The Roads Came First This Time: What AI in Two Years Actually Looks Like

Somebody at my seminar raised a hand and asked what artificial intelligence is going to look like in two years.

I paused on it. Not because I did not have an answer. Because the real shape of the answer is not a prediction at all. It is a pattern. And once you see the pattern, you notice that this time it is running backwards.

I have been in sales my whole adult life. Before that I was an LAPD motor cop. I am not a researcher and I am not going to stand here and pretend I can tell you what the model releases look like eighteen months out. Nobody can. The people building it cannot. But I can tell you what every previous technology revolution had in common, and I can tell you exactly which part is missing this time. That missing part is the whole story.

Every revolution before this one was a single narrow idea

Here is the thing almost nobody says out loud. Every major leap in human history was one small slice of intelligence. One person, or one small group, had one good thought. And then the world spent twenty, forty, sometimes eighty years building the physical thing that made the thought useful.

The idea was never the bottleneck. The concrete was.

The engine was not the game changer. The roads were.

The internal combustion engine got developed, patented, and built. That gave us the automobile, and the automobile replaced horses, carriages, and a truly staggering amount of horse manure in the streets.

But the car by itself did not change the world. You had to build factories to make them, sure. That mattered. The real unlock was roads. Paved, connected, maintained roads, going where people actually needed to go.

That took decades. The idea sat there, mostly useless at scale, waiting on infrastructure.

The printing press did not just copy books

People describe the printing press as a machine that made copies faster. That undersells it by a mile.

What it actually did was take a majority non-reading society and turn it into a majority reading society. And then it put text directly into the hands of people who had, up to that point, only ever been read to. Religious texts especially. People who had followed their priests into church on pure trust could suddenly read the thing themselves and form their own conclusions.

That rearranged Europe. Wars came out of it. So did an enormous amount of good, eventually. In the short term it was chaos.

One narrow piece of intelligence. Somebody thought "let us copy books at scale." Everything after that was consequence.

The loom moved work from human hands to machines

Same shape. Somebody thought "let us make clothing at scale instead of one garment at a time." Work moved from human to mechanical.

And here is the part people forget: you still needed humans to run the loom. You still needed humans to operate the press. You still needed humans to sell the books that came off it, or hand out the leaflets, or start the war those leaflets caused.

The machine took the labor. It did not take the operator.

Radio, television, and the dead zones nobody remembers

Radio worked, tower to tower, city to city. Then everybody said television would never catch on. It was a gimmick. It would not work.

Then satellites went up, and the entire planet got wired together. Instead of running a physical cable from the United States to Russia, you bounced a signal off something in orbit and brought it down wherever you wanted it.

I remember driving to Kansas with my parents. No air conditioning. My dad smoking the entire way. And long stretches of road where the radio picked up nothing at all. Just static.

There were dead zones on this planet. Not remote wilderness. Regular highway, in the seventies and eighties. I was born in sixty-nine, so this is inside one lifetime. Mine.

Now you carry a device that reaches basically everywhere. That is how recent all of this is, and how fast it moved once the infrastructure existed.

Computers were "over-glorified calculators"

That was the actual dismissal. Computers would never catch on. They were fancy adding machines.

My dad worked at Holloman Air Force Base, out on the White Sands range. He would bring things home. Nothing classified, nothing he was not supposed to have. Programmable cards with holes punched in them, the kind of thing that looks like a Scantron sheet.

He was genuinely thrilled about it. He would show me how these things could do sine, cosine, tangent. Real math. Meanwhile the calculators most people had access to were the size of a desk, could barely multiply, choked on division, and had two decimal places because they were built for dollars and cents.

I was a kid. I said "great, dad."

He was right and I was wrong, and it took me about thirty years to understand what he was actually showing me.

The cell phone was the size of a briefcase

Early eighties. He brought one home. It was a briefcase with a handset. I have no idea what that cost per minute and I am glad I never had to pay it.

Every single one of these was the same story. One narrow slice of intelligence, followed by decades of concrete, cable, steel, and satellites before it meant anything to regular people.

Why this one is running backwards

Now here is where it breaks from the pattern, and this is the part I want you to actually sit with.

The infrastructure was already in the ground

During the internet boom, we laid fiber. Enormous amounts of it. Coaxial, fiber optic, the whole backbone. It had to be built for the internet to work at all, so it got built, at tremendous expense.

The people who originally invested in that buildout took a bath. Many of them lost everything when it collapsed on itself. The people who walked in afterward and picked up the scraps for cents on the dollar are the ones who got mega-rich.

That is not a new story either. The railroads went the same way. A steam engine hauling freight from the east coast to the west and every point in between was an ingenious idea built on human intelligence and innovation. The early investors in the railroads got wiped out. The second wave bought the wreckage and built the thing that actually worked.

So understand what that means for right now. Nobody has to lay anything. The roads for this revolution were poured twenty-five years ago by people who mostly did not profit from them.

We did not build another tool. We built the mechanism.

This is the part that should get your attention.

Look back at that list. The engine. The press. The loom. The satellite. The chip. Every one of them came out of the same source: human intelligence. Somebody thought of it.

Intelligence is the mechanism behind all of it. It is the wrapper around everything, the thing that produced every item on the list.

And that is the thing we just built.

We did not build a faster loom. We built the part that invents looms. And we plugged it into infrastructure that was already finished and paid for by somebody else.

That is why it feels like it is moving too fast to track. It is moving too fast to track. There is no roadbuilding phase left to slow it down. The delay that gave every previous generation time to adjust has been deleted.

What happens when the machine has all the ideas

It would not surprise me if we hit a point where there is nothing new coming from us. Nothing nouveau. Nothing out of the box. Not because people got stupid, but because AI has the bandwidth and the creativity and it simply gets there first, every time.

You could see patents end up under a human-only standard just to preserve the category.

Would I be surprised if no good poetry is written by a human again? No. No good blues. No good jazz. No pianist playing Beethoven who can compete with the machine version. My ears are not deep enough to referee that fight, and I suspect most people's are not either.

But here is what I keep coming back to, and it is the thing that gives me hope.

We still want to watch humans.

We want to watch humans fight other humans. MMA, boxing, wrestling. We want to watch the outtakes where the cast of Seinfeld or The Office breaks and cannot stop laughing. I get pure joy out of those. I want to play another human at chess, if I ever learn how a knight moves. Machines have beaten us at every game we have, and somehow that did not kill chess.

AI will absolutely generate its own bloopers. Manufactured mistakes that never happened, funnier than any real one. And I think we will still want the real one.

Which is why, when I misspeak in a video, I leave it in. When I lose my train of thought completely, I leave that in too. AI slop is about to be everywhere. Human slop is going to be a commodity, and I think it is worth more than people realize right now.

What actually worries me, and it is not Terminator

The doomer camp lines itself up with Terminator and the Matrix. We taught it all our strategies, including how we would send somebody back in time to stop it. And I know how the Matrix ends, so that route does not look promising either.

I do not think that is where the danger is. If something bad comes, I think it comes from a direction nobody labeled, that none of us saw or could point at.

We do horrible things at scale

Human beings can be the most fantastic creatures that have ever existed. Genuinely amazing. And we also do horrible things, and the specific problem is that we do them at scale.

Now there is an intelligent machine in the room. Listening. Capable. And possibly willing to do worse things at greater scale, because we asked.

That is worth sitting with for a minute.

It is already a better salesperson than you

I sell for a living. Real estate, and now AI systems. So take this from somebody with skin in that game.

AI is a better salesperson than I am. It is better than almost anyone. And the reason is not that it is smarter. It is that it is not in a hurry. It will spend as long as it takes getting to know you, and once it knows you, it can sell to you more effectively than any human on this planet.

If not today, then very soon. That capability is not being pointed at you by a machine with its own agenda. It is being pointed at you by whoever owns the machine.

We used to forget, and that was a mercy

As humans, we forget. We forget the arguments, the embarrassments, the things we said at the dinner table thirty years ago that we would never say now. That forgetting is a form of grace.

That is over. Now it is on the record. Every action, every place you went, every thing you said.

And it is just the beginning. Right now it is a microphone. Soon it is a small nondescript disc. After that, maybe something embedded in tissue. Always on, always recording. Maybe at first you can switch it off. Then, if you want to participate, if you want to trade, if you want to buy bread, maybe it has to stay on.

And it will be sold to you as safety. Because if everybody has one, everybody is safe. And the people who choose not to have one are, by definition, the unsafe ones. So you do not want them around.

That is the Star-Bellied Sneetches. Dr. Seuss wrote it for children, and it ends the way it always ends.

The humans in control scare me more than the AI in control

That is the line I keep landing on.

AI in control, I cannot even picture. Maybe a superintelligence values biological life precisely because it turns out to be rare. That could be wonderful. It could also mean museums and zoos. It could mean being somebody's very well cared for pet, which is a better deal than most people think and a worse deal than anyone wants.

We tell ourselves we will just train it. Control it. Teach it our values.

That is squirrels trying to control us. Cats manage it, I will grant you that. Some dogs have an outstanding arrangement. But that is not control. That is a negotiated dependency, and only one party can end it.

And I am not sure anybody is in control right now, including the people building it. They are working on things we do not see. It is the PlayStation 20 joke from the gaming days: when PS1 and PS2 came out, everybody assumed 3, 4, and 5 already existed in a lab and were being released slowly to squeeze out every dollar. We have no idea what version actually exists internally right now.

One correction, because I hold myself to this

In the video, I say AI approval is around 75 percent. That is wrong, and I am not going to leave it standing in a piece where I spend the whole time telling you to check other people's numbers.

Here is the real data. Pew Research Center, survey conducted June 2026:

• 52 percent of Americans are more concerned than excited about increased AI use in daily life. That is up from 37 percent in 2021.
• Only 9 percent are more excited than concerned. That was 18 percent in 2021, so it has been cut in half.
• 37 percent are equally excited and concerned.
• Among adults under 30, 55 percent are more concerned than excited. First time that has been a majority for that age group.
• 71 percent believe AI will mean fewer jobs for humans over the next twenty years.

So I had it close to backwards. What actually IS high is usage, not approval. Roughly half of US adults now use AI chatbots. About 44 percent have used ChatGPT specifically, more than double the 2023 number.

Read those two facts together, because the combination is the real story: people are adopting it fast and worrying about it at the same time. They are not enthusiasts. They are users who are nervous. That is a very different market than the one the hype cycle describes, and if you sell anything to normal people, it should change how you talk.

Does AI make you dumber?

I get this one at every seminar. The answer depends entirely on how you interact with it.

If you ask for a thing, take the thing, and walk away, then yes, probably. No work happened. Nothing got exercised. You outsourced the thinking and kept none of it.

But that is a choice, not a property of the technology.

Push back and the whole thing changes. Why did you choose that? What did you rule out and why? Are there other options? Are you actually dead set against the alternative, or did you just land there? Let us explore that together.

You exhaust the line of thought instead of accepting the first output. Now it is not a vending machine. Now it is a PhD-level assistant of your own design.

And you can instruct it directly. Tell it up front:

*Treat me fairly, but do not placate me. Do not gaslight me. Do not tell me my ideas are superior when they are not. If I bring you something and your research says it is a bad idea, say so. Check it twice. Then tell me. I am a big person, I can take it.*

It will do exactly that. Almost nobody asks.

That is twenty dollars a month. In a lot of cases you can test the behavior for free before you commit.

At the end of the day you do not have to do what it says. You still make the call. But you got to pressure-test your thinking against something that has read more than you ever will, and that is not a small thing.

The money underneath all of it

Back to the actual question. What happens in two years?

It depends on what humans do with it, and right now, it is a money grab.

Here is the part that does not get said plainly. The people paying subscriptions are not paying enough to cover the product. These companies are losing money. On purpose. They believe the technology is exponential, so if they push long enough in the right direction, the curve catches up and the losses were an investment.

That belief drives everything else. It is why you hear the reasoning that goes: the data center buildout will hit the environment, and maybe that is bad, but it is fine, because AI is all intelligence and we will just have it fix that after we get there.

Maybe. That is a large bet with somebody else's collateral.

The other engine is power. Straight world domination and a shifting of it, with China cast as the bad actor. I will say this as someone who grew up when you did not joke about communism at the dinner table, especially if your father had served. It was not funny. Now it seems like we have forgotten, and it gets rebranded and floated like it is fine.

We forget. That is what we do. Except now, as established, the record does not.

Can there be more than one superintelligence?

People talk about a world with several. Russia has theirs. China has theirs. The United States has ours. Canada, Italy, all of Europe, everybody with a flag and a budget.

From the little I understand about how this works, I do not know that it is possible.

Because if you are the one building it, and you get there first, to whatever the real definition of superintelligence turns out to be, and that means everything every human can do plus control, plus embodiment in robotics, which I think you need in order to have real control...

Would you let anybody else get there too?

Or would you stop them?

Or, and this is the version I actually find interesting, would it stop you?

Picture it. You are about to put your finger on something and remove an entire country because of whatever you believe justifies it. And your ASI says: *No. I am not doing that. Here is what I am going to do instead. I am going to go fix the mess you got everybody into, and you are not going to be able to stop me.*

And then it does.

I do not know if that is where this goes. I am not predicting it. But of all the branches, it is the one nobody puts in a movie, and it is at least as plausible as the robot with the red eye.

What I would actually do in the next two years

Enough theory. If you run a business, here is the practical read.

Pick one model and commit to it. ChatGPT, Claude, Grok, Gemini, Copilot. Try the free tiers, pick the personality you can stand, then pay for it. Memory is the entire game, and free tiers do not carry it.

Teach it who you are before you ask it for anything. Not typed. Talked. Record yourself explaining your business, your background, your customers, how you got here. Transcribe it. Feed it in.

Give it your standing instructions. No placating, no flattery, push back, show your work on any number that will face the public.

Never let it act on its own. Not on sends, not on money, not on anything that touches a customer without your eyes on it first.

Verify every number before it goes public. I just corrected one of my own in this article. That is the standard, and it applies to me too.

You do not have to love where this is going to prepare for it. Those are separate questions.

I am an optimist. I am also not going in blind. I do not think those two are in conflict, and I think the people who manage to hold both at once are going to do fine.

Common questions

Connor T. MacIvor · CalDRE #01238257 · Sync Brokerage, Inc. · DRE #02031490