An MIT professor changed one number from 0.506127 to 0.506. The computer responded by producing an entirely different future.
He had changed the input by 0.025%
Edward Lorenz was running a weather simulation in 1961. To restart it midway, he copied a rounded number from the printout instead of entering the full value stored inside the machine. He expected the second run to retrace the first.
For a while, it did. Then the lines separated. Soon, the two forecasts had nothing in common.
The computer was not broken. The equations were not wrong. Lorenz had found something worse: a deterministic system that could follow every rule perfectly and still become impossible to predict.
That accident became chaos theory.
Wall Street and AI still sell the opposite idea. Give the machine more data, more parameters, and more compute, and the future should come into focus. But in a chaotic system, a tiny error in the state you observe can grow faster than any model can correct it.
Markets are full of states no machine observes completely: hidden leverage, forced sellers, crowded positions, and millions of people changing their behavior because they saw the same forecast.
The model can be right.
The prediction can still die.
Lorenz discovered the limit on a room-sized computer in 1961. AI can reduce the error. It cannot repeal what the error revealed.
In America they told me the football game starts at 1 PM.
I arrived at 1 PM.
I was five hours late.
The parking lot was already a city.
A man had built a living room beside his truck. Not metaphorically. He had a couch. A television. A chandelier powered by a generator the size of a small horse.
He was grilling enough meat to feed a village, and when I walked past he said, "You hungry?"
I said I had not been invited.
He looked at me the way you look at someone who has apologized for breathing.
"Brother," he said. "You're here. That's the invitation."
He handed me a plate. It was not a small plate. The brisket hung over the edges like a man sleeping in a bed he has outgrown.
I ate it standing beside a stranger's couch in a parking lot in October, and it was among the finest meals of my life.
A woman across the row had a tent, a smoker, a speaker system, and a flag so large it could have sheltered a family of five beneath it.
She had been here since 6 AM.
The game had not started.
The game, I began to realize, was not the point.
I asked the man what time he would go inside the stadium.
He said, "Depends."
I asked on what.
He said, "On whether the ribs are done."
I want to be clear. He had a ticket. He had driven four hours. He had assembled a small civilization from the back of a pickup truck.
And he was considering not going to the game.
Because the ribs were not ready.
In Japan we tailgate nothing. We do not build living rooms in parking lots. We do not grill for strangers.
We sit quietly on trains and think about whether we remembered to bow at the right angle.
I have since attended eleven tailgates.
I have never once cared who won the game.
Nobody has.
The game is inside the stadium.
America is in the parking lot.
Beyond Quantum Computation: Constructor Theory
From the point of view of fundamental physics, quantum computation is an unfinished revolution.
Something is still to be done to complete this conceptual revolution that started in the 1980s.
This has impacts on both theoretical physics and technology.
Constructor theory is connected to the way in which I imagine that this revolution can be completed.
~Conjecture Institute Senior Scientist Chiara Marletto for @foresightinst
The real risk today isn’t just Social Isolation or AI Psychosis- it’s the insidious theft of our unobserved minds.
And "spending more time with people IRL" won’t fix it, as offline socialization perpetuates the same observation pressures.
Studies confirm that the awareness of being observed boosts performance on simple tasks by 10–15%, but impairs complex task performance by 12–20%.
In short, Y'all need to medidtate. The future of innovation depends on it.
There is no reason to assume that a scientific theory of consciousness must eliminate experience in order to explain it. Quite the opposite. The task is to connect experience with function and mechanism without pretending the experience itself is unreal.
A Neural Network That Can Grow Like Living Tissue And Repair Itself After Damage
A Neural Cellular Automaton learns one tiny local rule and applies it everywhere. Each cell only sees its nearby neighbors, but together the grid grows a full pattern from a single seed.
In the animation, the learned tissue grows, gets wounded, and keeps running the same update rule to rebuild the missing region.
The strange part is that there is no separate healing algorithm. The regeneration comes from the same local neural rule that created the structure in the first place.
Paper: Growing Neural Cellular Automata
Authors: Alexander Mordvintsev, Ettore Randazzo, Eyvind Niklasson, Michael Levin
Year: 2020