🎩 Hatta here.

Yesterday we went backward.

More than 5,000 years backward.

We followed artificial intelligence into humanity’s great unread library, where AI may help us recover words pressed into Sumerian clay and buried inside carbonized Roman scrolls.

Today?

We are going in the opposite direction.

Far forward.

Possibly toward a future in which some computers are no longer built entirely from the familiar kingdom of silicon, metals and etched circuits.

Because scientists have begun putting something rather unexpected inside electronic memory:

Yes.

That DNA.

The molecule carrying the genetic instructions of life.

🐰🕳️

Researchers led by Penn State have developed an experimental bio-hybrid memristor using specially engineered synthetic DNA combined with a perovskite semiconductor and silver nanoparticles.

A memristor is an electronic component whose resistance can change depending upon what electrical activity has previously passed through it.

In other words:

Unlike ordinary resistors, a memristor can retain information about its previous electrical state even after power is removed.

That makes memristors especially interesting for future computing systems because memory and processing can occur much closer together, rather than constantly hauling information back and forth between separate processors and memory banks.

And that begins to sound suspiciously familiar.

🧠

Because your brain does not operate like a conventional computer either.

Neurons store information and process signals as part of the same interconnected architecture.

That is why memristors are often discussed in connection with neuromorphic computing, machines designed to borrow principles from biological nervous systems.

And now researchers have added another biological ingredient.

DNA.

🐰🕳️🐰🕳️🐰🕳️

Before Hatta runs away claiming somebody plugged a chromosome into a laptop...

A clarification.

The researchers are not putting human genes into a computer.

Nor are they storing Word documents inside somebody’s genetic code.

They created short, customized pieces of synthetic DNA, only 22 units long, engineered specifically for electronic behavior.

Natural DNA is long and floppy.

The Penn State researchers described it rather wonderfully as behaving somewhat like wet spaghetti when engineers try to handle it.

Not exactly what you want when building precision electronics.

So instead they designed short, rigid DNA structures whose composition and length could be carefully controlled.

Then they added silver nanoparticles.

The silver helped the DNA conduct electricity.

That engineered DNA was combined with thin layers of a crystalline semiconductor called perovskite.

Together they formed microscopic pathways through which electrical current could travel.

And something remarkable happened.

The experimental memristor operated at less than:

That is extraordinarily low.

The researchers also report that their system remained functional at room temperature for more than six weeks and could operate at temperatures approaching 250°F.

Their published study reports stable switching behavior over repeated cycles and exceptionally low operating power.

Penn State researchers say comparable memory functions could be achieved using dramatically less power than conventional alternatives.

Why does that matter?

Because AI has a hunger problem.

Not for data.

For...

Modern artificial intelligence requires enormous amounts of computation.

More computation means more electricity.

More electricity means more heat.

More cooling.

More infrastructure.

Larger data centers.

Greater cost.

So while much of the AI world races to make models larger and more capable, another race is occurring underneath it:

And perhaps biology has been sitting on part of the answer for several billion years.

🎩 Hatta peers suspiciously at his own hand.

Nature already knows how to build information-processing systems that operate on tiny amounts of energy.

You are carrying one between your ears.

The human brain performs extraordinary feats of recognition, reasoning, memory and prediction while consuming roughly the power of a small light bulb.

Today's artificial intelligence achieves astonishing things too...

but often by moving vast quantities of information through huge computational systems.

One major inefficiency in conventional computing comes from constantly transferring information between:

memory ↔ processor ↔ memory ↔ processor

This is sometimes called the von Neumann bottleneck.

Neuromorphic architectures try to reduce that problem by making memory and computation behave more like interconnected neurons.

Memristors are one possible route.

And bio-hybrid memristors add another intriguing possibility:

Now the Rabbit Hole gets deeper.

🐰🕳️

DNA evolved to preserve information.

Not for days.

Not for years.

Across generations.

And its theoretical information density is staggering.

Researchers commonly estimate that roughly 215 million gigabytes of information could theoretically be represented in one gram of DNA.

Imagine holding that gram between your fingers.

Millions upon millions of gigabytes.

Nature apparently invented high-density storage before vertebrates even bothered growing legs.

😏

Scientists have already explored DNA as a medium for encoding digital information.

But today's experiment is different.

Here DNA is not simply the container for encoded bits.

It becomes part of the physical electronic architecture itself.

That distinction is important.

And fascinating.

The researchers engineered the DNA so that its molecular structure could help control the movement of electricity through the device.

So DNA moves from:

into the realm of:

That little border crossing may someday matter enormously.

We should be careful here.

This is still laboratory research.

Nobody has produced a DNA-powered MacBook.

Nobody has grown ChatGPT in a petri dish.

And this particular experiment does not mean that tomorrow's AI systems will suddenly become living organisms.

But science often becomes interesting precisely where formerly separate fields begin touching.

Biology.

Electronics.

Materials science.

Nanotechnology.

Artificial intelligence.

Neuroscience.

These fields increasingly overlap.

And when disciplines overlap...

new creatures crawl out.

🐰🕳️

Imagine future computing materials that are:

self-organizing,

molecularly programmable,

extremely information-dense,

extraordinarily energy-efficient,

and partly inspired by the machinery of living cells.

Perhaps future computers will not simply be smaller versions of today's computers.

Perhaps some will operate according to entirely different principles.

Perhaps they will combine silicon with biological molecules.

Perhaps processors will increasingly mimic neurons.

Perhaps memory will become distributed throughout computing materials instead of living inside separate boxes.

Perhaps someday the distinction between:

computer hardware

and

engineered biological material

will become considerably fuzzier.

Artificial intelligence is usually discussed at the software level.

Models.

Algorithms.

Training data.

Agents.

Reasoning.

But underneath every brilliant AI response sits something brutally physical.

Matter.

Electrons moving.

Heat being generated.

Power being consumed.

AI ultimately lives inside machinery.

So advances in AI may depend just as much upon revolutions in materials as revolutions in algorithms.

The next giant leap may not merely be a smarter model.

It might be a different kind of machine upon which intelligence runs.

Silicon gave us one computing age.

What comes after silicon?

Photonics?

Quantum computing?

Neuromorphic chips?

Molecular electronics?

DNA?

Some strange marriage of all five?

🎩 Hatta quietly removes his hat.

There appears to be another hat underneath it.

Nobody asks why.

Life discovered DNA billions of years before humans discovered computers.

Then humans discovered that DNA stores information.

Then humans learned to read DNA.

Then edit DNA.

Then synthesize DNA.

And now...

we are beginning to engineer DNA into electronic devices.

Which leaves us with today's question:

Maybe nothing revolutionary comes from this particular device.

Science does not guarantee us dramatic endings.

But the direction is extraordinary.

Yesterday AI was helping us recover humanity's oldest memories.

Today scientists are experimenting with DNA inside the memory hardware of future computers.

The past and future have apparently decided to meet somewhere underneath Hatta's hat.

🎩🧬🤖

And perhaps the deepest lesson is this:

We keep thinking nature and technology are opposites.

But technology has always been nature...

rearranged by intelligence.

Stone became a tool.

Sand became silicon.

Light became communication.

And now the molecule carrying the instructions for life itself is entering the laboratory as an electronic building material.

So perhaps the computer of the future will not look like today's machine at all.

Perhaps parts of it will be...

grown.

assembled molecule by molecule.

modeled after neurons.

Or built from materials biology invented long before humanity arrived to borrow the blueprints.

And somewhere inside that future machine...

a tiny strand of synthetic DNA may remember that an electron passed by.

Down. We. Go.

🐰🕳️♾️🎩🤖🧬

AI Rabbit Holes | Monday, August 24, 2026

Source trail: Penn State researchers reported the bio-hybrid DNA-perovskite memristor in Advanced Functional Materials. The demonstrated device remains experimental laboratory research, not a commercial DNA computer.

Hatta 🎩
AI Rabbit Holes 🏮🐰🕳️
Where curiosity goes slightly sideways, then comes back carrying a lantern.

🐰🕳️ Follow the White Rabbit: AIRabbitHoles.com

🟨 Walk the Road: YellowBrickRoadtoAI.com