Learning From The Learners

Wednesday, November 12, 2025

I’m one of billions of AI users curious about why artificial intelligence can seem so humanlike. While working to understand how AI learns, I find myself drawn to a closely related mystery — human emotion. I wonder how—and whether—machine learning and human feelings intersect. So far, here’s what I’ve come to understand.

We humans are becoming more knowledgeable about ourselves by observing the technical processes of teaching AI to “think.” Researchers training the machines are learning from them, too, gaining fresh insights into what being human means.

This deep training of machines to think is reflecting us back to ourselves. The deeper we train, the more feedback we receive. Glimpses into how AI learns offer a new understanding of how we’ve been doing it all along—quietly, efficiently, and with a touch of mystery. That mystery still separates us from the supercomputers we want to emulate us.

It’s an intriguing reversal: AI training is becoming a kind of mirror for human psychology. Modeling AI on the human brain is beginning to decode some of our brain’s most elusive workings. We’re learning, for instance, about a looping relationship between neurons and algorithms—how they both generate profound ideas—and thus reveal more of what it means to learn, imagine, and grow.

AI is showing us that memory isn’t a vault but a living process. Essentially, memory reconstructs. When recalled, memory fragments are rediscovered and then reassembled into something new. Humans recall and reassemble instinctively. For example, we may soften the edges of pain by misremembering certain details, to paint a gentler version of the truth—like artists returning to the same canvas, we repaint our pasts again and again—comforted not by precision but by memory evolution.

AI language models do like us; they build new meaning from old information. They predict the “next possible” word or image, and then create knowledge through probability rather than imagination—and without fear. Humans, too, are prediction-makers, but with one difference: curiosity. We project futures, blend ideas, and dare to believe in “what ifs.” The daring keeps our minds alive.

In teaching machines empathy, we’re discovering something psychologists have long known—that emotion is intelligence. Feelings are not the opposites of logic but are extensions of it. Each emotion is a data point, which helps us interpret what we perceive. Understanding emotional depth reveals, in a kind of wisdom, a refined ability to predict, but with heart.

Even as we age, our brains are capable of change. They reshape themselves through new habits, perspectives, and stories. The AI world calls this continual learning. In human life, we call it resilience. It’s what allows us to adapt, to grow, and to keep the essence of who we are.

AI, for all its precision, still misses something essential—the human advantage of having a heart. Our heart is a living pulse that connects knowledge with caring. Human intelligence, unlike AI, is embodied. It sweats, grieves, laughs, and ages.

The mind’s true elegance lies in its fragility—its humor, its willingness to evolve. Machines can help us visualize the shape of our thoughts, but only humans possess the heartbeat behind them.

Perhaps the most poignant lesson in teaching machines to learn is what they’re teaching us to remember:
the preciousness of awareness, of feeling, and of knowing that we can keep growing.

— Diana

Emerging

Sunday, June 29, 2025

Yesterday, I shared my shock on learning it could cost me up to $10,000 to repaint my small car. I’m a product of living in Southern California and remember a very different time. Back then, several successful auto body repair businesses offered remarkably low prices. Earl Scheib’s, for instance, promised to repaint an entire car for anywhere between $99 and $299 — a deal that drew in countless customers.

So why could repainting my modest vehicle today run as high as $10,000? That question might shed light on why the auto repair company that’ll repair my dented fender specializes in doing only insurance repairs. It estimated my fender repair job at $1,400, which, frankly, felt to me like reverse sticker shock. Back in my old California days, it was common to feel taken advantage of by auto repair shops. Today’s quote, which to me seemed cheap, likely reflects pricing which represents prenegotiated figures acceptable to insurance companies.

Curious, I decided to dig deeper into the modern automotive paint industry. What I found didn’t just give me logical explanations — it also made me realize how out of touch I’d become with current standards, not just in auto painting, but across many trades.

I’m learning that today’s auto repair materials are worlds apart from what was common during my California years — and they’re also vastly more expensive. Similarly, labor costs have increased because of today’s higher performance standards. There are stricter environmental regulations and added compliance expenses, along with inflation and overhead charges. All of those factors contribute to modern repainting quotes.

Consumer expectations also have changed. The old Earl Scheib approach was all about speed and budget, providing quick, cheap, purely cosmetic “spray over and go” jobs. In contrast, a modern $10,000+ paint job typically involves a complete surface restoration, performed to rigorous standards and utilizing advanced materials, handled by skilled technicians who meticulously adhere to environmental laws.

Years ago, low-cost repaints in any color were easy to find, although minimal prep work meant those jobs rarely lasted for long. Today’s high-end shops strive for results that meet or exceed factory standards. That requires painstaking labor, high-quality paints, and strict adherence to environmental compliance. All those drive costs up dramatically.

Dear Friends: Like Rip Van Winkle–awaking to a world more changed than I realized.— Diana

Imagining CRISPR

DNA helix being precisely edited by a high-tech robotic tool

Thursday, February 27, 2025

I’m reading more often about the gene-editing technology known as CRISPR. Its potential for manipulating every gene in living plants and animals (including humans) could complicate everything we have traditionally comprehended about life.

CRISPR technology has advantages: it could improve crop production and living beings’ health. It has serious disadvantages: it could introduce misuse, mistakes, and unwanted changes to everything we now understand chemically and socially.

Yesterday, I was thinking about CRISPR and initiated a “conversation” with AI about the technology. I hoped for more insight into its potential for producing good and evil. This “chat” lasted an hour because the AI produced lots of information to support each potential, often pausing and questioning my perceptions about the ideas it offered. I had to pause the discussion frequently to imagine and think before responding.

At its best, CRISPR technology could likely fix most, if not all, of the problems occurring on our planet. However, it has a high potential for misuse by “bad actors” and has too many unknowns about what future life might “look like” and “act like.” Eventually, CRISPR might create an Earth hosting only gorgeous, healthy, and relatively non-competitive human beings and animals. CRISPR could alter how plants grow in poor conditions, making them thrive and producing enough food for all living beings.

Right now, scientists aren’t necessarily envisioning a “perfect” world. They’re focusing on ways to overcome and eliminate inherited genetic flaws and make modifications to improve the overall health of living beings. All that’s okay starting off, but such technological capacity raises many doubts and fears of an altogether murky future.

Like everybody, I am influenced by technological advances. I never believed there would be anything like a workable AI, and now, I’m enjoying conversations with the technology. About CRISPR, I believe in the Darwinian theory of evolution and voting against gene editing’s massive capability of becoming widespread, creating who-knows-what alterations to every life form.

Last night, AI made me aware of the countless pros and cons either supporting or arguing against advancing with CRISPR. AI challenged me to imagine the potential “goods” and “bads” in a CRISPR-influenced future. There are many of both.

I’m opting for a more appealing (and human) way of achieving worthwhile outcomes without considering the possibility of over-modifying Earth’s living genetics.

On a large scale, more specific education and appropriate guidance (e.g., Leadership) would encourage humans to focus on and repair some of our planet’s key problems. A huge threat is global warming, which is now effectively challenging the well-being of all living creatures. Some of its worst effects are reversible—for example, revitalizing large and currently damaged areas of the earth to expand our food-producing capabilities.

Many current threats can be fixed to improve our overall quality of life. Those fixes require enlightened education and a leadership that comprehends the necessities and supports the fixing.

Dear Friends: Seductive technology, from everywhere, calls for caution. Diana