another, but this is akin to saying that all Turing-complete programming languages are equivalent: true in theory, mostly irrelevant in practice. Different kinds of programs are written in Lisp than in Go. And even if you only ever use one programming language, you will benefit from learning others, because it will expand your understanding of what you’re doing. Thus also with human languages.

Neither of these are “extremely marginal”. And neither can be done with AI (because the whole point of both is what’s going on in your head).

[-]habryka5mo50

FWIW, on the second point, I am a native german speaker (plus obviously proficient english speaker), and I don't think I have gained approximately any benefit from the second point. Like, as far as I can tell I just have a strictly harder time expressing things in german than in english, after having mastered both.

This is partially because english has a much larger vocabulary, and so there is almost never a word that you can't say in english, but can express in german (and in the rare circumstances where that is not true, english has helpfully imported many of the words that have no equivalent as loan words).

The primary thing I would recommend people do is if they do not speak english, they learn english. It's honestly just a much more expressive language for thought than at least german (and I am pretty sure also polish which I am a bit familiar with). It's possible there are other languages that are even better, though I am skeptical. I would definitely not recommend anyone learning german today on the basis of the second point.

(On the first point, I have also been disappointed by the benefits of reading german philosophy in the original language. At least for the continental philosophers, I actually had a better time reading them in their english translations, because the translator had to do a bunch of cognitive labor to make them less obnoxious/weird/obscure, but I can imagine that there are other works where that is less true, and there is real benefit)

[-]Said Achmiz5mo31

On the first point, I have also been disappointed by the benefits of reading german philosophy in the original language

Well, yes, but German philosophy is famously obscurantist. Like, “German philosophy” is the paradigmatic example of “continental philosophy which is impenetrable and which, one strongly suspects, is barely saying anything at all even once you get past all the layers of bizarre formulations and idiosyncratic terminology”. So it’s no surprise that you’d be disappointed!

At least for the continental philosophers, I actually had a better time reading them in their english translations, because the translator had to do a bunch of cognitive labor to make them less obnoxious/weird/obscure

I can easily believe this. I think that this is probably related to the point that David Stove makes in his famous “What is Wrong with Our Thoughts?”:

And when every reasonable allowance has been made for the real difficulty of translating Plotinus, say, or Hegel, into English, this will scarcely even begin to explain what is wrong with the passages above. We cannot understand, indeed, how anyone would come to say the things that Plotinus or Hegel says. But that they were saying, in Greek or in German, the same baffling things as they are found saying in good modern English translations, cannot rationally be questioned. (It is a very striking fact, however, that I had to go to translations for my three quotations above. Nothing which was ever expressed originally in the English language resembles, except in the most distant way, the thought of Plotinus, or Hegel, or Foucault. I take this to be enormously to the credit of our language.)

English, I think, is a strictly superior language for doing analytic philosophy (i.e., real philosophy, rather than obscurantism) than (according to Stove, and I guess also you?) German, or (according to me) Russian.

But! Note that my point #1 did not talk about philosophy, but rather about “literature / poetry / etc.”. I am talking about aesthetics, not about precision of concrete ideas!

FWIW, on the second point, I am a native german speaker (plus obviously proficient english speaker), and I don’t think I have gained approximately any benefit from the second point.

Fair enough, but I’m a native Russian speaker, and I think I’ve gained lots of benefit from knowing both languages.

The primary thing I would recommend people do is if they do not speak english, they learn english.

I complete agree with this. Everyone should learn English. This one’s basically a no-brainer.

[-]habryka5mo20

Well, yes, but German philosophy is famously obscurantist. Like, “German philosophy” is the paradigmatic example of “continental philosophy which is impenetrable and which, one strongly suspects, is barely saying anything at all even once you get past all the layers of bizarre formulations and idiosyncratic terminology”. So it’s no surprise that you’d be disappointed!

Well, I was hoping that given the combination of both widespread popularity and reputation for subtlety/nuance/ineffability (and insistence by at least some of my friends and acquaintances who had read the english translations and got lots of value out of them) that this would be one domain where I would be exposed to a particularly high gradient of value, so it was a surprise to me!

Like, the thing that was most surprising to me is that I did get value out of the english translations I read. Like, I think a bunch of the things were reasonably useful, and not just nonsense, but extracting that usefulness was substantially easier in the english version than the german version.

[-]exmateriae5mo10

Well we'll have to disagree on that. I have not said that there were no other benefits but that they were nowhere near communication and reading. Saying that those were not very largely the main benefits of language learning simply seems untrue to me and your examples are only comforting this view.

Both are nice things that come with a new language but definitely not something that would motivate the immense majority of people (and people on lesswrong are definitely not normal people) to learn a language if they were the only reason. I'm sure that's a thing in lesswrong adjacent communities.

I do agree that "many" people benefit from the first example but that is almost always a side effect : what they want first and foremost is access to the content itself. You could not read the sequences in italian 10 years ago so you learned english, had they been translated you would not have learned it.(terrible example but you know what I mean)

[-]AnthonyC5mo*53

Others have already added some good ones, here's a few more.

A few that are likely familiar throughout this community even if you have never considered them this way, but still often not believed or noticed elsewhere:

Diagnostic algorithms outperforming human doctors
Autonomous vehicles safer than human drivers

In a similar vein:

EVs being as safe and reliable as ICE vehicles (fire risk is the worry I hear about most regularly)
Solar and wind being cheap and predictable enough to make a significant contribution to the grid
Sustainable agriculture being able to produce enough food per acre to in principle successfully replace conventional agriculture

Some that are more obscure but have come up in my professional work:

Making a 3d printed part that's mechanically comparable to a conventionally manufactured equivalent
Heat pumps good enough to replace combustion for cold climate HVAC or industrial steam production

[-]Gyrodiot5mo116

Related recent post: Intelligence Is Not Magic, But Your Threshold For "Magic" Is Pretty Low (similar point, focused on human short-timeframe feats rather than technological achievements).

[-]Trevor Hill-Hand5mo22

These two posts pair well together.

[-]Dagon5mo*60

As you note, Arthur Clarke identified this in 1973, and it holds up: "Any sufficiently advanced technology is indistinguishable from magic." GPT2 is pretty close to "just math". Current frontier models are sufficiently advanced and indistinguishable from magic. As we study them and learn more, and develop more methods for introspecting their behavior, they'll become "just math", in the same way that, say heavier-than-air flight is. But by then, there will be newer/bigger/harder-to-trace mechanisms that are indistinguishable from magic.

It's quite likely that human brains will become "just math" at some point in the future as well. Until then, "magic" is a fine handle for "amazing and coherent high-level behaviors that we can't explain from lower-level observations/mechanics.

[-]Dagon5mo32

The presumption of complete reducibility is, with some great certainty, unclear at best and at worst absolutely impossible,

Oh, I fully agree. But "complete" is not necessary to achieve the change in categorization from "unknown and magical" to "just (big/difficult) math".

I don't know how much playing people around here have done with Mandelbrot set coding, but it's a useful comparison in this. It's very clearly NOT magic in any literal sense - the calculation is trivial (and even the iterations to determine converge/diverge is pretty easy to understand). But the results remain captivating and astounding (to me) that they come from such simple rules.

In this sense, I suspect many complex systems will remain impressive and astounding, no matter how good we get at modeling and understanding their components. In the sense that knowing the underlying rules DOES turn it from "fully magical" into "an interesting corner of math", this will probably happen to current LLMs, and likely eventually to primate intelligence.

[-]Ben (Berlin)5mo10

Cross-posted from my blog on github: https://benwheatley.github.io/blog/2025/06/22-13.21.36.html

LESSWRONG
LW

LESSWRONG
LW

91

91

91