The Inevitable by Kevin Kelly

Kelly makes it clear at the beginning that he did not mean “inevitable” in the sense of a “specific, pre-ordained, defined to the last bit” destiny, but rather in the sense that the nature of technological progress makes certain general scenarios more likely to occur than others.

And not all change is going to be good change.

He also makes another observation that I absolutely agree with which is that we are inventing new things and technological progress is so rapid these days that “… the answer comes before the question …” – That is to say the technology gets introduced first and only a few years or (more likely) decades later, do we finalise the place that technology will occupy in our lives.

You know like 3D printing, AI, AR/VR, crypto and so on.

The reason the author is confident in his predictions is because he purports to base them on the “basic physics” of digital technology, so, like the foundational characteristics of how digital technology behaves.

The thing about the inevitable is that fighting it is sort of pointless because of its very definition, so Kelly recommends embracing and working with those changes is much better.

Kelly also makes a point that makes sense if you think about it, he says that because change is constant focusing on the processes that drive change is much more useful than focusing on the products that those processes produce.

If we understand the process we will be able to predict and therefore prepare for the products and their impact.

In the context of this book, this means looking at the processes we have decided to follow as a global community (such as the scientific method, democracy, inclusivity, (increasingly) sustainability and so on), as a global tech. community to predict the kind of products that will emerge.

This also has a corollary that if you want better end results (end products) you must get the processes right (though honestly, this is so fundamental that we can apply this to our entire lives).

Fitting then, that Kelly has divided his twelve forces into verbs that represent something ongoing like “accessing”, “tracking” or “becoming”. Also, these are not discrete but overlap and influence each other.

This is not the first of twelve predictions but is important enough to write about.

“In the next 30 years we will continue to take solid things—an automobile, a shoe—and turn them into intangible verbs.“.

Like how cars have become mobility services, food has become food delivery services, books have become Kindle unlimited, DVDs have become Netflix and so on.

Technology has led us to “protopia”: Which is a state of becoming, so it is a noun (“state”) and a verb (“becoming”) at the same time. Like a forever work in progress.

This is constant progress not in a day and night difference kind of way, but rather very incremental, to the point that you’ll not even realise it is happening when it is happening and will only see the difference when you compare your life today from a decade ago. In this way we’ll be “perpetual newbies” as we are presented with an ever evolving technological landscape to adapt to.

The author gives the example of how massive corporations and very intelligent people didn’t realise the power the internet would have and instead dismissed it. They could not see the change happening before their very eyes because it was messy and very incremental.

In simple terms Kelly is saying that technological progress is going to be a constant now, get used to technology constantly progressing (the “pro” in “protopia”).

And upending existing ways of doing business/living, do not be surprised when this happens to you, instead, adapt to it.

Or better yet, leverage it.

Similarly, there is change taking place in front of our eyes right now – like IoT and AI (the author references these without calling them by name) – that we may be dismissing as a fad.

Just like corporations in 1990s imagined the internet to be just another way to deliver TV, we should not imagine the internet in 2050 (or rather any general purpose technology) to be a linear extrapolation of their present day application.

This also means opportunities are ever present and the best time to start-up is right now.

Do not think that all the good ideas have been taken.

When it comes to technological progress, it is important to not make a linear extrapolation when trying to imagine the far future – because it will not be.

AI will be like WiFi – You’ll notice only when it’s not working.

Many things have become free in our digital age (music, news, encyclopaedias, maps etc.) – General purpose AI will also be free in the future.

AI is the force of forces, the change of changes in the technology we are to experience in the coming decades – the author says that AI will accelerate all the other eleven forces – The addition of AI to inert things has been called “cognifying“.

A good analogy is electricity – “… the business plans of the next 10,000 startups are easy to forecast: Take X and add AI“.

Electricity flows in a grid, AI will also flow in a grid (internet).

Electricity is not really at one place in the grid, it is everywhere, AI too will be everywhere on the internet, distributed and emerging through the interaction of parts.

Electricity is a necessity, a utility so we ensure it is available cheaply, easily and abundantly. So will AI. “You’ll simply plug into the grid and get AI as if it was electricity. It will enliven inert objects, much as electricity did more than a century past.“.

Electricity powers most of our machines today, so will AI.

As inventors yesterday took manual machines and added electricity to make them better, inventors tomorrow take electric machines and add AI.

Today, electricity is everywhere but we don’t really think about it, will we approach AI similarly. But if you do think about it, you can list out many things today that use electricity to practically make your life better, so you will be able to do for AI in 2060.

“My prediction: By 2026, Google’s main product will not be search but AI.“.

Writing this in 2025 I am surprised by how accurate this prediction is turning out.

Cheap parallel computing + Big data + Better algos = AI wave

General purpose AI will benefit from lots of data and also be subject to network effects.

Which makes Kelly think “… our AI future is likely to be ruled by an oligarchy of two or three large, general-purpose cloud-based commercial intelligences.“.

But, like electric utilities, the companies behind these general-purpose cloud-based commercial intelligences will really just be providing the foundation upon which inventors will build on – what we interact with IRL 99% of the time will not be general purpose AI, but a general purpose AI backbone that is then uniquely trained and modified to serve a specific function very well i.e. a specific purpose AI.

This will be AI that drives our cars, keeps our home climate optimal, ensures we’re eating right and gives the right meds to make us better again.

Just like we wouldn’t want a distracted human doing the above for us, we wouldn’t want that with an AI too – so these will be narrowly focused specific purpose AI.

They’ll sound like humans but behave like robots – manically obsessed about the tasks they are supposed to do.

Companies will tout “our AI is consciousness free” as a feature.

So you are assured that your self driving car AI is focusing on the road and not instead dreaming about how life as a robotic surgeon AI would have been.

In 2024 the most interaction with AI is in the digital world, but the robots are coming soon (and yes, they will take our jobs) – when AI is applied to machines as I wrote above. “coming” is incorrect because robots as already here, just that they work in areas most of us cannot see, like warehouses, factories and labs.

So really, the more accurate thing to say is that “the next three decades will see a ubiquity of robots in common settings doing (replacing) common jobs done today by humans”.

I suppose I should point out that “robot” does not mean the stereotypical humanoid form – “To demand that artificial intelligence be humanlike is the same flawed logic as demanding that artificial flying be birdlike…”.

Robotic floor sweepers do not look like human floor sweepers because we optimise robots to sweep floors, evolution does not.

So, we’ll have robots everywhere but not in the robot form that movies show us – “Ninety percent of your coworkers will be unseen machines. Most of what you do will not be possible without them.“.

Just like many of us possess personal computers, and some of us are experimenting with “personal AI” (by prompt engineering LLMs to occupy specific personas).

Kelly thinks we will have “personal robots” which will allow us to create and invent like never before but only those who can best optimise the process of working with bots and machines can expect success.

Robots will not only do the jobs that we didn’t really want to do in the first place but more importantly help us discover the jobs that we didn’t realise we wanted done all along.

As I realised in The Book of Humans we’re not the only smart ones around, animals and birds and even microbes are smart in their own right.

Similarly, we should realise that the AI we create will not be a single kind of artificial intelligence.

Kelly gives a few examples of the different kind of AIs we can create, like one that is made for replicating itself quickly and exactly, or one that enhances its human’s intelligence (transhumanism), or one that is slow and deliberate because it literally refers to all of its memory (storage) before deciding.

Kelly recommends that we create a taxonomy of these different kinds of intelligences so that we can put them to the right use. The kind of intelligence of an AI will dictate their economic value (that is, what its creators are able to charge for it) just like the intelligence of a human dictates how much salary they can earn.

Eventually we will encounter an intelligence that is truly alien to us (“alien intelligence”), one that thinks in a way and offers viewpoints that we cannot comprehend.

Not unlike meeting someone from a completely unknown culture, like the Sentinelese.

We may need to work alongside or completely rely on this alien intelligence to solve complex problems that we cannot solve alone.

It will be a humbling time for all humanity as it relegates it position as smartest one around (not that it hasn’t done it in several fields already, but this will be the final set of blows).

From the ones giving orders we will become the ones taking them.

What about humans then? What will we be good for?

Tremendous philosophical questions and a permanent shift away from the stance that “humans are valuable because we can do what no other can”.

AI will challenge our identity (as it has already started) and then perhaps humanity will be able to define a new one for itself: “The greatest benefit of the arrival of artificial intelligence is that AIs will help define humanity. We need AIs to tell us who we are.“.

Your work will get digitised, replicated and distributed for free, in real time – The way to extract value from your work will have to change.

Kelly likens the internet as to copier, that is, once anything is put on the internet it gets endlessly copied.

Let me explain – we have become very good at digitising things – and once things are digitised:

• They can be easily split into their components parts, like a digitised book can be broken down into its constituting sentences and digitised music can be broken down into its constituting harmonies.
• These digitised parts can be easily duplicated infinitely with next to zero cost.
• Duplicated parts can more easily flow through a medium (the internet being the main one), and distributed freely.
• The distributed parts can then be worked upon by individual humans that tweak and modify them to create something new.

Like how I am using the digitised version of this book.

A book that was duplicated for me on demand at price lower than a non digitised physical book (though in reality it took the duplication service fractions of a cent to do it so technically they could have given it to me for close to zero money), then downloaded to my computer (as it was to thousands others who bought the book even if they were at opposite ends of the world), and whose digitised component parts I now use to create something new which is this field note.

This combination of digitised, real time, free, fungible is called “flowing“, our lives flow today.

Very poetic but it really just means we expect many things to be delivered to us in real time (music streamed the moment we want it) or near real time (groceries delivered in ten minutes).

Modern technology enables this very easily and will continue to even more in the times to come.

So (to make this personal), if your work can be digitised and distributed for free (which it will if it’s good, no point fighting it) – how will you make any money?

Why will your customers pay for anything if they can get it for free?

Why people will pay in a digitised world

1. To get it fastest – Immediacy – Not when it is created, but rather when it is being created, like a livestream by an author as they put together chapter after chapter of their book and take you along the meandering journey.
2. To get it personalised – Personalisation – Like a free copy of a song that you pay extra to sound the best when played through your AirPods and heard through your ears alone, or Aspirin modified to suit your DNA. Needless to say this requires deep trust between creator and consumer.

3. To get it to make sense – Interpretation – Data, data everywhere, not an insight to be found. That is our world today, we are swimming data and get overwhelmed by it. Even giant corporations are not immune. You are golden if you can help make sense of the data and get people to arrive at decisions and actionable insights. Like a blood test whose report, full of numbers and parameters is free, but whose interpretation and meaning is chargeable.

4. To care for the creator – Patronage – If the creator really touches them on an emotional level they will not really care about the extra cost even if they can get the work for free. They want to help the one that helped them (so as to speak). Kelly points out that four things must be met for a successful payout though: paying must be easy, amount must be reasonable, they must get something in return, and te payment must directly benefit the creator.

5. To get a unique experience or physical memento along with – Embodiment – This applies to the more sentimentally important things in their lives (like the final album of an outgoing band they purchase in vinyl or the defining movie in a series they get on BluRay or a ticketed meet and greet with the author of a book whose ideas changed their lives) and also to the moments they want to “treat” themselves.

6. To get it all in one place – Accessibility – It is possible to piece together an entire music collection from hundreds of free tracks that you scraped, swiped, recorded or downloaded from a myriad of websites. But it is also too much work. If you can get everything in one place, high quality, with a mark of authenticity, and at a reasonable price, you will.

7. To find new things that interest them – Discoverability – This is applicable to content aggregators like Amazon Prime Video, YouTube, SoundCloud and so on. It lessens the cognitive load on the customer to keep himself entertained.

8. To know it’s the real thing – Authenticity – Free copies come with risks such as malware. Plus, there is a social statement that one makes when they say “I paid for mine”.

We too will come to expect screens on everything.

Because screens can provide information that is dynamic, real time and therefore, helpful.

In fact, it will be the addition of AI, IoT and screens to mundane, everyday objects that will transform their usefulness to us.

And when you add the power of AI to inert objects (“cognifying”), even if it is a very basic AI like the one you would find on a box from Amazon, it can interact with you through the screen (for example, the Amazon box telling you how its journey to your doorstep has been, what all check points it went through before reaching you and whether anyone opened/tampered with it).

Kelly takes an unnecessary detour (IMO) when in this chapter, he starts discussing the future of books – he says that with advancements in e-ink technology, we could make a book that looks exactly like a physical book, except that the pages are all blank because they are e-ink pages, and through some interface on the book itself or through an app, you are able to dynamically change the content of the book to what you want to read.

Not only will we watch screens, but screens will watch us too – that is, through a bevy of sensors the devices they are embedded in they will be able to understand how we are feeling at that moment, and serve up exactly what we want.

Kelly also briefly talks about artificial reality (AR) and how it will become a reality in the next 30 years.

Ownership will change from “something I buy” to “something I subscribe to”.

If you think about it, what really matters to you is whether you’re able to access a good (or service) at the time you need it. Otherwise, possession without use, is just a waste of money.

According to Kelly, there are five big forces that will help this transition.

From purchasing things to subscribing to things: Dematerialisation, Read time, Decentralisation, Platforms, Clouds.

This basically means that we are able to reduce the amount of material used without compromising on the benefits that the original configuration used to bestow upon us.

Dematerialisation can be thought of on a spectrum: from partial to complete.

You start with light forms of dematerialisation, which basically means reducing the weight/size/dimension/mass of a product without compromising on its value (Kelly gives the example of the aluminium soda can that has fallen in weight from 73g in the fifties to 13g today without any consequent loss in its utility to us).

Farther along the spectrum is a complete elimination of the some or all of material cause, this is enabled by digital technology, examples abound like partial elimination of the asset heavy taxicab model by Uber and the complete elimination of physical music and physical movies by Spotify and Netflix.

Dematerialisation caused by digitation makes us treat a “product” like social property – instead of everyone owning the privately owning copies of the product, the digitised product is not owned, but in a sense shared by customer.

Customers themselves become “prosumers”, which is a term that describes a customer who not only consumes the product or service, but also helps in its subsequent development and enhancement like a producer.

Digital technology has also enabled us to get many things on demand like transportation, food, groceries, medicines, music movies, and a lot more.

So it makes sense, therefore that one would not be too inclined to purchasing things and would just access them whenever he or she requires.

Besides, we have a lot of stuff now while the number of hours in a day is same as ever.

So, if you want to enjoy all of these things, it means the total amount of time you can be spend on one thing would be a fraction of what you could when there were not so many things around.

And if you’re not going be using something a whole lot then you should just rent it.

Decentralisation

According to Kelly, “We are at the midpoint in a hundred-year scramble towards greater decentralisation” which was inevitable the moment we put cables around the world and birthed the internet.

The author talks about Bitcoin and Blockchain, but the key here is that as things become digital, they tend to become shared and as they become shared, also become ownerless.

Clouds

Reliable and fast digital connectivity has enabled us to put a lot of computing power in a network of computers “braided together seamlessly to act as a single large computer“, and thus make our personal devices lighter and cheaper.

While not ownerless (yet) clouds are our shared property like Kelly posited.

As there is a move towards less ownership and more renting caused by dematerialisation and decentralisation, it stands to reason that the few remaining owners of the assets would want to utilise them fully and not keep them unoccupied.

This is where the platform comes in as it allows owners of assets to have a steady stream of demand for their product (by the way, it also allows the customers for those assets to have a ready placed to satisfy their demand).

So, as we see more dematerialisation and decentralisation, we will see the emergence of platforms that balance demand and supply, and in their own right catalyse the shift towards renting over owning.

Kelly talks about how, as digital living pervades, we’ll all get a lot more forthcoming in contributing effort to causes we believe in because digital technology makes four things a lot easier.

1. Sharing (since you don’t lose a digital object when you share it)
2. Cooperation
3. Collaboration (digital objects can be worked upon simultaneously by many)
4. Collectivism

If you think about it, this already happens in small tightly knit communities/families.

But as people start living in cities and the causes they care about diverge and the people they know but do not care about increase, this tendency to share and cooperate diminishes.

Digital living of the future will bring it back according to the author.

So crowd sourcing will be a legitimate staffing plan for organisations (or cooperatives).

But you won’t just contribute out of the goodness of your heart – you’ll be paid as well, wherever your work is used in whatever capacity.

We’ll increasingly rely on “cognified” systems to help us choose what to consume – Because there is so much content out there, it is impossible to keep up and so you need something to filter out the things that you’ll not be interested in.

This already exists in the recommendation engines behind Netflix, Spotify, Amazon and so on.

Just that it will get a lot more deeply ingrained in our lives – what we watch, what we eat, what we listen to, what we engage in/touch/interact with, where we stay – you get the idea.

And to ensure that we’re only given the things we want we’ll have “personal avatars” that are like ourselves, but digital.

Yes, these avatars have information on what we like and do not, our histories, but they also have information on what we look like, our body measurements, what looks good on us and so on.

We’ll be so used to this personalisation, that in the rare cases it is not present, we’ll notice.

Personalisation, first in the digital world and eventually with what is physical, will be the default.

Human attention, for one. A place of land in our minds is what the advertisers pay for.

So it makes sense that the filtering systems we come to rely on will make a lot of money, as they do today you know?

Google Search, Spotify, YouTube and the like – all have a thriving ads business.

Some interesting ideas related to scarcity: advertisers paying us for watching their ads, senders paying us for reading their emails and people having a “influence score” that defines how much they’re able to charge.

Here is an important point he makes “In the goodness of time any particular technological function will act as if it were free.” – which means that over a long enough period, the technology itself as well as what it makes should become very close to free (like, the information on the technology will be free and the physical product it produces will be “too cheap to meter”)

So, what is really scarce is attention yes, but the attention we give to experiences (because goods will be too cheap to pay attention to).

Which means that experiences and the people/systems who deliver those experiences will really command the premium.

More new things will be created by remixing existing stuff, we will all be remixers more than de novo creators.

First of all it is getting easier than ever to just create stuff. I suppose I do not have to explain how. So that means there is a lot of raw materials to be remixed.

Then, editing is also becoming easier – take Canva (a tool for both creation and editing) for example, creating beautiful and imaginative designs was once the purview of the professional, was expensive and took a lot of time. Now you can create “professional looking” designs in a matter of minutes. You can take something that someone else made and remix that into something new.

Then, indexing – very important since there is so much raw material, we need a way to quickly get to the relevant bits for our remixing – is also getting there, it has existed for text for a long time (table of contents, bibliography, footnotes) and not being built for images and video (in fact AI in 2024 has become pretty good at quickly identifying images and indexing them, its slowly doing the same for video too)

Finally, rewinding – not only will we use what “is here today” to remix, we will also use what “was here yesterday” – because everything most things digital will allow a back/undo button.

Remixing will be easy because when something is digitised (which will happen increasingly more, refer point 5 above) it becomes non-excludable and non-rivalrous by default (you have to make extra effort to prevent it from becoming so). But this also “plays havoc with traditional notions of property and ownership.” – once something is digitised, do you really own it if it can be so easily copied, distributed and used by people simultaneously without hampering use by others? And if whether you own it or not is fuzzy, what about the payment you’re owed when someone uses your stuff to remix theirs?

VR and AR will be mainstream: Kelly talks about how VR is now possible through the use of mobile phone screens and the other sensors present in phones like gyroscopes, accelerometers and so on can be easily (and cheaply) moved to a VR headset.

Technology isn’t there yet, but he envisions VR/AR to be the regular way of working like how we work on laptops today.

Advanced headsets will even look at your hands and be controlled by gestures, have eye tracking technology that will highlight objects that you are looking at, that you could then select with your voice (“grunts”).

Given that this was written in 2016 and the Apple Vision Pro uses most of this, it seems incredibly prescient.

“A person mumbling to herself while her hands dance in front of her will be the signal in the future that she is working on her computer.“.

Plus, a wearables will get more intimate with smart fabrics and even brain implants (BMIs).

You will be able to layer on more information to everyday objects by donning a AR headset (or maybe something like a focus from the game Horizon) – like the prices for objects displayed on store shelves or the piping beneath the streets.

VR harnesses will replace some home gyms since it will be possible to get some nice cardio while running in a VR world of your choosing. Removing the AR headset (or whatever implement it’ll eventually be) will be like going through the world with one eye taped shut.

The things we can interact with and the senses we use in the interactions will continue to increase – more senses, more intimacy, more immersion. “Implicit in VR is the fact that everything—without exception—that occurs in VR is tracked.“. – tracked, mind you, not necessarily surveilled.

You will track your body (many of us already do), the level of tracking will deepen and a lot more data about our bodies will be generated, and powered by personal AIs the insights we will be able to glean will also increase, resulting in all of us having a choice to take greater control of bodily health. Kelly calls this the “quantified self“.

But not just health, our entire lives (“lifestream“) will be tracked (“lifelogging“) – our bodies, our location, our moods, our sights, our hearing, our smelling, our touching – you get the idea.

And once again, powered by AI, you will be able to crunch this massive database to figure out a lot about yourself, and if you’re so inclined, what makes you “tick”.

Kelly calls this being able to “google” your life, but I think it’s more like being able to “gemini” your life.

BTW, using AI to make sense of your data will not be a power user’s tool, it will be necessary because the scale of this data will exceed anything humans can comprehend.

Not only will we track us, others will track us too.

This is where Kelly says humanity arrives at a crossroads, either it goes the panopticon way where “they” track us but we cannot track “them” or we get to a mutually agreed “coveillance” where both track each other (like how it used to be in the old days where everyone in the village knew what everyone else was doing).

The first option is a “1984” like scenario, the second option though, that is actually a good one.

Kelly points out how we already sort of open give away all sorts of tracking information, just see how we update the internet about each second of our lives on social media (livestreaming is increasing), “vanity trumps privacy” and that this sort of mutual surveillance is our natural state.

So, it will be in our interest to be more open than closed about the surveillance we (people, governments and organisations) are doing so that the right kinds of rules (social or legal) can be built around what cannot be surveilled.

Like how when mobile phones came out they were ringing everywhere and people used to talk loudly in them, but with time we developed rules about where phones are allowed to ring, where they are not allowed to ring, and where they are not allowed at all.

Not only will we be able to track ourselves and each other but also things, you know IoT.

If these are personal effects then there are further questions if tracking this by others is also a form of surveillance (it is I suppose).

Expect a many surprises about what can and can’t work, about what’s normal and what isn’t, indeed about what’s true and what isn’t.

Kelly becomes philosophical in the eleventh chapter of the book thinking about what is this that we’re creating/dealing with.

He talks about how the notion that legit businesses cannot survive on a “free” core offering has been upended by Google Maps (free directions), YouTube (free entertainment), Wikipedia (free knowledge), eBay (free classifieds), Spotify (free music).

“… we are headed to a world where the improbable is the new normal.“.

He calls this force “questioning” which means that we’re asking a lot of questions (mostly to Google but recently to conversational LLMs as well), yes, but also that we’re questioning this new world we’re creating/experiencing.

Enough of those questions have received answers very different to what we’ve long assumed to be true (like, “Can I stay at home all day in my pyjamas and still be considered a productive, engaged member of society?” – if you ask the thousands of influencers and the millions working from home, yes, yes you can) that, in this age of information, we’re less certain than we used to be.

Speaking of asking questions, many of us are saying “let me google that” (I am more likely to say “let me ask Gemini”) – and the answers come within seconds and mostly arranged according to what we’re looking for. So, in a sense, the answers are cheap while the questions are the expensive thing.

This is a reversal of how we’ve lived since forever – getting the right answer has been a big thing and something to be taken seriously unless you want to end up hurt or dead.