The Intelligence of Things
The Internet of Things connected things to the Internet.
The Intelligence of Things connects things to intelligence.
That sounds like a small change in language. It is not. It represents a fundamentally different way of thinking about every physical object around us.
Start with something painfully ordinary: a parking meter.
The Parking Meter Problem
Everyone knows what to do when they see a parking meter.
Or at least we pretend we do.
Walk up to one and look closely. There is usually a display, a card reader, perhaps a contactless payment symbol, a collection of buttons, instructions, warnings, rates, time limits, municipal regulations, QR codes, and sometimes a sticker explaining which app you are supposed to download.
Then come the rules.
Parking might cost one amount before 4:00 PM and another amount afterward. You might be limited to two hours. Wednesday could be street-cleaning day. Certain restrictions may apply only during particular hours. Your credit card might need to be inserted differently than it does everywhere else because the physical geometry of a parking meter forced the manufacturer to invent another interface.
We have become so accustomed to this that we rarely stop to notice how ridiculous it is.
The city has taken a simple transaction — renting a small piece of pavement for a short period of time — and turned it into a miniature computer-training exercise.
The problem is not parking.
The problem is the interface.
We Keep Adding Technology to the Thing
Look at a modern parking meter and you can see the history of computing attached to it.
The old mechanical meter is still underneath. Then we added electronics. Then a screen. Then payment processing. Then wireless connectivity. Then solar power. Then contactless payments.
We kept improving the parking meter because we assumed the parking meter needed to communicate directly with the human.
That assumption is about to disappear.
The Intelligence of Things begins with a different idea:
Stop improving the interface on the thing. Let the thing communicate with my intelligence.
The parking meter does not need to become easier for me to operate.
I should not have to operate it at all.
Bring Your Own Intelligence
Imagine walking toward the parking meter in 2030.
You already have your phone. More importantly, you already have your personal intelligence — your stack — with you.
Your stack assists you throughout your life. It knows your calendar. It knows where you are going. It knows how long your appointment is scheduled to last. It knows how you prefer to pay. It knows how much time you normally allow yourself to walk back to the car.
You do not open a parking app.
You do not search for the correct city.
You do not create an account.
You do not type in your license plate.
You do not read the parking meter.
You simply bring your phone close enough for the parking meter and your stack to recognize one another.
We already understand this physical gesture. We use it every day with Apple Pay and other contactless systems. Sometimes we move the phone around for a second until we find the reader. Then the phone takes over.
The same basic interaction introduces the Intelligence of Things.
The parking meter identifies itself.
Not to you.
To your intelligence.
It communicates something like:
I am parking space 1847.
Parking is currently permitted.
The rate is $4 per hour.
This space becomes unavailable at 5:00 PM.
These payment methods are accepted.
Those are facts about the thing and its current environment.
Your stack combines those facts with facts about you.
John has an appointment beginning in 18 minutes.
The appointment is scheduled for two hours.
The destination is three blocks away.
Allow enough time to walk there and return.
The result presented to you might be astonishingly simple:
Park here for 2½ hours — $10?
Yes.
Done.
The Interface Is Generated for You
This is where the Intelligence of Things becomes much more interesting than contactless payment.
That little interface was not designed by the parking-meter manufacturer.
It was not created by the city.
It was not downloaded from an app store.
It may never have existed before that moment.
Your intelligence generated it for you.
Someone else's interaction with exactly the same parking meter could be completely different.
Perhaps she prefers voice. Her earbuds simply say:
“This space works. It will cost $12 and covers you through the end of your meeting with 30 minutes to spare.”
She says, “That's fine.”
Done.
Another person's stack may know that he is particularly price-sensitive. It might say:
“This space is $16. There is a public garage one block farther away for $7.”
The parking meter has not changed.
The intelligence has.
This is why the Intelligence of Things is not merely another generation of user-interface design.
Eventually, the thing may not need a human interface at all.
The Thing Declares. Intelligence Decides.
There is an important architectural distinction here.
The parking meter does not need to know everything about you.
Your personal stack does not need complete control of the city's parking infrastructure.
Each participant knows what it is supposed to know.
The parking meter knows the parking space.
The city's enterprise intelligence knows the larger parking system.
Your personal intelligence knows you.
The thing declares its identity, current state, capabilities, rules, requirements, and available actions.
Your intelligence determines which of those facts matter.
Then it acts on your behalf.
Once you approve the parking transaction, your stack handles payment. The municipal system receives what it needs to know: this vehicle is authorized to occupy this space until this time.
The transaction is complete.
Three layers have collaborated without forcing the human to learn the internal machinery of any of them.
That is the Intelligence of Things.
This Is Not the Internet of Things
The Internet of Things was an important step.
We connected thermostats, cameras, appliances, automobiles, industrial equipment, sensors, machines, and billions of other physical objects to networks.
But connecting a thing to the Internet did not necessarily make the thing easier to use.
Sometimes it made things worse.
Now the thermostat had an app.
The doorbell had an app.
The lightbulbs had an app.
The car had an app.
The washing machine had an app.
The parking system had an app.
Every manufacturer created another little digital world that the human was expected to enter, understand, configure, maintain, update, password-protect, and remember.
We connected the things, but we left the human responsible for integration.
The Intelligence of Things changes that relationship.
You bring the integration layer with you.
Your stack does that.
Products Do Not Need Apps
Once you see the parking meter this way, the idea expands quickly.
A hotel-room thermostat does not need you to learn its controls.
Your intelligence already knows that you prefer the room at 68 degrees while sleeping.
A rental bicycle does not need you to download the bicycle company's app.
The bicycle can tell your intelligence what it costs, whether it is available, where it can be returned, and what actions are permitted.
A washing machine does not need a complicated touchscreen containing twenty-seven washing modes.
It can declare its capabilities. Your intelligence knows what you put inside it and what you are trying to accomplish.
A power tool does not necessarily need a thick instruction manual.
A lawnmower does not need its own proprietary software experience.
A hotel door does not need a hotel app.
A vending machine does not need a touch interface.
A rental car does not need a dashboard full of unfamiliar menus.
The physical object needs to be able to communicate what it is, what it knows, what it can do, what it requires, and what rules govern the interaction.
Your intelligence handles the rest.
That is a radically different design principle.
Walk Into Walmart
Now expand the idea.
Walk into Walmart.
Look around.
Thousands upon thousands of products surround you.
Today those products are mostly silent.
A box of cereal communicates through packaging. A television communicates through a printed specification card. A drill communicates through marketing copy on the box. A bottle of shampoo communicates through a label designed to capture your attention for three seconds.
The burden is on you to interpret all of it.
You compare.
You search.
You read.
You scan QR codes.
You pull out your phone and look up reviews.
You try to remember whether this is the model somebody recommended.
You wonder whether the part will fit the thing you already own.
You wonder whether the ingredients are appropriate for what you are cooking tonight.
The Intelligence of Things reverses that.
Every product becomes capable of interacting with your intelligence.
Not necessarily with a screen.
Not necessarily with a microphone.
Not necessarily with an app.
It simply needs an identity and a way of making useful information and permitted actions available to intelligence.
Pick up a drill.
Your stack already knows that you are repairing a deck this weekend.
It knows the tools you own.
It knows the batteries you already have.
The drill identifies itself.
Your stack says:
“This is compatible with the batteries you already own. It is more powerful than you need for Saturday's project. The less expensive model two shelves down is sufficient.”
Put it back.
Pick up another.
“This one uses a different battery platform. I would not buy it.”
That is a very different shopping experience.
Even a Tomato Can Participate
It does not stop with durable goods.
Eventually, even simple consumer goods and perishables participate.
Consider a tomato.
The tomato does not need artificial intelligence embedded inside it.
That would miss the point completely.
The tomato needs enough identity for intelligence to understand what it is.
Variety.
Source.
Harvest date.
Lot.
Price.
Relevant food-safety information.
Perhaps ripeness or storage history where useful.
Your stack supplies the intelligence.
You are standing in the produce aisle. Your stack knows what you are cooking. It knows what is already in your refrigerator. It knows that you are leaving town Thursday and should not buy food that will spoil while you are gone.
You pick up the tomatoes.
Your intelligence tells you:
“Buy two. You need them for dinner tonight. You already have three at home, but they will not be enough for the recipe.”
You are no longer browsing a database.
You are moving through the physical world accompanied by intelligence.
That distinction matters.
The Intelligence Is Not in the Thing
The phrase Intelligence of Things can create the wrong mental picture.
It sounds as though we are going to put a large language model inside every parking meter, refrigerator, tomato, pallet, and toothbrush.
We are not.
That would be both unnecessary and absurd.
The intelligence does not have to live inside the thing.
The thing needs to participate in an intelligent environment.
This is much closer to how the human world already works.
A stop sign is not intelligent.
It carries a small amount of information in a standardized format. The intelligence is in the driver.
A barcode is not intelligent.
A key is not intelligent.
A shipping label is not intelligent.
The value comes from the larger system that knows how to interpret them.
The Intelligence of Things follows the same principle, except the interpreter is no longer merely a human following fixed rules.
The interpreter is your intelligence.
And increasingly, it understands context.
Context Changes Everything
This is the crucial difference.
Traditional software asks:
What can this product do?
Intelligence asks:
What matters right now?
Those are not the same question.
The parking meter may expose twenty-seven facts.
You may need one.
The refrigerator may expose 400 data points.
You may care about none of them.
The automobile may contain thousands of sensor readings.
You do not want to see a dashboard displaying all of them.
Your intelligence watches the complexity and surfaces only the artifact that matters.
“Your rear tire is losing pressure faster than normal. There is a tire shop 1.2 miles ahead. They have your size in stock and can take you in now.”
That is useful.
A screen showing tire-pressure telemetry is merely data.
The Intelligence of Things is not about giving humans more information.
It is about eliminating information that humans no longer need to process.
The End of the Universal Interface
For the last fifty years, product designers have faced an impossible problem.
They have had to design one interface for everyone.
Young people.
Old people.
Experts.
Beginners.
People who speak different languages.
People with different physical abilities.
People in a hurry.
People who want every advanced option.
People who want one button.
That is why interfaces become complicated.
Every possible user has to be represented in advance.
Intelligence breaks that constraint.
There is no longer one interface.
There is your interface.
Generated when needed.
Generated from context.
Generated by the intelligence that already knows you.
For one person, the interaction is visual.
For another, spoken.
For another, almost entirely automatic.
For an expert, the system may expose tremendous detail.
For a child, it may expose almost none.
The manufacturer no longer has to predict the human.
The human brings the intelligence.
There Will Be Standards
For this world to emerge, things will need common ways to identify themselves and communicate.
Products will need standards for identity, capabilities, state, permissions, transactions, security, provenance, and trust.
Cities will need them.
Manufacturers will need them.
Retailers will need them.
Logistics companies will need them.
Hospitals will need them.
Automobiles will need them.
Homes will need them.
The standards will evolve, just as the standards of the Internet evolved.
But the important shift is conceptual before it is technical.
We need to stop assuming that every thing needs its own application.
It does not.
We need to stop assuming that every thing needs a sophisticated user interface.
It does not.
We need to stop assuming that every manufacturer needs to own the relationship between its product and the human using it.
It does not.
The thing needs to be understandable to intelligence.
That is enough.
From Smart Things to an Intelligent World
For years we have talked about smart homes, smart cities, smart appliances, smart cars, and smart factories.
The adjective has always been attached to the thing.
Smart thermostat.
Smart refrigerator.
Smart watch.
Smart meter.
That made sense when intelligence was scarce.
We tried to put a little bit of computing into each object and make the object smarter.
But intelligence is becoming abundant.
Once everyone carries powerful personal intelligence with them, the architecture changes.
The refrigerator does not have to be brilliant.
The parking meter does not have to be brilliant.
The washing machine does not have to be brilliant.
The product does not need to know you.
Your intelligence knows you.
The enterprise intelligence knows the enterprise.
The thing knows the thing.
And they collaborate.
That is a far cleaner world.
The Parking Meter Disappears
Return to the parking meter.
Today we see an object covered in the accumulated history of interface design.
A mechanical housing.
A digital screen.
A card reader.
A solar panel.
Payment instructions.
Warnings.
Labels.
Phone numbers.
Rules.
Every layer was added because the human was still standing in front of the machine trying to figure out what to do.
The Intelligence of Things removes that burden.
Eventually, what remains may barely resemble the parking meter we know today.
Perhaps it is simply a small durable device associated with a physical space.
It identifies the space.
It senses what it needs to sense.
It communicates its rules and state.
It accepts authorized transactions.
That is enough.
You walk up.
Your intelligence recognizes it.
“This spot works for your appointment. $10 for 2½ hours.”
Yes.
You keep walking.
The greatest improvement to the parking-meter interface will not be a better parking-meter interface.
It will be the day we no longer need one.