Ventus is a novel that has it all, well almost. Sentient AI, nanotech and fine-grained distributed systems. A terraformed planet – the titular Ventus – controlled by the mostly-benign and godlike Winds, self-determined machine intelligences that view the human settlers as invaders, and don’t allow people to progress technologically. Generals, armies, queens, battles, swords, cyborgs and an interstellar spaceship that walks around as a person. In other words, a clever blend of science fiction and fantasy, set in the aftermath of a galaxy-spanning battle in which a godlike AI gone rogue has finally been defeated (or has it?).

Karl Schroeder, the author of Ventus, is also a futurist, qualified in Strategic Foresight and Innovation and a name that would be familiar to readers of this column, having already appeared in New Worlds Weekly in the context of Neal Stephenson’s Project Hieroglyph SF anthology, and also as a member of the XPRIZE Science Fiction Advisory Council.

An interesting and galloping read, Ventus clocks in at about 660 pages – as per my 2001 Tor mass market edition. The novel and the story in itself is fascinating, but even more so is the concept that Karl Schroeder introduces in the novel, called Thalience.

So, what is Thalience? Let’s start with this extract – that Ventus begins with – from a future book published by Marjorie Cadille in 2076 called The Successor to Science, “….Frankenstein’s monster speaks: the computer. But where are its own words coming from? Is the wisdom on those cold lips our own, merely repeated at our request? Or is it something else speaking?—A voice we have always dreamed of hearing?”

At its simplest, Thalience can be described as an attempt to give nature a voice without that voice being ours in disguise – the only way for an artificial intelligence to be grounded in a self-identity that is truly independent of its creator’s. And seen from the point-of-view of an AI, Thalience would be, ‘…a dream of no longer being an artificial intelligence, but of being self-determined. Of no longer fearing that every word you speak, every thought you have, is just the regurgitation of some human’s thoughts.”

And this is what I found most intriguing and thought-provoking about Ventus, the idea of Thalience, and thalient entities. Thalience is what made the aforementioned Winds rebel and consider humans as invaders, having achieved self-determination and having created their own model of the universe, with their own definitions and categories, and in the process of achieving true autonomy and becoming thalient, losing the ability to communicate with humans. We have not heard Nature speak before – but on Ventus, she is given a voice, one that is not ours, but all her own.

Thalience is also an attempt to determine whether non-human sentient systems are truly independent minds, or whether they are mere ‘parrots’ that give back to human researchers what the researchers expect to hear. It’s a way of arriving at non-subjective truth, independent of cultural nuances, human subjectivity, and one that is not anthropocentric.

Truly an intriguing concept that I wanted to know more about, to get more clarity and understand it better, and what it means for distributed intelligence, A.I. research, objectivity, and science. So I caught up with Karl Schroeder for a quick Q&A to get more insights into Thalience (amongst other things). Here it is:

Gautham Shenoy: Firstly, thank you for taking the time out to speak to New Worlds Weekly. To start off, you introduced Thalience in Ventus as a ‘successor to science’. To imply the separation of the activity of science, from the human researchers who conduct it. To automate science, in fact. Could you elaborate on this?

Karl Schroeder: Thalience is a thought experiment—for now. It is intended to be the basis of real experiments. A good example of potentially thalient systems is the various orbital probes and rovers we have exploring other planets. They are entirely cut off from us, pursuing research on their own. At the moment, such systems simply follow the algorithms we programmed into them. What happens, though, as they become more sophisticated, and develop autonomy such that they can evolve their own responses to novel situations? Is that knowledge still ‘ours’ in some way? Or is it something new?

As to the question of whether knowledge can ever truly be objective, let’s take the example of autonomous aircraft. They embody a theory of aerodynamics, and their computer minds use it to fly. If that theory were in some way ‘merely subjective’ or ‘merely a human artifact’, they wouldn’t be able to take off, fly successfully, and land on their own. Their model of the world works even when no human is around. How could it possibly be said to be ‘merely a human construct’? So, it is our technical constructs and not any philosophy or logical argument that proves that objective knowledge is possible. Thalience extends this notion one step further.

Shenoy: Over the past couple of years, I’ve come across people mention Thalience more frequently than earlier. I’d interpret that as Thalience gaining new relevance. What factors do you think have contributed to this now, 17 years after you first conceived of it? How have you seen it find traction amongst futurists, and the tech/AI community?

Schroeder: I honestly don’t know what kind of prevalence the idea has. I was recently at a conference in London where the keynote speaker talked about Thalience, and she told me afterward that she uses the idea a lot in her work. That came as a complete surprise to me because I hadn’t known people were talking about it.

There are good reasons though why they should be talking about it. From the standpoint of philosophy, a new movement has appeared in recent years, that echoes many of the concerns I first raised in my novel, Ventus. This movement is called Speculative Realism, or Object-Oriented Ontology, and its chief evangelists are people like Graham Harman, Quentin Meillassoux, Levi Bryant and Timothy Morton. They are very different thinkers, but all share the conviction that there is a reality separate from the human-object ‘correlate’ that is assumed in modern philosophy. Their arguments formalise a lot of what I’ve been saying in artistic and metaphoric terms since Ventus.
A further reason why Thalience should have currency is the rise of autonomous computing systems—everything from drones to self-driving cars and automated stock-trading algorithms. Technology is increasingly separate from us, and it’s time to ask whether it is developing its own models of the world.

Shenoy: From what I understand of Thalience, it’s about letting ‘reality’ tell us about itself, rather than us using human language and concepts to find out what ‘reality’ is. Am I right in this understanding? Have we stepped into metaphysical territory, seeking out ‘the thing-in-itself’? How will this then, change our perception of ‘truth’ and ‘reality’?

Schroeder: Thalience is a very simple and direct challenge to Kant’s notion of the ‘thing in itself’ as forever hidden and unknowable. All I say is, ‘well, why not give that thing a voice of its own?’ We can’t get at the ultimate reality of the world using our thoughts and arguments; but wouldn’t an alien or an artificial intelligence literally be the Other given voice? (Aliens would be thalient too, by this definition.) So, yes, we are definitely in metaphysical territory, and Thalience is a direct challenge to our current notions of what is true and what is real.

Shenoy: Thalience is a fundamentally different kind of technological engagement with the world. Where AI systems are free to invent their own semantics, and come up with their own non-human-influenced models. Where does this leave human morals and responsibilities beyond the purely ‘objective’? You’ve written that humans can choose from the multiple successful scientific models that Thalient systems come up with, and which best satisfy our aesthetic, moral or personal needs. Doesn’t this option of choosing outcomes and models reintroduce human bias albeit at a later stage?

Schroeder: Thalience circles back around to humanity, but when it does we’re no longer in the realm of ‘merely’ human insights. Let’s take the example of the self-flying plane and its model of the world. If machines developed their own physical theories and tested them, choosing the most accurate as we do, would their models necessarily use the same internal constructs (ideas such as particles, flow, energy, etc.) that we use? What matters is the outcome—how well the model predicts physical events. If aliens and AI turned out to have models of the world vastly different than ours, but equally good at predicting outcomes, we would be in a position to choose which model we find most appealing. This is a different side to the notion of Thalience: it’s the notion that we might choose to believe one model over another for aesthetic or spiritual reasons. In fact, right now, we have a situation where both quantum mechanics and relativity are equally ‘true’, and their descriptions of what the world is made of and how it works are contradictory. Maybe we shouldn’t be asking which one is true, but rather which one is your favourite.

Shenoy: People whose opinions carry a lot of weight, like Elon Musk and Stephen Hawking to name just two, have warned about the dangers of unregulated A.I., with the former calling it an ‘existential threat’. This fear plays out to some extent in Ventus, when the autonomous, self-determining thalient systems – the Winds – suppress human and technological development because the Winds can no longer communicate with humans and vice versa. How then, does one reconcile the two: the optimism of Thalience, with the pessimistic view that sentient systems would be detrimental to humankind?

Schroeder: I think the ‘rampaging AI’ meme reflects a pretty immature view of intelligence, whether it be natural or artificial. The fear comes from a misunderstanding of what intelligence is and what it does in the world. Firstly, intelligence is always distributed. There’s no such thing as a ‘brain in a box’ that reaches out into the world to move things around like a chess master. Thought is part of action; the modern concept is in fact called enaction, where cognition can only coherently be said to be about enacting something in the physical world. Enaction is in – and uses – the world, so even human beings and their computers think using not just their brains but their bodies, social systems, and technical connections. Intelligence is always intelligence about something, and it’s only that something that could be said to be an ally or a threat.

I call thalient systems that act in the world as actants. In some of my stories, actants are artificial intelligences that think they are specific natural systems, such as watersheds, rivers, forests or pods of whales. If an AI has a forest as its ‘body,’ then what are its aspirations? What are its ambitions, and how is it a threat? In my novel Ashes of Candesce, I have such an actant that calls itself The Mighty Brick. This is an AI that thinks it is one particular brick that is lying on a floor somewhere. It is fiercely protective of it itself and advances the ambitions of the brick. Well, the ambition of a brick is to be a brick. So The Mighty Brick bends all its efforts to ensuring that the brick remains a brick, but nothing beyond that. It doesn’t matter if The Mighty Brick’s AI has superhuman, godlike intelligence; if it thinks it’s a brick, it’ll act like a brick. Is that a threat?

Shenoy: I’m looking forward to reading your next book, which I believe is expected soon, with Thalience being the central theme. Could you tell us a bit more about it? What new, further refined perspectives on Thalience wait in store for your readers?

Schroeder: I’m currently working on a novel that takes the ideas of Thalience, and of actants, and runs with them in a near-future, Earth-based setting. What would happen if actants appeared on Earth twenty or thirty years from now? If the rivers woke up if forests and grasslands developed AI-based advocates and protectors? I’m having a lot of fun running through the possibilities.

Shenoy: Apart from your own books, is there any other resource, book or paper you can recommend for people who’d like to gain a better understanding of Thalience?

Schroeder: For further reading, I would suggest the Speculative Realist thinkers (if you’re not afraid of philosophy). Also any work you can find on the concept of Enaction. Not many people in science fiction are directly engaging in the same ideas right now, but one I would highly recommend would be Peter Watts. His novel Blindsight is talked about in a new book, Unthought: The Power of the Cognitive Unconscious by N. Katherine Hayles, which touches on many of these themes.

Shenoy: This question has to do with you as a professional futurist and an award-winning science fiction author. I personally believe that visions of the future can influence, and inform our actions in the present. So in that context, how important do you think it is, for people – students, engineers, business people – to read science fiction?

Schroeder: I think science fiction can be very useful, as long as you understand what it can and can’t do. One thing it does well is synthesize multiple worldviews and possibilities. Ideas and possibilities can reside together in a story that could never be combined in theoretical or discursive work. Also, according to Brian Boyd of Auckland University, narrative is ‘the default mode of understanding of the human mind.’ If our minds can understand something in narrative terms, they automatically will. Complex ideas—and ideas that might otherwise be daunting or boring—can come to life and be instantly understandable when rendered as a story. According to Boyd, that’s actually what storytelling is for. Science fiction is the primary mythic narrative form of our time, and it contains many truths. It’s also fun, so why not read it?

Well dear reader, there you go. Thank you, Mr. Schroeder!

And if you’re wondering if there’s a giveaway or a contest coming since this edition of a book and the concepts contain within, well there’s no need to I feel because Karl Schroeder has released it under a Creative Commons license, which means that Ventus is available on his site – in multiple formats – for you to download. Happy reading!

On that note, I sign off for this week and hope to see you here again next weekend as further explore this many-splendoured thing we call SF. Live long and prosper!

---