Are computers alive?

Before you answer — take the question seriously for a minute. Your brain is an electrical machine. Your computer is an electrical machine. A computer can copy itself, mutate, evolve, learn, and draw power. A flame can grow, eat, and reproduce. So where is the line?

Here is the honest scientific answer, up front: no — a computer is not alive, by every definition scientists actually use12. But the parallels are real enough that an entire research field exists to argue about exactly where they break. Tap the numbered markers for sources.

1. The checklist

In 2002, the biochemist Daniel Koshland published an attempt at a universal definition of life — built so it could cover artificial and even extraterrestrial life, not just Earth biology. His seven pillars, abbreviated PICERAS1:

NASA's working definition draws the same boundary in different words: life is "a self-sustained chemical system capable of undergoing Darwinian evolution"9. Note the two quiet killers inside that sentence: chemical and self-sustained. And the classic trap case — fire grows, consumes, and "reproduces." It scores roughly one of seven above. Nobody alive argues that fire is alive, which tells you a checklist isn't the whole story — but it tells you where the gaps are.

2. Neurons vs circuits

Your laptop has more than twice as many switches as you have neurons. Tap each stat.

The raw counts make the computer look like it wins. The architecture says otherwise, in four ways:

Neurons are slow but massively parallel — and they rewire themselves. A transistor flips a billion times a second and never changes what it's connected to. A neuron fires a few hundred times a second, and its connections physically strengthen, weaken, grow, and prune every day of your life. Learning is literally the hardware reshaping itself.

The signal is electrochemical, not just electric. A nerve impulse is a wave of charged ions — sodium and potassium — pouring across a membrane, with neurotransmitter molecules physically crossing synapses. Electrons never carry the message alone; chemistry carries it halfway2.

There is no hardware-software split. Your computer stores its programs in memory and executes them on a processor — two different places. In a brain, memory is the wiring. Change the strength of a synapse and you've changed both the data and the program at once. No neuroscientist has ever found the brain's hard drive.

A transistor is a switch. A neuron computes. One neuron can integrate thousands of inputs and perform operations a whole circuit of transistors would need to simulate. The comparison "neurons = transistors" flatters the transistor and insults the neuron.

3. Electricity and blood

Leon put it best: compare electricity and blood. Both are the body's wiring — but only one of them is alive.

A wire that carries power to your GPU carries only power. A vessel carrying blood carries the fuel, the building materials, the maintenance crew, the waste disposal, and the immune system — all in the same tube, going in every direction at once. Your heart pumps about 7,500 liters of that a day through 100,000 kilometers of vessels your body grew itself.

That last part is the kill shot for the comparison: a body builds its own wires and seals its own cuts. Blood clots. Vessels regrow. Nerves re-sprout. A computer's wiring never heals — a nicked trace is broken forever until a human technician replaces it. Electricity is the delivery truck; blood is the delivery truck, the cargo, the factory, and the road crew.

"An organism's astonishing gift of concentrating a 'stream of order' on itself and thus escaping the decay into atomic chaos."— Erwin Schrödinger, What Is Life? (1944)

That stream of order isn't just electricity. It's electricity plus matter that rebuilds itself.

4. The hard cases

The checklist cleared computers off the table. But three contenders deserve a real hearing — and one scientist built from a computer file.

Computer viruses. They replicate. They mutate. In the wild, variants compete and the fitter spread — genuinely Darwinian. But they consume no energy, run no metabolism, and cannot exist without a host computer. Most biologists file them with biological viruses: half-alive at best.

Things that "learn." An AI model that adjusts its weights is adapting — but within a single lifetime, not across generations. It never rewrites its own program, never builds offspring, never dies to make room for variants. That's adaptability, not evolution.

Tierra. In 1991, ecologist Tom Ray built a virtual world where small computer programs bred and competed for memory4. Real Darwinian evolution happened in the machine: his original program was 80 lines long; within a day, 22-line descendants dominated — parasites that hijacked other programs' copying code had evolved from nothing. Richard Lenski, the microbiologist who ran the longest-running evolution experiment in biology, tried the successor system Avida and said it "had the smell of life"5.

The computer that became a cell. In 2010, the Venter Institute synthesized an entire bacterial genome — designed on a computer, built in a lab — and used it to boot up a living, dividing cell6. The first organism in Earth's history with an entirely synthetic genome. That one is alive — and the fact that we can see the boundary clearly there shows the boundary is real, not semantic.

And then there's the philosopher Carol Cleland's warning: maybe we're asking the wrong question. She argues we can't define life before we have a theory of it — the way nobody could define "water" before chemistry could say "H₂O"3. We are, right now, in the water-before-chemistry era of understanding life.

This post summarizes peer-reviewed findings for a general audience. It is not a claim about machine consciousness or rights.

5. My take

A computer is not alive. It has no body it maintains, no offspring it births, no chemistry that flows against entropy on its own account. By the checklist, it fails the pillars that matter most — the self-maintenance ones.

But here is why the question deserves to be asked: the computer is the best mirror we have ever built for life — and mirrors don't breathe. Every pillar it passes makes us more precise about what the failing pillars actually mean. We learned what regeneration is by meeting a machine that lacks it. We understand evolution better because Tierra evolved in a box we could watch. The machine doesn't get to be alive. It does get to show us what alive is.

Sources

• [1] Koshland, D. E., Jr. (2002). The seven pillars of life. Science, 295(5563), 2215–2216.
• [2] Azevedo, F. A. C., et al. (2009). Equal numbers of neuronal and nonneuronal cells make the human brain an isometrically scaled-up primate brain. J. Comp. Neurol., 513, 532–541. (≈86 billion neurons)
• [3] Cleland, C. E., & Chyba, C. F. (2002). Defining "life". Origins Life Evol. Biosph., 32(4), 387–393. (The water-before-chemistry argument.)
• [4] Ray, T. S. (1991). An approach to the synthesis of life. In Langton, C. G., et al. (Eds.), Artificial Life II. Addison-Wesley. (Tierra.)
• [5] Ofria, C., & Wilke, C. O. (2004). Avida: A software platform for research in computational evolutionary biology. Artificial Life, 10, 145–156.
• [6] Gibson, D. G., et al. (2010). Creation of a bacterial cell controlled by a chemically synthesized genome. Science, 329(5987), 52–56. (JCVI-syn1.0.)
• [7] UCLA Brain Research Institute: ~86 billion neurons, ~100 trillion synapses; brain ≈20 W (2% of mass, ~20% of body energy). (Reputable explainer; not peer-reviewed.)
• [8] NVIDIA Blackwell B200 GPU — 208 billion transistors across two chiplets, unveiled at GTC 2024. (Industry source.)
• [9] NASA astrobiology working definition: "a self-sustained chemical system capable of undergoing Darwinian evolution" — see Chyba & McDonald (1995), Annu. Rev. Earth Planet. Sci. (Peer-reviewed formulation.)