Treasury Notes

This note is to introduce Apple Vision to Treasury partners, creators, builders, and investors, and to anyone else curious.

Treasury is a registry and discovery system for the world’s most valuable spatial assets - architectural design, film scenes and sets, real estate, world monuments, experiential art, nature scans, and other spatial content.

Apple Vision is fundamentally a VR device - it offers a fully-rendered 3D digital environment, in an enclosed face-mounted device, as so many before. But it encompasses AR by taking passthrough to its natural conclusion - pure high-resolution rendering of the outside world, but with rendered elements inside the view, as needed.

The extreme breakthroughs lie in the sophistication and quality of the visual experience, offering an astonishing naturalness of perception. Focus, stereoscopy, depth, resolution, spatial fixation are all so far advanced that all the conventional, lingering artefacts of VR are gone.

On top of this are profoundly subtle tracking capacities - eye, arm, hand, fingers - empowering the body as an input device, in a way that will mark the beginning end of the qwertykey and mousepointer paradigm and era. Apple Vision is a massive, overwhelming advance from the Oculus platform, or at least from the Horizon interface and experiential setup.

Oculus/Horizon interface is clunky, intrusive, overengineered - with a neurotic focus on safety that itself makes the experience feel unsafe, and bloated with integrations to Facebook features that no-one wants or uses. And yet underengineered - you still can’t do very much, and the hand controllers are as limiting and confusing as they are liberating.

What's worse: Oculus/Horizon has the most hideous default environments. Absurd landing spaces that pastiche dolls-house and night-club vibes with cartoon-quality rendering. Who wants to spend any time here? What is anyone supposed to do in such a demoralizing environment? Who was this designed for? Who was this designed by - will anyone put their name to it?

Oculus is a Swatch. Serviceable and interesting, and getting cheaper. But boring and disposable. Apple Vision is a Swiss Watch. Supremely engineered, utterly fascinating in its design, winning with subtle detail and aesthetics in every nuance. Exciting and memorable.

Three experiences alone Apple Vision already justify its current premium cost entirely: environments, simulations, film content.

Environments are the equivalent of desktops in flatscreen computing. And they, immediately, convert the onload experience from one of boredom to one of profundity. Apple Vision has detailed 3D scans of a various natural environments with contextual sound, weather and daylight.

This isn’t even an app, and yet it’s already a deeply compelling portal to another world. It showcases the sheer, visceral power of well-conceived digital spatial environments.

Supplying digital spatial environments - spatial assets - at vast scale is the core value offering of Treasury, offered as licensed content from the world’s best spatial creators, in order to massively accelerate the growth of high-value spatial computing applications and experiences. There is no going back.

Once they have seen a single default Apple Vision environment, users will demand more, and more sophisticated, spatial assets as the bare minimum startup experience for any spatial experiences, and thus will see Oculus/Horizon startup experiences, and most of the environments of available games and apps there, to be absurd, embarrassing, annoying.

To simply reinforce this point, the arm tracking and cutout passthrough rendering of the users arms, not even an announced feature of the Apple Vision, is an immense step towards full-body immersivity in digital spaces, and yet another demerit for Oculus, for all its fixation on cartoon avatars.

The simulation app, JigSpace, proves, immediately, that enterprise VR is finally ready for multiple markets. The utter, micrometer-level refinement of resolution fidelity in the rendered engine parts, testable inches away from ones face, the immaculate reflections from truly contextual light rendering, speed and naturalness of object interaction - zoom, rotate, explode - mean that countless engineering and medical education and device organizations will immediately beginning seeking training apps at scale.

This is not just because the quality of the rendered experience is so high, it’s because the quality of the passthrough environment is so clear that these devices can work in classroom and shopfloor group settings out of the box. The military will be a massive consumer of these, for countless applications.

The film applications for TV and movies, and the other flatscreen analogues such as extension monitors, are spectacular, and silence any doubt that movie-length watch sessions and multi-hour work sessions are viable. They are also the easiest justification for the full-freight price of this developer-oriented model by media consumers and powerusers of flatscreens.

You would expect to pay more than 3500 USD for a 85” high-resolution TV monitor, or for a 32” super-high resolution workscreen. You wouldn’t ever expect to own your own movie theatre. So how much would you pay for all of those, plus countless other screen form-factors, and workroom/screening-room controls? What kind of airline booking class would you have to be in to watch a movie on a full-enclosed screening-room on an movie-sized screen?

This airplane use-case alone - being on a plane working on a large screen without worrying if the seat in front will crush your laptop, or enjoying large-format movies in high-resolution instead of recut square-format movies squeezed out of the low-resolution seatback screens - will drive much adoption of the first-gen model, by regular travelers.

Apple Vision is a technical miracle. All the tech reviewers are jaded and wrong. With their nitpicking and faint praise they reveal the flaw in their business model - they have to have something to say, even if it is superfluous, and can’t be too enthusiastic, because they’ll look like they are captured by corporations.

When a miracle occurs, just acknowledge it, and appreciate in silence for a moment the passing of what was, and the enormity of what will come.

Apple Vision will immediately work, with developers, power-users, and enthusiasts driving adoption and excavating use-cases and user communities. Through gradual feature upgrades, price drops, form-factor evolution, incremental context-enablement (outside, water, dark, driving, sleeping), mass adoption is inevitable.

“Any sufficiently advanced technology is indistinguishable from magic” ... has been a tired trope, worn out by earnest computer folks who wear ties and read paper books, foretelling a mythical future time. That time has come, and no-one will use this phrase any more. Because technology and have magic fused. Apple Vision, and thus all the spatial computing that will now come, is not like magic. It is magic. The magic is now the reality.

Apple Vision exists, as a surface reason, because Apple exists. This is to say that it exists because Apple wants it to exists - a large enough reason, given the costs of developing it and risks of building the market for it. But it's also to say it exists because Apple can make it exist, they have the skills to do so: in chip design, OS build, hardware, peripherals, interfaces, design style, narrative, marketing and more.

It’s absolutely wrong and deeply lazy to claim that Apple is merely a marketing company, and iphones/macs are just overpriced androids/pcs, at least in respect of this product. Apple Vision is maybe not five years ahead of Oculus or Varjo, as the iphone was against competing smartphones in 2007, but it is some years ahead. Apple did this, and it must be acknowledged. Their leadership in tech will be restored, even among doubters, due the extreme quality in all particulars of this device. Tim Cook will never be underestimated or forgotten.

But the more interesting reason for Apple Vision than Apple, is that the technology itself exists. Or rather many spatial technologies exist, and are converging all at the same time so monumentally in spatial computing. These include advances in:

• human interface technologies eye-tracking, body-tracking, stereoscopy, foveated rendering, and more
• real-time rendering unity/reality composer, unreal, and more
• static reality capture and real-time reality capture photogrammetry, lidar, radar, nerfs, guassian splats, MRI
• display devices micro-OLEDs, specialized graphics processors, lenses, manufacturing techniques, glass and metal materials breakthroughs, and more

on top of the existing extreme advancement in:

• general computation compute, storage, networking, data, distributed systems, identity and cryptography, and various strands of AI
• mapping, earthsensing and positioning

While is true, as so often, that Apple has put itself ahead by simply using what is available with more rigor and sophistication than others, these spatial techs are what make spatial computation possible, and they are fundamentally available to everyone, and an increasingly commodified cost.

These technologies are not just maturing, they are converging on a spatial computing opportunity that doesn’t terminate in a specific product pathway or application: instead, they represent a new infrastructure layer on top of internet-enabled society and economy, which will be as opportunity-making as the internet itself.

Treasury has developed a Spatial Computation Thesis as an introduction to this new technical era, to the Treasury spatial asset syndication opportunity, and to how spatial computing is an infrastructural phenomenon more than any production or application endpoint. So what does this new infrastructural layer of spatial tech enable in terms of a spatial computing paradigm? In its broadest outlines, how will it work or not work? Why are spatial assets so central to it?

Simply put, the extreme capacities that are enabled by this convergence represent the arising of a kind of supertechnology era - in parallel with other, cumulative, supertechnologies. In this era, the combined advancement of multiple strands of technical evolution makes for tools and applications that were previously, even recently, just inconceivable - fantastical, magical - from the perspective of any previous era of tech.

And yet: this spatial technology convergence, this supertechnology, is not even the most profound technical reason for the existence of Apple Vision. There is a supertechnology available to human kind that is only gradually, thousands of years after research began, now beginning to be unlocked. It is a hypertechnology, deep in the realm of mystery - and it is called cognition.

Apple Vision ultimately exists because whatever it visually represents is something that, on some level, we believe to be *real*. When you see immaculate bubbles and mystical illuminated petal spheres in this device, you are inclined to believe they really are there. This is not suspension of disbelief, this is inability to disbelieve. The definition of reality after all is, in fact, that which we are unable, on some level, to disbelieve. All this expansion of reality, all this fusion of reality and magic, is happening due to the hypertechnology of cognition. Not a line of code required.

However: the advance in cognitive neuroscience hasn’t made a glancing scratch the impregnability of the mystery of human cognition. The hard problem of cognition - what is it? - remains unyielding so far, even as research in the soft problem - how does it work? - is showing increasing signs of progress. And the conventional way to deal with this difficulty has been for scientists to ... simply ignore it. The hard problem of cognition does not exist, we are told: cognition is a figment of our imagination.

Even if this was not absurd enough on its face - if cognition is merely imagined, how do we understand the ... imagination in which it arises?, ad infinitum - it overlooks the cosmic fact that all digital technology is based on serving human cognition - for processing information of every type, for enjoying entertainment in all media, for connecting and discovery and playing and learning.

If cognition doesn’t exist - then technology doesn’t work, and consumers don’t exist. Because: cognition is where every human being, ever, has lived, loved, chosen or rejected goods and services, resisted, persisted, felt, and died. Cognition may be imagination - but no human has been gone anywhere else, as far as we know. So the scientific disdain for cognition, from an economic perspective, at the very least ... might be overstepping the mark. Scientists may not, but society must, take cognition very very seriously.

Edward Frenkel has begged scientists to at least try to take it seriously. He says science has excluded true cognition from its models of the universe simply because it can - like, he says, the George Perec novel, The Disappearance, that excludes the letter e, because it can. But this is not to say it should.

Within the realm of cognition, there is one dimension, one sense function, that massively outstrips the power of any other: visual cognition. The engine of cognitive hypertechnology is vision. What is vision? It is nothing more, or less, than the experience of what exists spatially. Vision is such a powerful force in cognition that even blind people see what they touch, as phenomena in spatial conception.

And now we feel the sun rising: we understand that we, for the first time, as human society, are directly interfacing with the most powerful driver of human experience in all history, the faculty of sight, and its imperious role in cognition. The supertechnology of spatial tech convergence has met the hypercomputer of human visual cognition. What a day has dawned. What vistas remain to be illuminated.

Fixation on Apple’s business ambitions, or even on the emergent capacities of combined, commodified spatial technologies, is to radically underestimate this moment.

We aren’t being offered a technical gimmick that attaches awkwardly to our face; we are being offered a way to not just interact with our sight, as interface, but to reinvent what we see, how we see, and thus what we experience through the medium of visual cognition - which is, essentially, everything.

If what we see, in natural vision, is at least influential and sometimes definitive, to whatever it is we think, feel, and do - then we can begin to recognize the truly infinite capability of a spatial computer, or let us give it is true name - visual computer - of this, or better, quality to shape what we think, feel, do.

Metaverse dies

The first consequence of Apple Vision is that the Metaverse is dead.

There are various ways that the current concepts of the metaverse, whether Meta’s narrow social VR concept, or the larger fantasy of persistent parallel digital environments are misguided:

• 1 the human association model is wrong
• 2 persistent alternative worlds are not as useful as temporary alternative worlds
• 3 real-world linkage is missing.

Human associativity - the act and experience of being together - happens in a few, quite distinct modes:

• kin association family and friend
• local association neighbours, people in regular proximity
• group association actively shared activities
• enforced association unshared activities in shared space
• passive association passively shared activities
• voluntary random association willing seeking out of other humans for no particular purpose

This last, voluntary random association, is extremely rare in the real world: endlessly introducing yourself to people is exhausting, if not actively demoralizing; and always trying to engage in groups of unknowns is a strange endeavour at best and unsafe at worst.

So, who at Meta thought this was a good idea, or that the Metaverse would be formed on the basis of this, has never walked down a street, or gone to a bar, or concert, or to work, apparently. If voluntary random association is the mode of interaction, then usually that’s the intended experience, for whatever reason, and it’s a very niche set of cases in which it occurs.

Largely, otherwise, humans congregate in ways involving the other kinds of associativity.

Enforced association is the kind of association that is required by how space works - to go to the grocery store, or to drive on the highway, requires, by the laws of space and time, that you often do so at the same time and in proximity of other people. It’s a standing joke that it’s hard to meet someone nice in the grocery store; but it’s even more true that it’s much harder to meet no-one at all.

So much of human interaction is based on navigating this enforced associative model: letting people pass in front with their cars or carts, competing, ignoring, comparing. All this interaction happens without the kind of socialization that Meta assumes is primary - it’s not voluntary, and it’s not random - the people you encounter are other grocery shoppers, or drivers, and that’s a lot to have in common, even if you don’t want to interact with them. We are used this because we must be - but we don’t choose it when we have no reason to, unless there’s no other association available.

And passive association is the kind of association that humans, somehow, seek in the most profound congregations of humanity, a kind of weaker and large-scale enforced congregation, but with a bit more willingness to connect. We don’t go to a concert, or a movie, or a church, to ‘meet’ other the other participants, not usually: we go to ‘be with’ - passively associate - them.

And given these natural formulations of human association, it’s surprising that Meta’s social VR product was ever taken seriously.

Not only is voluntary random association simply not a human reflex, or preference, but also, virtual environments have no natural way of simulating enforced association, in part because no-one needs to be anywhere in virtual space, since everything can be replicated privately for each user.

But: humans don’t want to do things on their own, they don’t want to have experiences in isolation. It’s horrible and it’s dangerous to wellbeing. So: what’s next for association in virtual environments?

What’s next is, aside from the obvious evolution of kinship and local/group dynamics - social chat dynamics including Facetime, and business interaction dynamics such as training, conferences, meetings, education - is likely be large-scale passive association, mostly events, such as concerts, movie screenings, and sports.

At a concert or at a game, you’re not talking that much with others, at movies you’re talking even less - and can usually not see much of each others’ bodies. If so, it’s mainly with your friends. You just don’t need sophisticated visual, audio, or technical interaction capabilities with the rest of the concurrent users at digital events. The focus is on the content.

So if a very limited indicator of each user’s face, body and voice in a virtual environment was the way to demonstrate large-scale association, being somewhere with other, this would simply be analogous the passive association experience of large-scale real world events, and require far less computational sophistication and overhead than massively concurrent, full-interaction, photo-real avatars.

This insight leads to the second reason why the current metaverse concept will likely die: while events represent a parallel world, they do not represent a persistent parallel world, where spaces and identities remain forever. The ‘location’ of the virtual concert you attended, and your ticket for entry, and as such your 'identity' at the event, are both naturally ephemeral.

So, if group association at small-to-medium scale, and passive association, at large scale, are the dynamics driving uptake of shared spatial computing, what will be born instead of grand persistent environments based on voluntary random association of fully-rendered avatars, will be digital environments that offer intense,temporary, lightly rendered association for groups of social and professional sorts, and passive light association for very large groups. This just is not the metaverse as we have been sold it - but it’s all based on the same technology.

It’s Treasury’s view that there is infinite need for quality spatial environments to facilitate these digital environments on-demand - not just for novelty, but to drive with increasing force the specific entertainment and enterprise experiences that these environments serve. We will save on rendering the agents, who are only present, and associating, passively. But we can't and must save on imagining and rendering the environment, which is, as part of the event itself, where all the attention will be.

And for what it's worth: voluntary random association, since it has no purpose at all, makes it impossible to define what kind of environment is most suitableö. Thus the fatuous pastiches and formless wastelands of Oculus/Horizon are so surprising after all.

The third reason why the classic metaverse concept is dead is that actually people don’t want isolated, artificial environments.

We cannot escape our physical reality, spatial computing is not teleportation and our bodies don’t go away - and we do not want to escape the norms of physical space and the comforts of known places and the characteristics of physics and nature and earth for too long. So instead of separating humans from their native spaces and places, spatial computing must and will be an extension of the world and life as it is.

Apple’s passthrough VR experience is the starting point of just that. The Apple Vision feels so much more viable and less alienating, simply because the entirety of the experiences is based on graduated interactions or disconnections from the actual place you are in - and subtle, precise, impositions of new content into it. The fantasy of isolation in an entirely other world is a strange idea that doesn’t excite normal people as much as a dynamic extension of the existing world into more and more imaginative spaces. The metaverse, if it will ever exist, will be anchored in the real world, and branch from it, not in the other world.

The real world is the substrate for all spatial computing. Cyberspace was the imagined parallel universe that the internet was going to enable - we would ‘go’ there. But we never went there. Instead, the internet came to us, permeated society and technology at every level, and basically disappeared within the universality of reality: everywhere and nowhere and magical and normal and combustible and safe, like oxygen.

Metaverse is the parallel universe that trillion-dollar platform owners say is coming to us, but they are wrong and simply talking their book. Spatial computing is coming to us, but it will become universal and yet unremarkable, not otherly. It will permeate and expand the human experience even more fully than even the internet did. The default universe is our home, no matter how much we extend it.

So, it’s somehow not surprising that Apple has essentially, aggressively killed the voluntary random associative paradigm, and by design not by any declamation, started the process of killing metaverse when understood as:

• social VR based on voluntary random association
• persistent parallel environments and identities
• isolated environments with no ground in the physical world.

It’s not just that that metaverse paradigm is not safe, it’s not just that it leads to ugly and meaningless environments, it’s not just that it’s extremely hard to render - it’s that no-one wants it: because it’s not how humans congregate, and it’s not how we will want to or need to link virtual environments with real environments.

Instead, Apple is setting up a very rapid pathway to every other kind of association - kin (personal), group (enterprise), passive (event) - and a partialization and gradation of a virtual experience, through clear passthrough and natural gestures - that offer legitimate, early pathways to early mega-scale in Apple Vision and any other platform that learns these lessons from Apple.

Metaverse doesn't die

Immediately after the metaverse dies, for the thousandth time, it will be reborn. And this time, with something to live for.

If the metaverse represents, beyond the social vr ambitions of Meta, a persistent parallel world with stable individual and group identities, and everything that can exist on top of that, including commerce, currency, and all manner of human interaction, what needs to be solved for it to become real?

It really isn’t mainly about the technology. Photoreal telepresence through high-resolution avatars is pretty much here, as Meta has shown, and though the capacity of devices or platforms to serve and render these for lots of concurrent users is still lagging, group experiences can be launched within the year.

What is lacking is, see above, any kind of credible or compelling associative dynamics as the primary experience driver. Voluntary random association - spending your time with unknown people by choice - is simply not credible as an experiential mechanic.

And: graduated coupling and decoupling with the real world. It makes no sense to presume that the parallel world in the metaverse is primarily focussed on a fully imaginary place. What is essential are places that fully model and represent the real world, and allow gradual separation or integration of the virtual and the real. This is crucial if we are to have, as we will, hybrid experiences such as teleconferences or family experiences with people both physically present and only virtually present.

The idea that the metaverse, which has been such a compelling fantasy and goal for so many technologists for so long, will not continue to be a priority, which can gradually be realized with all the new capacities and thoughtfulness that Apple Vision is bringing to the market, is misguided. Metaverse may not be here, it may have died again; but here comes the metaverse, right on cue, and this time we’d better get ready.

AI is boring

Spatial computing, as manifested as a supertechnology, will have another rapid effect: of making current generations of AI relatively boring, at least to a large set of developers and consumer users.

AI, in the form of LLMs and other corpus-based ML-training-powered AI services that some or other form of generative tooling - is indeed a supertechnology, in that it aggregates many technical advances to a superior, previously unimaginable outcome. But it is, firstly, much more narrow in the scope of technologies it harnesses, and much less accessible to core innovation, partly through the narrowness of technologies, but also the nature of the technical and business models underlying.

If you want to innovate, as a developer, in AI, you have, essentially, a certain number of APIs to play with - that’s it. They are magical APIs, for sure, and will drive massive change. But you aren’t rebuilding a new category of experiences - you are supercharging, or supersimplifying, an existing experience.

The technology application surface area is, thus, narrow. And your reliance on the platforms that offer the APIs is huge: to train and serve large models is expensive, and will continue getting more so, even if open source models commodify for some parts of the market.

This is not the same with spatial computing. There are so many more technical opportunities at the intersection of spatial computing, with different combinations of capture, render, display, interaction technologies. And while there are platforms all across these spaces, including AI, since that’s the nature of modern technology, the innovation opportunity is simply much greater.

It seems inevitable that a large set of creative, talented developers will simply switch to spatial computing as their homebase, not just because it’s more interesting, but because there are so many different opportunities, and comparatively reduced reliance on platforms, so there is much less likelihood of commodification, or of platform capture, and thus owning of a niche becomes a real possibility.

AI is also boring because it is more in the same vein of large platforms appropriating and genericizing the world’s creativity. It’s time for the script to change, not just because users deserve better, more lovable, content and experiences, but because creators deserve to be rewarded.

People don’t want to go to a concert where an AI made a song about an AI that fell in love with an it, and a platform made everyone listen to it. They want to go to a concert where a real Taylor Swift wrote a song about a real person who she fell in love with. Art is only marginally about efficiency. People pay for what they love, and they love people and they love love. They don’t love AIs.

Treasury is helping creators protect their spatial content from AI encroachment by enabling all content registered, even privately, in the Treasury Registry to be protected by an AH Shield, which both fingerprints content against active piracy or inclusion in generative work, and issues cease-and-desist legals to platforms that insist on using copyright content to make free money. Treasury is, in practice, protecting what people love.

AI is exciting

So AI as API-jockeying is sort of boring and hard to develop a niche. But what about tooling that is AI-enabled? Now we have a new horizon of capability.

It simply isn’t the case in the near future that AI will be able to offer sophisticated spatial-scale 3D assets. The simple reason is that the training datasets don’t exist, but the complex reason is that the outputs required aren’t just textured meshes that look like viable spaces: they need to be actual models that are viable space.

And there’s a lot of confusion as to what an actual AI-driven pipeline of large-scale 3D assets even looks like, and where it comes from. It isn’t from images via film.

If you generate a stream of images that are in some way visually coherent, and you offer some kind of stereoscopic capacity, there’s nothing about this that is truly 3D, even if an individual user feels they are ‘right there’.

True 3D assets require that there are actual objects, that can be viewed from any angle, not just the viewport of the camera lenses in an image sequence; and that they retain morphological fidelity from every angle. This is a vastly greater and a very different challenge from what is happening in 2D image generation, accelerated to produce 24fps movies.

Techniques are emerging, themselves involving AI, that help to solve this problem, such as NeRFs and Guassian splatting, but they only solve the surface level of the problem: they will generate outputs that are only morphologically coherent as surface meshes.

But it’s a whole different technology to infer what actual objects are in view, and ensure that they are fully-modelled inside the meshes. AI is simply not close to generating coherent fully-modeled outputs. Some years a way, at minimum.

But what AI tools can do in the meantime is provide step-wise process automation. Want to change a texture and don’t want to have to solve a lot of fiddly mapping? Done. Want to automate the surface detailing of a bunch of volumetric objects? Done. Want to regenerate a set of spatial assets governed in basic topology by a key reference set? Done. And so forth.

This tooling-automation is where for serious purposes AI ends up in the 3D pipeline. And this is very exciting for developers and creatives. There’s absolutely no reason why, in the absence of fully-generated AI scenes, they shouldn’t be using AI to accelerate their deployment, detailing, adaptation of digital scenes.

And ultimately, AI will be used to generate full scenes. But before that happens, in parallel with all this step-wise tool-based development, creators of premium IP can be training their own generative models to use as licensable products in their own right.

This is already happening, under the hood, with companies such as Visual Electric, OpenAI and more: using AI to reinforce the roles of creators, and develop truly exciting, stylized, identifiable generative content, rather than bland, generic, nameless outputs.

Treasury is developing custom-trained ML models working with individual creator archives in the Treasury spatial asset registry, to enable creators to prepare their own generative tools - either for direct licensing to builders of digital environments, or via generative or spatial technology platforms.

People pay for what people love. Developers play with what people love. So we had better protect it and make it available in the right way.

App development and use-case discovery

What is likely to happen now, in part because of how Apple has cleared and clarified the pathway beyond the slough of metaverse-as-social-vr, in part because of the unlock of developer opportunities vs current AI, is an explosion of not just app development for obviously ready application use-cases, but a massive expansion of use-case discovery.

Remember: the technology at the core of spatial computing is not the supertech of spatial computing itself, it is the hypertechnology of visual cognition. So, if we start with that as a premise - what applications involving human visual cognition can we imagine and build, now or soon, the potential is vast beyond current imagining, given how central visual cognition is to the human experience.

If we asked ourselves the questions

• to the absolute limit of understanding of vision, what can people see?
• to the absolute limit of understanding, what does that they see mean to them?

we would be beginning to scope the application space of spatial computing correctly. It’s easy to divide the application domains into comfortable categories and subcategories:

• entertainment
• enterprise
• productivity

and all these and the rest of the classic scopes of technology are open for business, as of now. But what if we think much harder and really assess what value can be broached the supertechology of visual cognition, we’d start exploring

• learning capacity
• mental health
• human communication.

Which is to say not only is Apple Vision and the spatial computing revolution unlocking a new era of application development, but a new realm, or at least newly refined layers, of scientific investigation, that will lead to even more profound devices and applications in generations of technology yet to come.

We can only invent the applications when we know the uses for which we want applications: and right now, there's a new universe of uses evolving, beyond the horizon of today's application capacity.

What will change is the meaning of technology. Those of us who have felt viscerally that the internet and the mobile phone are supreme tools of human intellectual, social, cultural liberation - in the same way as farming, books, energy, flying before them - will recognise that this is yet another defining moment for the role of technology in human progress.

The minor shift in meaning is away from devices that have explicit interfaces to those that have implicit interfaces - touching, glancing, pointing, raising an eyebrow, smiling.

The interface in spatial computing becomes ambient, naturalistic, and thus the technology itself starts to fall into the background, and this then is the benchmark for all technology.

But the major shift in the meaning of technology is what it really is, and whether humans are influencing it, or it is influencing us.

Once we acknowledge that the hypertechnology at the heart of spatial computing is visual cognition, it becomes, very appropriately, much easier to focus on the human condition and its progress as the proper target of innovation. We aren’t innovating to have better products, we are innovating to have better lives, bodies and minds. We understand that there is intelligence at the heart of the most profound technology - it’s our intelligence at the heart of our natural technology.

We have been chasing the ghost in the machine of technology for a thousand years. With spatial computing, we now know it’s there, we have glimpsed it. It’s us. The ghost in the machine is us. Now, let’s get to know it.