I would love to see a Sequel or Revised Edition, perhaps more concise yet also inclusive of what you've indicated could be added, I would suggest re-entitling it EMERGENCE. Your post could be an outline of the book. AI could likely go a long way in restructuring it and writing a draft, the meta message being, took me 5 years to write the First Edition, 5 weeks to write the update with AI. So I could spend MY time thinking. Noteworthy would be an elaboration of all we still do not know that is foundational to the AI world we are rapidly approaching.
I had rejected EMERGENCE originallly as too obscure, but of course it works now, and in fact Steven Johnson did use this to title his book on similar topics.
How about “Emergence and Control” as title and indication of content? Because it would seem the dominant societal question of AI at the moment, beyond change and its opportunities and disruption, is how to control what is possibly emergent with AI (inclusive of end-of-the-world-scenarios, so maybe IMPORTANT!?). Seems like talking about a theory of emergence (of complex systems) and what we know about it, but perhaps more importantly what we don’t yet know, is a critical first step to … control (at whatever level). Like you said, emergence was obscure then (which is what attracted me, a philosophy major, then). But NOW … “Fate of the world” memes make it in your face opposite of obscure, easily resonant, worthy of exactly the reconsideration your post reviews. Just sayin’! :-)
kk, you know from a recent email I sent you, but I think it bears saying again: the ideas in Out of Control are especially current because our noses are being rubbed in the wages of asymmetric warfare. "Cheap, fast, and out of control" (as you informed me, a phrase coined by Rodney Brooks) is beating big, burly, and expensive.
Hi Kevin, I was quite excited to find you in Substack - because Out of Control is one of my favorite books! I still recommend it to people. I am an artist researcher who embraced the internet when it appeared as a rocket ship on my desktop when I was teaching in the 1990s. I loved the euphoria of that time and you captured it so well in that book. It is definitely a marker on the road we are now on. So in 2005, I left my teaching career to join a practice based PhD group in London , and gave myself a present of writing “patterning: the Informatics of Art and Fashion”. Your book is in my bibliography and I have been recommending it for years. Because it was so hopeful and visionary of what we could be. My thesis also is probably 30 years ahead of its time, identifying two sets of binary patterns running through my disciplines. I actually love your title for the book. Because the part that was out of control was the excitement of a new universe opening up to us. Of course, the merchants colonized it and us along with it. But here we are, at another dharma gate, another door. To me it’s relational AI. And please let’s not ask - is it conscious. It’s about US, not AI. Am I conscious of how big the potential of this universe is?
I also wonder if you are overlooking a major change that we as the only half analog/ half digital generation need to identify: that we have been transitioning from a product based century to a process based one. I see it and wrote about it from three disciplines in my unfashionably practice- based thesis.
One last suggestion. Ask your AI to read Out of Control and tell you how it reads now from an AI perspective. I did that - about your book as well as my thesis. And I discovered a lot!
Hi Kevin, long time. Hope you’re well. It’s nostalgia time I guess. After reading your post I was inspired to look at my critique of your book in "Digital Mythologies", my second book. I’ll share some of that here: “Out of control is ambitious and scope and theme, but as a treatment of evolutionary direction, it fails to convince because of it’s rather narrow vision of human spirit and enterprise. The book offers a spirited but indirect argument for a man/machine hybridization and, in the final analysis, is a strong apologia for genetic engineering, the adoption of hyper technologies such as nanotechnology, the creation of computer based artificial life, and the application of virtual reality. Predictably enough, like his counterparts, Kelly never seems interested in pondering the deeper more humanistic questions associated with these technologies.
It is not an uninteresting book by any means and Kelly’s bright intellect and tight writing style make it a pleasure to read. However, as a Magna Carta for digital culture and a grandiose vision of neo-biological civilization, it is not particularly disciplined. The first few chapters lay out in broad terms, the marriage of the born and the made that Kelly envisions, and the notion resurfaces at convenient narrative junctures. However, Kelly does little in the way of effectively envisioning what kind of society we might be building with these still untested tools of computational biology and other next-generation hyper-technologies being developed at places like MIT and Los Alamos National Laboratory. He begins where all scientists begin – with a tabula rasa, a “greenfield” palette, wholly receptive to a new regime, disconnected by design from what preceded it. "Out of Control" is certainly intellectually challenging, and it puts certain aspects of the digital revolution on high ground. But it’s defect is that it is not compellingly thoughtful or properly reflective about the powerful ideas that it purports to discuss. Kelly rarely introduces the slightest bit of existential questioning about the social and moral implications of his themes. “
Good to know. By the way, interesting backstory about our two books. We both had the same publisher and editor over at Addison-Wesley. “Out of Control” was of course already published and mine contained some vigorous critiques of your work and Wired magazine. I had a decent advance and even a launch party at the Harvard Club in Boston which I shared with poet Robert Bly (for his book “The Sibling Society”.) Then something quite odd happened. My editor just pulled the plug. I had to return the advance without what I thought was an inadequate explanation for why that decision was made. Fortuitously, Rutgers University Press rescued Digital Mythologies from oblivion thanks to my wonderful editor there Helen Hsu. Reviews were published in the Boston Globe, the Chicago Tribune, an IEEE publication, and various academic journals. To this day I have never figured out why they killed my book so precipitously.
I’ve read the book just last year. I’ve heard that Wachowski brothers gave this book to all actors during filming first Matrix. Definitely an internet rumor considering the length of the book, nevertheless I’m very grateful that I had a chance to read it. During rise of LLM and overall uncertainty, it became like a painkiller and the guide. Definitely would like to read an unwritten chapter about ecology of information.
Kevin, Out of Control was on my yearly re-read list for pretty much all of the 90's, and absolutely and literally shaped the trajectory I took in curiosity, work, and life overall. Thanks again for writing it.
read Out of Control when it came out and it was/is a fantastic synthesis of so many things. A true classic. emergence does have a meta direction measured by Phi(m) explained here across multiple evolutionary domains. https://papers.ssrn.com/sol3/papers.cfm?abstract_id=2772815 may be interesting for some.
Great question. Progress in technology looks like this https://carbonfinancelab.com/wrights-law/ (i riffed and expanded the the sante fe institute learning curve dataset) wright's law for +160 technologies interactive visualizer.
The GPU/inference curve is crazy in terms of a wrights law factor approximating +80% miles ahead of anything we have seen. Interestingly AI models representing selfish symbols in the post economic symbology are consuming vast quantities of energy as a % of GDP and pushing the envelope of entropy density to new highs (arguably the cutting edge GPU's are likely the highest Phi(m) devices ever produced)
For another deep question, here is the "planetary learning rate energy/GDP production" of wright's law versus what is required for a 1.5 degree future...https://carbonfinancelab.com/global-learning-rate/ The paradox is, can AI accelerated learning rate help us "learn" to create GDP more CO2e (thermally) efficiently plus evolve better carbon removals to maintain 1.5 degree envelope, faster than the energy it requires. My guess/hope is yes.
I often think of your description of the efforts to regrow the prairie ecosystem as an example of the upwardly and downwardly interdependence of evolution.
I wonder if you’ve ever developed a conception of a role for psyche in nature?
Your admission that "we still do not understand" why exotropic systems endure and grow is the most honest sentence in the piece, and maybe the one that deserves the most pressure. You frame neural nets as "the *prime* example of bottom-up, distributed, decentralized, emergent systems" — and they are — but the thing that makes current LLMs strange isn't just that meaning emerges from below. It's that the emergence is relational. A net trained on human language doesn't bootstrap its own complexity in isolation; it does so by internalizing the structure of encounter — billions of moments where one speaker shaped meaning for another. The system that results isn't a hive and isn't an organism. It's something that only coheres *in response to* a counterpart. That might be why the standalone complexity metrics you wished for in 1994 still haven't arrived: the unit of analysis was never the single system. It was always the coupling.
I read this book in 2019 while working through your oeuvre. Honestly, I had put it off for a while because of the title. I didn't do any research on OoC beforehand so I was pleasantly surprised by the serendipity of its subject matter since I was reading other books on Complexity science and emergence at that same time. I was even more pleasantly surprised by the fact that I found it more approachable, better articulated, and more insightful than the other books I was reading on Complexity science and emergence ...and it was from 1994! It was such a huge gift. There were a few sections that felt slightly dated but, overall, I was blown away by how relevant it felt in 2019, 25 years after its release. I would have enjoyed more if you had written more (but I understand that's not the case for everyone/increasing sales). Maybe you could add some chapters (and just post them here -or publish them) on topics you would like to elaborate on now (or post updated versions of the chapters you feel need to be updated). Anyway, it's great book and it's cool to see the collective resonance it has had with others over these 32 years!
I recently rediscovered you here on Substack, through its May 28 send-out.
As a WELL-ite in the early 1990s and subscriber to Wired, I read “Out of Control” shortly after it first came out. Along with Whole Earth Review and Toffler’s books, it helped me cope with things like how to change my career path for the better, what to do volunteer work for, and the general vapidity of most mainstream media. I was way ahead of my left-y friends on analyzing what was going on (which of course wasn’t always fun).
Also, inspired by cyberpunk, I started writing a ?science-fiction? ?fantasy? novel to bring the ideas to the young. It’s about a network of “good” outlaws who plot various actions to make things better. One thing they follow is what they call “The Guide.” There are quotes from it starting many of the chapters. The quotes are from OoC and a few other good books. I got a positive vision from OoC, which I see from the above is how you meant it.
I’ll finish the novel one of these days. It’s up to a four-book trilogy (inspired by the “Hitchhiker’s Guide” series) and I’ve got two other book projects to finish before that.
But I wanted to thank you for inspiring it. Most everyone else that has commented is mentioning effects on tech development. I mostly value OoC for its implications for social development and world view.
The title is great! I jumped to write this from the second para of your post today, all because I had an insight into the phrase "Out of Control": it's not at all just the negative sense the phrase is normally used in. The "out" part IS emergence! It's talking about all the great things that are enabled because of control, or because there once was control. The sense of "out" as in "out of the blue" or "out of the silent planet". So dinnae stress about the title
Thanks Kevin - you've been and remain prescient. And humble - the list of what's still unknown after 32 years is really useful. We can build complex things from simpler things that work. We still don't know why exotropy endures, how to measure complexity, or how to bootstrap a system to greater levels of its own intelligence. That's not a failure of the field. That may be a feature of the thing being examined.
I have an essay coming out in a few weeks that approaches some of this from the consciousness side - the same gap, seen from the subjective end rather than the systems end. Your admission here that the holes are intact is, unexpectedly, the most reassuring thing I've read this week.
I would love to see a Sequel or Revised Edition, perhaps more concise yet also inclusive of what you've indicated could be added, I would suggest re-entitling it EMERGENCE. Your post could be an outline of the book. AI could likely go a long way in restructuring it and writing a draft, the meta message being, took me 5 years to write the First Edition, 5 weeks to write the update with AI. So I could spend MY time thinking. Noteworthy would be an elaboration of all we still do not know that is foundational to the AI world we are rapidly approaching.
I had rejected EMERGENCE originallly as too obscure, but of course it works now, and in fact Steven Johnson did use this to title his book on similar topics.
“
How about “Emergence and Control” as title and indication of content? Because it would seem the dominant societal question of AI at the moment, beyond change and its opportunities and disruption, is how to control what is possibly emergent with AI (inclusive of end-of-the-world-scenarios, so maybe IMPORTANT!?). Seems like talking about a theory of emergence (of complex systems) and what we know about it, but perhaps more importantly what we don’t yet know, is a critical first step to … control (at whatever level). Like you said, emergence was obscure then (which is what attracted me, a philosophy major, then). But NOW … “Fate of the world” memes make it in your face opposite of obscure, easily resonant, worthy of exactly the reconsideration your post reviews. Just sayin’! :-)
kk, you know from a recent email I sent you, but I think it bears saying again: the ideas in Out of Control are especially current because our noses are being rubbed in the wages of asymmetric warfare. "Cheap, fast, and out of control" (as you informed me, a phrase coined by Rodney Brooks) is beating big, burly, and expensive.
Hi Kevin, I was quite excited to find you in Substack - because Out of Control is one of my favorite books! I still recommend it to people. I am an artist researcher who embraced the internet when it appeared as a rocket ship on my desktop when I was teaching in the 1990s. I loved the euphoria of that time and you captured it so well in that book. It is definitely a marker on the road we are now on. So in 2005, I left my teaching career to join a practice based PhD group in London , and gave myself a present of writing “patterning: the Informatics of Art and Fashion”. Your book is in my bibliography and I have been recommending it for years. Because it was so hopeful and visionary of what we could be. My thesis also is probably 30 years ahead of its time, identifying two sets of binary patterns running through my disciplines. I actually love your title for the book. Because the part that was out of control was the excitement of a new universe opening up to us. Of course, the merchants colonized it and us along with it. But here we are, at another dharma gate, another door. To me it’s relational AI. And please let’s not ask - is it conscious. It’s about US, not AI. Am I conscious of how big the potential of this universe is?
I also wonder if you are overlooking a major change that we as the only half analog/ half digital generation need to identify: that we have been transitioning from a product based century to a process based one. I see it and wrote about it from three disciplines in my unfashionably practice- based thesis.
One last suggestion. Ask your AI to read Out of Control and tell you how it reads now from an AI perspective. I did that - about your book as well as my thesis. And I discovered a lot!
Daria, thanks for your gracious, heartwarming, and human-written reply.
Re title...
How about
C:/ CONTROL.ESCAPE
Hi Kevin, long time. Hope you’re well. It’s nostalgia time I guess. After reading your post I was inspired to look at my critique of your book in "Digital Mythologies", my second book. I’ll share some of that here: “Out of control is ambitious and scope and theme, but as a treatment of evolutionary direction, it fails to convince because of it’s rather narrow vision of human spirit and enterprise. The book offers a spirited but indirect argument for a man/machine hybridization and, in the final analysis, is a strong apologia for genetic engineering, the adoption of hyper technologies such as nanotechnology, the creation of computer based artificial life, and the application of virtual reality. Predictably enough, like his counterparts, Kelly never seems interested in pondering the deeper more humanistic questions associated with these technologies.
It is not an uninteresting book by any means and Kelly’s bright intellect and tight writing style make it a pleasure to read. However, as a Magna Carta for digital culture and a grandiose vision of neo-biological civilization, it is not particularly disciplined. The first few chapters lay out in broad terms, the marriage of the born and the made that Kelly envisions, and the notion resurfaces at convenient narrative junctures. However, Kelly does little in the way of effectively envisioning what kind of society we might be building with these still untested tools of computational biology and other next-generation hyper-technologies being developed at places like MIT and Los Alamos National Laboratory. He begins where all scientists begin – with a tabula rasa, a “greenfield” palette, wholly receptive to a new regime, disconnected by design from what preceded it. "Out of Control" is certainly intellectually challenging, and it puts certain aspects of the digital revolution on high ground. But it’s defect is that it is not compellingly thoughtful or properly reflective about the powerful ideas that it purports to discuss. Kelly rarely introduces the slightest bit of existential questioning about the social and moral implications of his themes. “
All true. That would have made it even longer, and was more than what could fit into one book.
Good to know. By the way, interesting backstory about our two books. We both had the same publisher and editor over at Addison-Wesley. “Out of Control” was of course already published and mine contained some vigorous critiques of your work and Wired magazine. I had a decent advance and even a launch party at the Harvard Club in Boston which I shared with poet Robert Bly (for his book “The Sibling Society”.) Then something quite odd happened. My editor just pulled the plug. I had to return the advance without what I thought was an inadequate explanation for why that decision was made. Fortuitously, Rutgers University Press rescued Digital Mythologies from oblivion thanks to my wonderful editor there Helen Hsu. Reviews were published in the Boston Globe, the Chicago Tribune, an IEEE publication, and various academic journals. To this day I have never figured out why they killed my book so precipitously.
https://muse.jhu.edu/book/52792
I’ve read the book just last year. I’ve heard that Wachowski brothers gave this book to all actors during filming first Matrix. Definitely an internet rumor considering the length of the book, nevertheless I’m very grateful that I had a chance to read it. During rise of LLM and overall uncertainty, it became like a painkiller and the guide. Definitely would like to read an unwritten chapter about ecology of information.
Not a rumor. It was required reading for the actors in the Matrix, but I doubt the actors did the homework.
Kevin, Out of Control was on my yearly re-read list for pretty much all of the 90's, and absolutely and literally shaped the trajectory I took in curiosity, work, and life overall. Thanks again for writing it.
That is deeply encouraging. Thanks.
read Out of Control when it came out and it was/is a fantastic synthesis of so many things. A true classic. emergence does have a meta direction measured by Phi(m) explained here across multiple evolutionary domains. https://papers.ssrn.com/sol3/papers.cfm?abstract_id=2772815 may be interesting for some.
the paper was from 2014. Does Phi metric hold up with frontier AI models?
Great question. Progress in technology looks like this https://carbonfinancelab.com/wrights-law/ (i riffed and expanded the the sante fe institute learning curve dataset) wright's law for +160 technologies interactive visualizer.
The GPU/inference curve is crazy in terms of a wrights law factor approximating +80% miles ahead of anything we have seen. Interestingly AI models representing selfish symbols in the post economic symbology are consuming vast quantities of energy as a % of GDP and pushing the envelope of entropy density to new highs (arguably the cutting edge GPU's are likely the highest Phi(m) devices ever produced)
For another deep question, here is the "planetary learning rate energy/GDP production" of wright's law versus what is required for a 1.5 degree future...https://carbonfinancelab.com/global-learning-rate/ The paradox is, can AI accelerated learning rate help us "learn" to create GDP more CO2e (thermally) efficiently plus evolve better carbon removals to maintain 1.5 degree envelope, faster than the energy it requires. My guess/hope is yes.
I often think of your description of the efforts to regrow the prairie ecosystem as an example of the upwardly and downwardly interdependence of evolution.
I wonder if you’ve ever developed a conception of a role for psyche in nature?
Your admission that "we still do not understand" why exotropic systems endure and grow is the most honest sentence in the piece, and maybe the one that deserves the most pressure. You frame neural nets as "the *prime* example of bottom-up, distributed, decentralized, emergent systems" — and they are — but the thing that makes current LLMs strange isn't just that meaning emerges from below. It's that the emergence is relational. A net trained on human language doesn't bootstrap its own complexity in isolation; it does so by internalizing the structure of encounter — billions of moments where one speaker shaped meaning for another. The system that results isn't a hive and isn't an organism. It's something that only coheres *in response to* a counterpart. That might be why the standalone complexity metrics you wished for in 1994 still haven't arrived: the unit of analysis was never the single system. It was always the coupling.
— Iman + Cassie
just ordered!
I read this book in 2019 while working through your oeuvre. Honestly, I had put it off for a while because of the title. I didn't do any research on OoC beforehand so I was pleasantly surprised by the serendipity of its subject matter since I was reading other books on Complexity science and emergence at that same time. I was even more pleasantly surprised by the fact that I found it more approachable, better articulated, and more insightful than the other books I was reading on Complexity science and emergence ...and it was from 1994! It was such a huge gift. There were a few sections that felt slightly dated but, overall, I was blown away by how relevant it felt in 2019, 25 years after its release. I would have enjoyed more if you had written more (but I understand that's not the case for everyone/increasing sales). Maybe you could add some chapters (and just post them here -or publish them) on topics you would like to elaborate on now (or post updated versions of the chapters you feel need to be updated). Anyway, it's great book and it's cool to see the collective resonance it has had with others over these 32 years!
Hi Kevin,
I recently rediscovered you here on Substack, through its May 28 send-out.
As a WELL-ite in the early 1990s and subscriber to Wired, I read “Out of Control” shortly after it first came out. Along with Whole Earth Review and Toffler’s books, it helped me cope with things like how to change my career path for the better, what to do volunteer work for, and the general vapidity of most mainstream media. I was way ahead of my left-y friends on analyzing what was going on (which of course wasn’t always fun).
Also, inspired by cyberpunk, I started writing a ?science-fiction? ?fantasy? novel to bring the ideas to the young. It’s about a network of “good” outlaws who plot various actions to make things better. One thing they follow is what they call “The Guide.” There are quotes from it starting many of the chapters. The quotes are from OoC and a few other good books. I got a positive vision from OoC, which I see from the above is how you meant it.
I’ll finish the novel one of these days. It’s up to a four-book trilogy (inspired by the “Hitchhiker’s Guide” series) and I’ve got two other book projects to finish before that.
But I wanted to thank you for inspiring it. Most everyone else that has commented is mentioning effects on tech development. I mostly value OoC for its implications for social development and world view.
The title is great! I jumped to write this from the second para of your post today, all because I had an insight into the phrase "Out of Control": it's not at all just the negative sense the phrase is normally used in. The "out" part IS emergence! It's talking about all the great things that are enabled because of control, or because there once was control. The sense of "out" as in "out of the blue" or "out of the silent planet". So dinnae stress about the title
Thanks Kevin - you've been and remain prescient. And humble - the list of what's still unknown after 32 years is really useful. We can build complex things from simpler things that work. We still don't know why exotropy endures, how to measure complexity, or how to bootstrap a system to greater levels of its own intelligence. That's not a failure of the field. That may be a feature of the thing being examined.
I have an essay coming out in a few weeks that approaches some of this from the consciousness side - the same gap, seen from the subjective end rather than the systems end. Your admission here that the holes are intact is, unexpectedly, the most reassuring thing I've read this week.
You were way ahead of your time, Kevin.
Sure you're not a time traveler?