At 83, I see better than I did at 38.
That is not a figure of speech. Two lens replacements — the clouded originals out, engineered lenses in, centered by laser to a precision no surgeon's hand could match. The world came back sharper than nature ever gave it to me. Colors I'd forgotten. Edges I'd lost. Technology did not restore my sight. It improved on the equipment I was born with.
I mention this because it is the smaller of the two things I want to tell you. And because it is the key to the larger one.
The detour
Two years ago I put down the things that had defined my later years — travel, paddling, pickleball, poker — and picked up AI. I have not put it down since. My life has moved to a different segment of the spectrum.
Two years ago, the tools that consumed me were chatbots. One question, one answer, a few seconds, and when they reached the edge of what they knew, they hallucinated something with perfect confidence. Impressive. Dangerous.
What sits in front of me now is a different animal. Not a chatbot — a society of tools. In November 2024, Anthropic released a protocol that let these systems reach out and pick up other instruments. By April 2025, Google had released another that let them talk to one another. Add models that now stop and reason before they answer, and coding agents that write and run their own code, and you no longer have a thing that responds. You have a thing that acts.
People keep asking how much faster it has gotten. Faster is the wrong word. It did not get faster. It changed kind. The chatbot became a workforce.
And yet
For all the billions pouring into this — and it is billions, fought over like territory — there is one thing none of it has bought. Not a dollar of it. Foresight.
These machines read the past with superhuman fluency. They cannot see the future. Whatever they call prediction is extrapolation — drawing the line forward from where the data already sits. That is not the same as sensing what has not yet happened. And on that count, a two-year-old beats every model on earth.
Watch a newborn fawn. For its first days it does one thing supremely well: nothing. It lies in the grass, motionless, and when danger nears it will flatten its ears, drop its head, and still its own breathing. It is born almost without scent. Stillness is its entire defense — and it holds that stillness until it receives one specific signal: its mother's call. Not noise. Not movement. The right signal, from the one source it trusts. Everything else, it ignores.
A days-old infant does its own version. Within weeks a child parses the world into solid objects, expects them to persist, senses cause and effect, and — long before it can speak — pulls away from a thing that is wrong before anyone explains why. The alarm fires ahead of the reasoning. That is not learned from data. It is older than data.
I have spent my life in the water and the open country — solo across the Molokai and Catalina channels, the depths of the Grand Canyon. You learn out there to respect the thing that goes quiet and still a half-second before you understand why. That thing is not in the machine. It may never be.
What the builder concedes
Listen to Demis Hassabis — a genuine scientist, a Nobel laureate, one of the most careful minds building any of this. When he describes what his systems are reaching for, he reaches for the same image I am: a child's intuitive grasp of how the world works, not a physicist's equations. He admits the line keeps surprising him — that the machines have absorbed more of that intuition, faster, than he ever expected.
But notice what he is chasing. He is chasing the present — how liquids flow, how light falls, how the world is right now. That is a lens on what is. It is a magnificent thing to build. It is not the fawn. The fawn is not modeling the physics of the grass. It is anticipating a threat that has not arrived, on a signal that has not come. Hassabis can be surprised a hundred more times by how much of the now the machine can hold, and the missing piece stands untouched. The missing piece was never the present. It was the not-yet.
Teaching versus learning
So what do we do with a machine that sees the present better than any human ever could, but cannot see ahead?
We put it where sight is the bottleneck.
We put it in the classroom.
"The bottleneck that defeated Bloom for forty years was never the knowledge. It was sight at scale. That is exactly what the machine supplies."
In 1984 the educational psychologist Benjamin Bloom measured something extraordinary. A child given one-on-one tutoring performed two standard deviations above a child in an ordinary classroom. The average tutored student outscored 98 of every 100 of his peers. I am an actuary; two standard deviations is not a slogan to me, it is an earthquake. Bloom called it a problem — not because it failed, but because it worked and could not be afforded. You cannot buy thirty children thirty tutors. For forty years that gold standard sat on the shelf, known and out of reach.
It is no longer out of reach
A teacher can now assess a child every day — see exactly where she is, what she grasped, what she missed, how fast she moves — and build her the next lesson, hers alone, this morning. Not the lesson aimed at the middle of the room, the quick ones bored and the slow ones lost. Her lesson. Thirty children, thirty paths, one room. The bottleneck that defeated Bloom for forty years was never the knowledge. It was sight at scale. That is exactly what the machine supplies.
This is the difference between teaching and learning. Teaching was a broadcast. Learning is personal. The machine enables the broadcast — the assessing, the pacing, the endless individual adjustment — and hands the hours back to the human, for the one thing the machine cannot do — learn.
And that one thing is the whole point. What we grow in the child is critical thinking — judgment, the capacity to choose under uncertainty, to sense danger and opportunity before the data arrives. The fawn's faculty. The toddler's faculty. Foresight. The teacher observes the child and assesses; the child becomes the navigator. We finally automate sight so we can spend ourselves growing foresight — in the one creature that can hold it. It will be the bedrock the worker of the next decades stands on, more than at any time in history.
There is a fair objection, and it belongs inside this argument rather than against it. A machine can personalize, the critics say, but it cannot be personal — no relationship, no spark, no judgment. Exactly right. That gap is not the flaw in the plan. It is the plan. The machine personalizes. The human supplies the personal.
And for the child the old classroom would have quietly given up on — the one the front of the room was never aimed at — the single teacher, now with the machine, will finally have the hours to build individual plans suited to the child's DNA and daily growth — hours she never had before. It does not sort her. It widens her aperture. One teacher can now surface many individual paths at once; the lens meets her exactly where she stands, and lets her choose. No child written off because the broadcast was not built for her.
What I have built
This has been built. Three dimensions — data, analysis, and AI — working together to do two things at once: build each child a new lesson every day, and, with the teacher's judgment in the loop, appraise the best path for her to learn. Now. The protocols exist. The models exist. The science — Bloom's, and a generation since — exists. The only thing missing is the will, and the capital, to put the lens in front of every child instead of the few whose families can buy it privately.
That is what I have built. H-EDU.solutions exists to do precisely this: to give every child the daily, individual sight Bloom proved would change everything, and to spend the freed human hours growing the one faculty the machine will never have. It is ready for someone with the vision — and the dollars — to see what I see.
Which brings me back to my eyes.
At 83, with engineered lenses, I see more sharply than I did at 38. Technology can manufacture sight beyond anything nature hands you. That is real, it is a gift, and I use it every day.
"But it cannot manufacture foresight. No lens does. No model does. That stays the work of the living."
But it cannot manufacture foresight. No lens does. No model does. That stays the work of the living — the fawn in the grass, the child in the room, the old man who has learned to trust the quiet half-second before he understands why.
Please visit the website https://case.h-edu.solutions and help our efforts to improve teaching and learning.
Don't lament. Engage.
———