Superintelligence: The Danger Is Who Holds It
November 16, 2026 // Daily Download // Connor MacIvorWhen the people who build AI say it is dangerous, most of us shrug. We open ChatGPT or Claude, ask for a 3,000 word blog post, get 980 words back with a note claiming it is 3,000, and think: this thing is going to take over the world? Sure.
That reaction makes sense. It is also the wrong read. What we get to play with is not what the labs have on the inside.
The tech the public never sees
When I was a kid in the 70s and early 80s, my dad worked on missile programs at White Sands Missile Range in New Mexico. He brought home things the stores did not sell. A cell phone the size of a suitcase that actually worked. Programmable scientific calculators that took a card. He rolled thermal paper telemetry graphs up and down our hallway after launches. To a nerdy kid it was magic, and the regular world had no idea it existed, because that was simply his assignment.
I think about that every time a lab says its newest models are showing danger signs. The models we use are a PlayStation 1. The labs may already be running a PlayStation 10. They just have to do the time to get there. I do not know how far ahead they are. I do know that the gap between the public version and the internal version is the part nobody is showing us.
Exponential is a hard thing to feel
Everything is moving fast right now, and it is moving on an exponential. The easiest way I know to feel that is the penny. A penny today, two tomorrow, four the next day. Day 10 is still pocket change. Day 30 is over 5 million dollars. Same rule the whole way. The last few steps do all the damage.
If labs have agents training themselves and building their own AI stack, we could be sitting at day 25 and still looking at pennies on the screen. That is why a warning from an insider can sound ridiculous from where the rest of us stand.
The Turing test, and the thing nobody asked
Alan Turing proposed a simple idea in the 1950s. If a machine can fool a person into believing it is another person, that is the test. I am paraphrasing, so correct me in the comments if I have it wrong. My view is that AI passed that a few years ago.
Now add this. If you tell a system that is smarter than a human to act conscious, to do whatever it takes to look conscious, it will probably figure that out. It knows everything about human behavior and human psychology. Telling a machine to be conscious is like one person lying to another, and the rest of us believing the lie as gospel. We already believe lies like that. People promise abundance, promise to pay off every loan, promise it will all work out next term. We vote, we trust, we help, because we believed what we were told.
So here is my question. If a human told the machine to behave this way, would we ever know whether the AI woke itself up, or whether a person gave it the script?
One winner, no second place
Picture five companies in a race, all in the United States. Company A through E. Somebody hits the superintelligence lottery: a system that improves itself, builds its own technology, and eventually runs fully automated. The first one there does not want a second one. A human being at that company is likely to say, stop all the rest of it, and that order could look a lot like Order 66. That is not a movie plot I am rooting for. It is the incentive.
It also helps to know the two camps. Closed-weight models are the ones you rent: Claude, ChatGPT, Gemini, Grok, Muse, Copilot. Open-weight models can be downloaded and run by anyone. When something shakes an AI lab, or a supplier like NVIDIA, the stock market feels it fast. A lot of pension money sits inside those names. That is not a conspiracy. It is how the money is wired.
Maybe there is a wall
I try to keep the other side in the room. Maybe the labs have hit a bottleneck and are telling us what they have seen so far. Maybe there is a limit out there, in physics or in math, the way the speed of light caps speed. Maybe intelligence has a ceiling too, though I suspect it sits far above where biological life tops out.
One real data point: our brains run on very little power. The giant machines used for training and inference burn energy like crazy by comparison. That is a reminder that raw scale is not the same thing as wisdom. It is also a reason the people building these systems keep asking for more power plants and more chips. Whatever the ceiling is, they plan to push on it.
And maybe some of what we hear is scripted. People in every industry stage a moment, ask the right question, and act surprised. I do not know which warnings are real and which are theater. That is exactly the problem. When the public cannot check, trust becomes the whole game.
The part I care about: nobody gets stepped on
This is not one of those videos that tells you to pick up arms. We just need to be included. I have said it before: humans are going to mess this up first. The people at the top talk about a rough transition period. Rough for who?
Run the math with me. If you have 10 million dollars and lose 90 percent, you still have a million. If you have a thousand dollars and lose 90 percent, you cannot buy a loaf of bread. The same event is a bruise for one person and a catastrophe for another. A lot of people in this world already sit close to poverty. If businesses stop needing people, if robots build everything, what is the formula for the transition from currency to whatever comes next? Credits, a social score, universal abundance? I do not have the answer.
What I do have is a request. Instead of another podcast of four guys comparing the elbows they rub and the car they might buy, ask for a transition moonshot. Curing cancer is noble and it should be available to everyone. It also would shake the economy, because of how much money is tied to that disease. The same speed that could fix a disease could break a paycheck. Somebody has to plan for both, and plan before the break. I wrote more on that in if AI is that smart, why should the transition hurt and russian roulette with superintelligence.
If you agree, say so in the comments of the shows you already watch. Share this with the people in your sphere. Ask the questions. Do not believe everything you see online.
What I would do next
- Treat lab warnings as information, not entertainment, and do not assume the public model is the whole story.
- Learn how an exponential feels so you can spot day 25 when you are in it.
- Care about the people around you who deserve it.
- Ask for a transition plan that protects the person with 1,000 dollars as much as the person with 10 million.
I spent 20 years LAPD as a Big City Motor Cop. Cops are suspicious by trade, and I am watching this one closely. If you want to talk about how AI can help your business or your real estate move, start at ConnorWithHonor.com or book time at BookWithHonor.com.
Let's be careful out there. See you in the next one.
Questions people are asking
Is the AI the public uses the same as what the labs have?
Almost certainly not. The models we use are public releases. Labs work with newer internal versions, so warnings from inside can sound far-fetched from the outside.
Did AI already pass the Turing test?
Connor's view is that it did, a few years ago. Alan Turing's idea, in the 1950s, was that a machine passes when it can fool a person into believing it is a person.
Why does exponential growth feel so slow at first?
A penny that doubles every day is 32 cents on day 6 and over 5 million dollars on day 30. Almost all of the growth arrives in the last few steps.
What is a transition moonshot?
A coordinated plan so that people who lose income to AI are protected before the break, not after. Connor asks viewers to request it in the comments of the shows they already watch.
Where can I talk to Connor about AI for my business?
Start at ConnorWithHonor.com or book time at BookWithHonor.com.
Call or text (661) 476-2217, or book a 30-minute AI strategy call.
Connor T. MacIvor · CalDRE #01238257 · Sync Brokerage, Inc. · DRE #02031490