Nice to see e-flux, one of the oldest netart sites on the internet, still publishing. For a deeper dive into fascinating origins of materialism in Ancient Indian philosophy, the following provides a decent intro.
"However, in importing these labels from Western philosophy to the classical Indian philosophical systems, one needs to exercise caution because the concepts of nature, science, scientific method, etc. do not smoothly converge in the two theoretical traditions."
Despite the way it is often framed and critiqued, artificial intelligence is not really “artificial” or “alien”: in the usual mystification process of ideology, it appears to be a deus ex machina that descends to the world like in ancient theater. But this hides the fact that it actually emerges from the intelligence of this world.
What people call “AI” is actually a long historical process of crystallizing collective behavior, personal data, and individual labor into privatized algorithms that are used for the automation of complex tasks: from driving to translation, from object recognition to music composition. Just as much as the machines of the industrial age grew out of experimentation, know-how, and the labor of skilled workers, engineers, and craftsmen, the statistical models of AI grow out of the data produced by collective intelligence.
Quite a nice excerpt, will be reading the article now.
We almost always discount the work done by our predecessors: NIH-syndrome/"reinventing the wheel" vs "standing in the shoulders of giants", and this was a humble reminder for me.
> We almost always discount the work done by our predecessors: NIH-syndrome/"reinventing the wheel" vs "standing in the shoulders of giants", and this was a humble reminder for me.
"We" being computer scientists (or computing) which for some reason has this disease, perhaps due to the field's newness? When I worked in the life sciences (pharmaceutical chemistry) the literature was the first place we looked.
Whereas when working at a research lab in my 20s: I would often go down to our library and read CACM, IEEE journal et al and would sometimes implement things I read there. Even though I would brandish a photocopy of the journal article and reference it in the source code my boss and colleagues would praise me for "thinking up" a solution to a given problem.
Ironically the Internet changed things dramatically in the physical sciences (in Physics especially* but in other physical sciences by making it much easier to both search and read the literature) whereas in non-academic computing it's been the small conferences and, to my surprise, Youtube.
* Physics has long been an innovator in information dissemination; when journal publishing got too slow and letters too point-to-point it gave rise to Physics Letters, which itself became a pair of journals....
Great to see this on hacker news. E-flux publishes a lot of voices (artists, continental philosophers, and others) that have sound, intellectual criticisms about te ways we incorporate technology into our lives—and I think it’s important for tehnologists to hear them. It’s not common for these voices to reach traditional comp-sci/mega tech spheres.
They put out a nice series of readers based on their journals and other material. I recently read a great one that collected some of the works of Hito Steyerl.
I'm trying to work in that direction on many fronts. Hackernews is not really one of them but this post was really in Hackernews style so I decided to post it here.
That’s great. This is something I’m interested in as well. I found your critical essays repo on github. I have committed time to studying the philosophy of technology and am particularly familiar with the works of the Frankfurt school, some of Paul Virilio’s work, and the work of Lewis Mumford. I’d be happy to contribute if you’re looking for contributors for any critical projects.
"Even at the end of the twentieth century, no one would have ever thought to call a truck driver a “cognitive worker,” an intellectual. At the beginning of the twenty-first century, the use of machine learning in the development of self-driving vehicles has led to a new understanding of manual skills such as driving, revealing how the most valuable component of work, generally speaking, has never been merely manual, but also social and cognitive"
At the risk of sounding like a broken record I must point out that AI, neural
networks, machine learning, Perceptrons, algorithms and deep learning are not
interchangeable terms, any more than "maths" is interchangeable with
"calculus", "trigonometry", "geometry", "algebra" and "arithmetic".
Unfortunately this article is typically using such terms as synonyms and,
while it's probably meant to help the lay reader understand a technical
subject, will only manage to confuse the issue further.
It's clear that there is intense interest in deep learning and machine
learning in the last 7 years or so, but, especially when talking of "AI", it
is important to remember that the majority of work in AI has _not_ been on
machine learning, but on inference and reasoning, and using symbolic,
logic-based techniques (including Bayesian reasoning, as in Solomonoff's
Inductive Inference) rather than "statistical" methods (in the sense that the
term is used in AI to mean everything from probabilities to calculus).
Finally, there are historcial inaccuracies in this article. For example,
Rosenblatt's Perceptron was not "the first machine-learning algorithm' as the
article says: Arthur Samuel's self-playing checkers agents and Minsky's SNARC
are two systems predating the Perceptron by a few years and research on
machine learning techniques like Markov chains goes back to the 1910's.
While agree with you on the need to straighten the narrative and usage of these terms, the author is a philosopher and uses AI not in the common sense nor in the Computer Science sense. It is indeed confusing but he's very very rigorous in the way he writes (this is extremely simple writing for his usual style) so don't believe he's using the terms intercheangeably
Most of these philosophical thinkpiece authors are charlatans. Charlatan philosophers can get along very well, even in professional circles, because there's almost no way to tell the difference between a true belief and a false belief in some areas of philosophy. However, they meet their downfall in articles like these, when they make the mistake of talking about something that is relatively easy to verify - revealing that they are sloppy thinkers with great English skills.
I don't get what you're talking about. I'm a Machine Learning engineers and I find his work one of the best being conducted on the topic. You can say he's snob or inaccessible, but for me Pasquinelli is in the top 1% of the tech critique, especially in the theoretical part.
Also he's not even a native speaker and I don't find his way writing very compelling. Quite the contrary.
I don't know anything about the author of the piece. I understand that it's a sketch of a larger article that he intends to write for publication. I hope that he will have the time to refine his scholarship until then.
I'm not sure what to do of the ideas discussed in the article, themselves. I am honestly confused about what he means when he's talking about "AI", given his loose use of terminology, so I'm not sure I understand the article clearly enough to agree or disagree with it.
To clarify, I'd be all for reading a strong critique of the modern (last decade) use of machine learning, including neural networks and deep learning (but not only). If he's talking about AI as a whole, however, I don't see how his critique can really work. I don't think I've ever heard of anyone using a classical planner or a constraint solver to drive an automated truck, say.
The start of the article, with the discussion of the Agnicayana ritual was very intersting.
Yeah, of the things that he writes about that I know of, he seems to not quite get them right, but this is still an interesting article with a lot of big ideas to think about.
If he sounds like he's bending things, that should sound good to your ears. Otherwise he would be just reinforcing what you already believe and how you already see the world and therefore not doing any good philosophy.
I know you’re being sarcastic, but I’ve integrated this idea into my bullshit detector. When I’m dealing with someone who purports to be an expert in a field I know little about, I will try to steer the conversation towards things that I know a lot about—without disclosing that I know a lot about them.
People will show you who they are if you let them.
I vaguely recall reading a paper about oracle bones and hunter-gatherers, with the central theme being a study of generating randomness by the means of an "algorithm". I recall it was on arxiv but that doesn't seem to be the case anymore.
Algorithm being defined as: take a bunch of animal bones, burn them until they are essentially cremated, and then interpret the ashes somehow to point you in a random direction away from your village.
The idea was that this is such a noisy and chaotic system that you get decent enough randomization to distribute your hunter-gatherers into wilderness. The reason this is economical is it allows you to distribute away from concentrating too much effort into just one particular zone. A key idea here is that while maybe a certain river or patch of grass might have been historically lucrative, eventually the population will be exhausted. Without doing ample surveys of the wilderness all the time, hunter-gatherers have scarcely a clue that they are exhausting its resources.
So a simple strategy is to randomly distribute your collection of resources to allow areas to recover. At the same time, the model of relying on an oracle bone also makes it simpler than doing an exhaustive search of your "food-space" and if anything goes awry, it's because you've misinterpreted the divine spirits or you've done something wrong or bad. It's easy to cooperate if you all blame the same thing.
I think the article's tone is a bit on what I would wholeheartedly dismiss as outlandish numerology mysticism - were I to read this when I was just beginning to study mathematics. Now that I've read a bit more history, I have a better appreciation that humans by nature have always tried to impose order, symmetry and harmony upon the universe and have codified that language of understanding as poetry in all forms. I think the article is ambitious towards grasping at the scale and scope of mathematics not merely as a tool of human will, or a language science, or philosophy of philosophies, but something that is very palpably and tangibly human in form. There is a decent treatment towards the ambiguity of what an "algorithm" even is.
My overall impression of the article is that it is trying to describe something beautiful about how the process of mathematics has acted upon itself. Models such as cellular automata or recursive processes show how complexity arises from simple rules - order from chaos, chaos from order.
At the end of the day, I love to nitpick the treatment of technical subjects and their sort of phenomenal attributes in society - but it's posts like these on HackerNews that make me happy that anyone even bothers to write about this topic at all.
It’s interesting to consider how mathematics was viewed in ancient times, often interwoven with other disciplines and seen as the mystical language of nature rather than the sterile way it is often taught today. The legacy of Pythagorus is as much about reincarnation, music, and philosophy as it is about geometry. Numerology and geometry has played a huge role in religious practices from the I Ching to Gothic cathederals. Most divinitory practices are essentially algorithms for generating randomness, not only granting the benefits you mention but also allowing the guide to tap into their intuition and semi/unconscious mind, which may produce insights not as easily accessible via analytical methods.
25 comments
[ 3.4 ms ] story [ 21.0 ms ] thread"However, in importing these labels from Western philosophy to the classical Indian philosophical systems, one needs to exercise caution because the concepts of nature, science, scientific method, etc. do not smoothly converge in the two theoretical traditions."
Naturalism in Classical Indian Philosophy
https://plato.stanford.edu/entries/naturalism-india/
What people call “AI” is actually a long historical process of crystallizing collective behavior, personal data, and individual labor into privatized algorithms that are used for the automation of complex tasks: from driving to translation, from object recognition to music composition. Just as much as the machines of the industrial age grew out of experimentation, know-how, and the labor of skilled workers, engineers, and craftsmen, the statistical models of AI grow out of the data produced by collective intelligence.
I can totally resonate with that part
We almost always discount the work done by our predecessors: NIH-syndrome/"reinventing the wheel" vs "standing in the shoulders of giants", and this was a humble reminder for me.
"We" being computer scientists (or computing) which for some reason has this disease, perhaps due to the field's newness? When I worked in the life sciences (pharmaceutical chemistry) the literature was the first place we looked.
Whereas when working at a research lab in my 20s: I would often go down to our library and read CACM, IEEE journal et al and would sometimes implement things I read there. Even though I would brandish a photocopy of the journal article and reference it in the source code my boss and colleagues would praise me for "thinking up" a solution to a given problem.
Ironically the Internet changed things dramatically in the physical sciences (in Physics especially* but in other physical sciences by making it much easier to both search and read the literature) whereas in non-academic computing it's been the small conferences and, to my surprise, Youtube.
* Physics has long been an innovator in information dissemination; when journal publishing got too slow and letters too point-to-point it gave rise to Physics Letters, which itself became a pair of journals....
They put out a nice series of readers based on their journals and other material. I recently read a great one that collected some of the works of Hito Steyerl.
My projects are very local: Tech Worker Coalition in Berlin e Gambe.ro in Italy, so I don't think there's any room in those.
This is the best quote from the article.
It's clear that there is intense interest in deep learning and machine learning in the last 7 years or so, but, especially when talking of "AI", it is important to remember that the majority of work in AI has _not_ been on machine learning, but on inference and reasoning, and using symbolic, logic-based techniques (including Bayesian reasoning, as in Solomonoff's Inductive Inference) rather than "statistical" methods (in the sense that the term is used in AI to mean everything from probabilities to calculus).
Finally, there are historcial inaccuracies in this article. For example, Rosenblatt's Perceptron was not "the first machine-learning algorithm' as the article says: Arthur Samuel's self-playing checkers agents and Minsky's SNARC are two systems predating the Perceptron by a few years and research on machine learning techniques like Markov chains goes back to the 1910's.
Also he's not even a native speaker and I don't find his way writing very compelling. Quite the contrary.
Why do you think he's a charlatan?
I'm not sure what to do of the ideas discussed in the article, themselves. I am honestly confused about what he means when he's talking about "AI", given his loose use of terminology, so I'm not sure I understand the article clearly enough to agree or disagree with it.
To clarify, I'd be all for reading a strong critique of the modern (last decade) use of machine learning, including neural networks and deep learning (but not only). If he's talking about AI as a whole, however, I don't see how his critique can really work. I don't think I've ever heard of anyone using a classical planner or a constraint solver to drive an automated truck, say.
The start of the article, with the discussion of the Agnicayana ritual was very intersting.
This might help you to bridge your understanding of AI with how it's used elsewhere.
People will show you who they are if you let them.
Algorithm being defined as: take a bunch of animal bones, burn them until they are essentially cremated, and then interpret the ashes somehow to point you in a random direction away from your village.
The idea was that this is such a noisy and chaotic system that you get decent enough randomization to distribute your hunter-gatherers into wilderness. The reason this is economical is it allows you to distribute away from concentrating too much effort into just one particular zone. A key idea here is that while maybe a certain river or patch of grass might have been historically lucrative, eventually the population will be exhausted. Without doing ample surveys of the wilderness all the time, hunter-gatherers have scarcely a clue that they are exhausting its resources.
So a simple strategy is to randomly distribute your collection of resources to allow areas to recover. At the same time, the model of relying on an oracle bone also makes it simpler than doing an exhaustive search of your "food-space" and if anything goes awry, it's because you've misinterpreted the divine spirits or you've done something wrong or bad. It's easy to cooperate if you all blame the same thing.
I think the article's tone is a bit on what I would wholeheartedly dismiss as outlandish numerology mysticism - were I to read this when I was just beginning to study mathematics. Now that I've read a bit more history, I have a better appreciation that humans by nature have always tried to impose order, symmetry and harmony upon the universe and have codified that language of understanding as poetry in all forms. I think the article is ambitious towards grasping at the scale and scope of mathematics not merely as a tool of human will, or a language science, or philosophy of philosophies, but something that is very palpably and tangibly human in form. There is a decent treatment towards the ambiguity of what an "algorithm" even is.
My overall impression of the article is that it is trying to describe something beautiful about how the process of mathematics has acted upon itself. Models such as cellular automata or recursive processes show how complexity arises from simple rules - order from chaos, chaos from order.
At the end of the day, I love to nitpick the treatment of technical subjects and their sort of phenomenal attributes in society - but it's posts like these on HackerNews that make me happy that anyone even bothers to write about this topic at all.