"and then crushes it with a pressure equivalent to an adult Arabian horse standing on a quarter", what a bizarre comparison, this seems like an incomprehensible reference point for most people (myself included)
You just need to find out how much an Arabian horse weighs, and divide the area of the thing they're standing on by 4, and the calculations are pretty straightforward.
I'll look it up. The horse weighs 360 to 450 kg, the quarter in this context is an American coin with diameter 24.26 mm .. I think it's pretty clear now.
Cross checking with units : A coin of diameter 25.4mm=0.0254m has an area of pi(0.0254m/2)^2=0.00016129m^2, and 9.81N/kg450kg/4=1103N, so that gives 1103N/(0.00016m^2)=6.819MPa. Sounds right, pascals are quite small. :)
... I had the 'pleasure' of driving American cars and if there's one thing they don't have it's good suspension. Though maybe they finally stopped using leaf springs for newer models.
In this context, "inch" is a verb, not a noun. It represents slow movement, not a precise measurement. There are many words like that in English.
To answer your question, "moving" would have been a perfectly acceptable substitute. You could also use "crawling", "creeping", "worming", or "wriggling".
But that doesn't answer my question, as it appears there is no metric equivalent. I'm just highlighting how ridiculous the GP comment is, to conclude that using an analogy of a horse and a 25-cent coin has anything to do with metric measurement.
It does answer your question. You just don't like the answer, which is that there need not be a metric equivalent when the word does not represent a measurement.
Here's another example: what's the metric equivalent of "footing" the bill? The verb "to foot", in this example, has the exact same etymology as the noun "foot", but it isn't used as a measurement.
It's like asking what the imperial equivalent is for "the poem's meter". The question is nonsensical, unlike the comment you were replying to.
Sure, the comment was whimsical and flippant, but it was based on a legitimate objection: why express the pressure in an incredibly obscure way? The objection has sense, and the joke is related to it. Your knee-jerk reaction to the joke, however, conflates disparate categories.
This is a human watermark to ensure readers an LLM didn’t author the article. There’s no reasonable prompt that could create this comparison without hallucinating the rest of the article.
The venerable 'Half A Giraffe' subreddit covers the frequent use of nonstandard journalistic measurement units --
https://old.reddit.com/r/HalfAGiraffe/
AI used to be autonomously controlled systems. Now AI is re-brand of machine learning by marketing departments, etc.
This seems more like AI than the very recent rebrand of ML. Machine learning can be a sub-system of AI systems but in itself was not thought of as AI for a long time.
isn't it? I think it might be. It's definitely artificial so it depends on your definition of intelligence. And that's a rather confused subject.
I favor a model of intelligence where one measures problem-solving capability along multiple axes and defines an N-dimensional shape that describes a hypervolume of 'general' intelligence. One could certainly argue that because this program solves problems that can only meaningfully be measured along one axis that it has too few dimensions to be called intelligence. The trouble with that is software generally works best when specialized for problems with a single axis, so by that definition 'general intelligence' is either unlikely to happen in that space at all or won't be energy efficient if it does.
What would likely have a higher intelligence is a 'cyborg' - which could just be a human accessing this program via conventional means - who adds this program's problem solving capability along one axis to their existing hypervolume of intelligence. So maybe not artificial intelligence, but definitely an artificial intelligence expander.
If you mean the particular subtype of AI that is the currently fashionable shiny object people are chasing, then no, it is not.
If you mean it's AGI, then no, definitely not, but then, nothing is. (Whoever was about to stand up and say "how do you know that ChatGPT isn't AGI", don't. You'll thank me in three years that those words are not attached to your name in public.)
If you mean, does it fall into what people would have described as AI before marketers, executives, and other similarly ignorant people started using "AI" to refer specifically to generative models, then yes, it absolutely is.
AI was never well defined. AB tree pruning chess engines were called AI at one point. Game programmers call A* and behaviour trees - AI.
At this point it means nothing.
Call it machine learning or automated optimization. Arguably you could do the same thing with a greedy algorithm iterating over all solutions. Would that be AI too?
Which is why we should stop using it, and refer to A* as A*, LLMs as LLMs, evolutionary algorithms as evolutionary algorithms, et cetera. I doubt this will happen in my lifetime.
Unfortunately the moment you get dollar bills attached to particular words, everyone starts using them. AI sells. Evolutionary algorithms do not. So everyone chasing the sweet dollar will say AI and other buzzwords.
> The robot creates a small plastic structure with a 3D printer, records its shape and size, moves it to a flat metal surface—and then crushes it with a pressure equivalent to an adult Arabian horse standing on a quarter. The robot then measures how much energy the structure absorbed, how its shape changed after being squashed, and records every detail in a vast database.
Yes, but the robot is presumably also ranking them and doing some clever "ai" thing to actually generate the model, like a genetic algorithm or something. It's also working 24/7, and I think if you make humans work 24/7 they get angry and ask to go to the bathroom and such
I mean you're welcome to manually do all the calculations for the gradient descent stuff, redo the mesh yourself and carve it out of a block of plastic with a chisel, I'll stick to getting a computer to do it for me
I'm curious what the difference between this and an "automated scientist" like sama/openai says they want to build would be. This is what I imagine when they mention that goal.
I’m interested in the structure it created - is that it in the hero image? The article needs to highlight it more! There should be at least a paragraph with a figure of this 75% structure petal looking thing that an Arabian horse could crush.
59 comments
[ 3.6 ms ] story [ 138 ms ] threadhttps://thehorse.com/tools/adult-horse-weight-calculator/
(reference: https://old.reddit.com/r/Metric/comments/cpbi2g/americans_wi... )
What is the metric equivalent of this statement? (from TFA)
To answer your question, "moving" would have been a perfectly acceptable substitute. You could also use "crawling", "creeping", "worming", or "wriggling".
The one that I specifically described as a meme comment? You're saying it's technically inaccurate? Thanks, criticism noted.
Here's another example: what's the metric equivalent of "footing" the bill? The verb "to foot", in this example, has the exact same etymology as the noun "foot", but it isn't used as a measurement.
It's like asking what the imperial equivalent is for "the poem's meter". The question is nonsensical, unlike the comment you were replying to.
Sure, the comment was whimsical and flippant, but it was based on a legitimate objection: why express the pressure in an incredibly obscure way? The objection has sense, and the joke is related to it. Your knee-jerk reaction to the joke, however, conflates disparate categories.
They don’t have centimeter worms in the metric world I guess?
If we had a machine to convert that horse into energy we could power Britain for about 20 years. (Britain uses a few quadrillion BTUs annually).
It is not a language model, and it doesn't run on a GPU, but it is definitely a mechanical implementation of searching a solution in a space.
This seems more like AI than the very recent rebrand of ML. Machine learning can be a sub-system of AI systems but in itself was not thought of as AI for a long time.
I favor a model of intelligence where one measures problem-solving capability along multiple axes and defines an N-dimensional shape that describes a hypervolume of 'general' intelligence. One could certainly argue that because this program solves problems that can only meaningfully be measured along one axis that it has too few dimensions to be called intelligence. The trouble with that is software generally works best when specialized for problems with a single axis, so by that definition 'general intelligence' is either unlikely to happen in that space at all or won't be energy efficient if it does.
What would likely have a higher intelligence is a 'cyborg' - which could just be a human accessing this program via conventional means - who adds this program's problem solving capability along one axis to their existing hypervolume of intelligence. So maybe not artificial intelligence, but definitely an artificial intelligence expander.
If you mean it's AGI, then no, definitely not, but then, nothing is. (Whoever was about to stand up and say "how do you know that ChatGPT isn't AGI", don't. You'll thank me in three years that those words are not attached to your name in public.)
If you mean, does it fall into what people would have described as AI before marketers, executives, and other similarly ignorant people started using "AI" to refer specifically to generative models, then yes, it absolutely is.
Call it machine learning or automated optimization. Arguably you could do the same thing with a greedy algorithm iterating over all solutions. Would that be AI too?
Which is why we should stop using it, and refer to A* as A*, LLMs as LLMs, evolutionary algorithms as evolutionary algorithms, et cetera. I doubt this will happen in my lifetime.
Try, test, fail, change parameters, repeat.
The 'AI', to me at least would suggest some kind of intelligence to 'make informed assumptions' at least.. But not sure. Its confusing.
https://www.nature.com/articles/s41467-024-48534-4 (Open-Access)
I had wondered why they weren't doing this in simulation but didn't realise that Finite Element Analysis gets weird in the regions they're exploring.
Reminds me a lot of the "Chemputer" concept for automated chemical synthesis expriments. https://www.chem.gla.ac.uk/cronin/
So, something that humans absolutely could do.
I'm only focusing on that. We could use the exact same process and tools minus AI and do it, which would be slightly less efficient, that's it.
People invent tools to increase our abilities. That doesn’t mean when a human uses a tool, the tool does the work.
A human made this thing. The AI was part of the process but a human did it.
In a very literal sense, tools are "doing the work" to precisely the degree to which it's not food calories that empower them.
Good job.