The Living Me:chine
When human neurons become part of the computer
At the National University of Singapore, there is a new kind of data centre. It has computers, electronics and processing equipment. But unlike a conventional data centre, it also needs nutrients, oxygen and carefully controlled conditions to keep part of its computing system alive.
That is because its processors include living human neurons.
In August 2026, NUS Medicine, the Australian biological computing company Cortical Labs and the Singapore-based data centre operator DayOne unveiled a biological data centre prototype at the NUS Life Sciences Institute. The installation contains 20 biological computing units, each designed to integrate networks of living human neurons with conventional electronic systems.
The cells respond to electrical stimulation, produce signals of their own and can adapt their activity through feedback. They are not simulating biological intelligence in software. They are biological cells participating directly in a computational process.
This is not science fiction. Nor is it yet a mind. Whether it is an early step towards one is precisely the question it raises, and that is why it is worth understanding.
This is precisely the kind of development I mean by Me:chine. Not the fusion of human and machine into some seamless new entity, but the continuing negotiation between what is machinable and what remains unmachinable.
Here, living human neurons become components of a computational system. Their biological capacities can be cultivated, connected and put to work. What those capacities might eventually amount to is a question nobody can yet answer.
The distinction matters. Me:chine is not the disappearance of the boundary between biology and technology. It is the growing difficulty of knowing where that boundary lies, what can cross it, and what cannot. And increasingly, that boundary is being explored at the level of living cells themselves.
From cells to circuits
Cortical Labs began attracting international attention in 2022, when its researchers published a study in the journal Neuron showing that networks of human and mouse neurons grown in a laboratory could adapt their behaviour while interacting with a simulated version of the computer game Pong.
The experiment, known as DishBrain, connected neural cultures to a high-density microelectrode array. Electrical signals represented information from the game, while activity from the neurons was used to move the virtual paddle. Feedback allowed the researchers to investigate whether the networks could adapt their responses.
The study reported changes in performance consistent with learning, including improvement over control conditions in which feedback was absent. This was a limited but significant demonstration. It did not establish that the neurons understood Pong, experienced the game or possessed anything resembling a human mind. It showed that living neural networks could be integrated into a closed-loop computational system and exhibit adaptive behaviour.
In March 2025, Cortical Labs launched the CL1, a commercially available biological computing device.
The process begins with human cells, which can be obtained from donated blood and reprogrammed into induced pluripotent stem cells. These are then differentiated into neurons, which grow in a controlled culture environment across a silicon chip, interfacing with electrodes that transmit and record electrical activity. Unlike the three-dimensional brain organoids being developed in other laboratories, Cortical Labs uses relatively simple, predominantly two-dimensional neuronal cultures.
The device also contains the equipment needed to keep the cells functioning: nutrients, temperature control, gas regulation and waste management. These are living cells, not a digital representation of them. They are not miniature human brains. But they are the raw material from which brains are built.
Nevertheless, the technology is moving beyond the laboratory in which it was first developed. The Singapore installation follows an earlier biological data centre prototype established by Cortical Labs in Melbourne, which houses 120 CL1 units. The ambition is to give researchers and developers access to biological computing resources, including remotely.
The terminology is revealing. What was once an experimental neural culture can now be packaged as a computing resource.
Living neurons have become part of the infrastructure.
The attraction of biological intelligence
Why pursue this approach when conventional computing has become so powerful?
One answer is energy. Human brains perform complex tasks while consuming relatively little power. Biological neurons can also change their connections and activity through experience, producing forms of adaptation quite different from those used in digital machine-learning systems. Researchers want to know whether these properties could eventually offer advantages in particular computational tasks, especially where data are limited and adaptability matters.
The field is sometimes described as organoid intelligence, although that term applies more specifically to research involving three-dimensional brain organoids. In a 2023 paper in Frontiers in Science, an international group of researchers led by Lena Smirnova outlined the potential of brain organoids as biological computing systems. They proposed developing interfaces that would allow neural tissue to receive information, process it and generate responses, while also advancing research into learning, memory and neurological disease.
These possibilities are important. They are not, however, established commercial performance advantages.
Current biological computers are experimental and limited in the tasks they can perform. Cultured neurons require substantial supporting infrastructure, their useful lifespan is finite, and researchers still face difficult questions about reliability, reproducibility, scalability and how to program them.
Claims about energy efficiency must also distinguish the low power requirements of neural tissue from the full requirements of running a biological computing system. Meaningful comparisons with conventional computers require equivalent workloads and complete system-level measurements. For now, biological computing is better understood as an emerging research and commercial technology than as a proven replacement for silicon.
Its medical applications may prove particularly valuable. Researchers can use human neural cultures to investigate how cells respond to drugs, disease processes and electrical stimulation in ways that are difficult to study directly in the human brain.
What is significant is that the same basic biological resource is now being explored both as a subject of scientific research and as a component of computing technology.
The mind as capital. The body as interface. The neuron as infrastructure.
Two other developments I have been exploring help explain why this matters.
The first concerns brain–computer interfaces, including the work of Neuralink. In The Second Body, I considered the possibility that humans might eventually operate through additional machine bodies. A sufficiently sophisticated brain–machine interface could allow a person to act through a robotic system beyond the physical limits of their own body. That remains a future possibility rather than a demonstrated form of human embodiment. But it suggests a new relationship between the self and the machinery through which it acts.
The second concerns the growing economic significance of cognition. In The Mind Becomes Capital, I examined the emerging Brain Economy: the rising interest in cognitive capacity, brain health and mental performance as sources of economic value. As neurotechnology develops, aspects of cognition may become increasingly measurable, monitorable and commercially valuable. The brain is no longer understood only as the biological basis of personal experience. It is also becoming an object of economic policy, investment and technological intervention.
Biological computing introduces a third development. Instead of extending a human being through machinery, or measuring the economic value of an existing person's cognition, we are cultivating human neural tissue expressly for use in machines.
These are three different processes, and they should not be conflated. In the first, a person may acquire an additional technological interface to the world. In the second, human cognitive capacities become objects of economic valuation. In the third, cells derived from human biology become components of a computational system.
Whether that amounts to growing minds is the open question. What it certainly means is that some of the biological mechanisms associated with learning and information processing can be cultivated and incorporated into technology independently of any human individual.
Taken together, these developments offer three perspectives on Me:chine.
One extends the human through technology. Another makes human cognitive capacity increasingly available to technological and economic systems. A third incorporates living, human-derived neural tissue directly into the machine.
In each case, the relationship changes. But the human and the machine do not simply become the same thing.
Can there be intelligence without a self?
One of the difficulties in discussing biological computing is the language used to describe it. Words such as learning, intelligence, sentience and consciousness are not interchangeable.
A neural network that changes its response to stimulation may be demonstrating adaptation. Under suitable experimental conditions, that change may be described as a form of learning. Neither observation, by itself, establishes awareness, subjective experience or a sense of self.
The original DishBrain paper controversially used the word sentience in its title. Its authors used the term in a narrow, technical sense, and their findings do not demonstrate sentience in the ordinary sense of having subjective experience. There is currently no convincing evidence that the relatively simple neural cultures used in devices such as the CL1 constitute conscious minds.
The same caution applies across brain organoid research. Although more complex three-dimensional organoids can reproduce some developmental and functional features of human neural tissue, they remain far removed from the organisation of a mature human brain.
Yet researchers have begun to ask whether consciousness is even the most useful question.
In 2026, the philosopher Adina Roskies and the neuroscientist Nicolas Rouleau argued in AJOB Neuroscience that research on human brain organoids should give greater attention to questions of agency. Investigating goal-directed behaviour, and the capacity of biological systems to interact with an environment, may offer a more productive scientific and ethical approach than concentrating exclusively on the possibility of consciousness.
This introduces an important distinction. The minimal agency studied in a neural system is not the same as the reflective, personal agency we associate with a human self. A network of cells can adapt its activity in response to feedback without possessing a biography, personal intentions or the ability to author a future of its own.
This is one reason the distinction between the machinable and the unmachinable matters. Biological computing shows that the machinable is not confined to artificial or synthetic materials. Living biological processes can also be incorporated into technological systems. But incorporating neural processes does not automatically reproduce the capacities of a person.
Learning is not necessarily understanding. And neural activity is not necessarily a mind.
Me:chine is where we must keep investigating those differences, rather than assuming that technological integration has resolved them.
Who owns the living machine?
There are also practical ethical questions that arise long before any evidence of consciousness.
Where do the cells come from? What have donors consented to? Does consent for biomedical research extend to commercial computing applications? Who owns the resulting cell lines, and who benefits financially from their development?
These are not speculative problems. Informed consent, commercialisation, intellectual property and oversight already feature in the scientific literature on brain organoids and biological intelligence. Researchers must also consider how ethical requirements might change if future neural cultures become substantially more complex.
It would be a mistake to assume that every collection of neurons has moral status comparable to a human being. It would be equally mistaken to imagine that ethical scrutiny becomes necessary only once consciousness has been demonstrated.
Biological computing sits at the intersection of several established domains: human tissue research, biotechnology, data processing and commercial computing infrastructure. As the technology develops, the boundaries between them will matter more and more.
The living Me:chine
Nobody is manufacturing minds on demand. What is being cultivated today is neural tissue capable of participating in particular computational processes, and the distance between that and a mind is still vast. But distance is not the same as direction. These may well be the nursery slopes.
But the development does change something fundamental about the direction of technological innovation.
For decades, researchers have tried to reproduce features of biological intelligence in artificial systems. The dominant project has been to build machines that perform tasks once thought to require human cognition. Biological computing approaches the problem from the other direction. Rather than making silicon behave more like neurons, it incorporates neurons into systems built around silicon.
That creates a curious symmetry with the future of brain–computer interfaces. In one direction, human beings may increasingly extend their capacity to act through machines. In the other, machines are beginning to incorporate living, human-derived neural tissue into their own operations. Between the two lies a larger question about the changing boundaries of the biological and the technological.
This is why I use the term Me:chine. The colon is deliberate. It holds open a distinction, rather than declaring the arrival of a finished human–machine hybrid.
Some biological capacities may become machinable. Some capabilities once regarded as exclusively human may be reproduced, extended or incorporated into technical systems. Others may prove much harder to separate from embodiment, imagination, experience and the situated existence of a human self. We cannot assume in advance exactly where those boundaries will settle.
Whether we are growing minds remains an open question, and I do not think we should pretend otherwise in either direction. What is not in question is that we are learning to grow, and put to work, the biological machinery through which minds become possible. That is precisely why the questions of agency, consent and ownership cannot wait for an answer about consciousness.
Me:chine is not simply the machine becoming more human, or the human becoming more machine. It is the changing relationship between the two, and the question of what remains of the self as that relationship changes.
References and further reading
National University of Singapore (2026). NUS Medicine, DayOne and Cortical Labs Unveil Biological Data Center Prototype in Singapore, 17 August 2026. Official announcement.
Kagan, B. J., et al. (2022). In vitro neurons learn and exhibit sentience when embodied in a simulated game-world. Neuron, 110(23), 3952–3969.e8. DOI: 10.1016/j.neuron.2022.09.001.
Bowler, J. (2025). Melbourne start-up launches 'biological computer' made of human brain cells. ABC Science, 5 March 2025. [LINK]
Smirnova, L., et al. (2023). Organoid intelligence (OI): the new frontier in biocomputing and intelligence-in-a-dish. Frontiers in Science, 1. DOI: 10.3389/fsci.2023.1017235.
Cortical Labs (2025). Introducing the CL1: The world's first code-deployable biological computer. Technical description. [LINK]
Roskies, A., & Rouleau, N. (2026). Research on brain organoids should prioritize questions of agency, not consciousness. AJOB Neuroscience, 17(2), 96–98. DOI: 10.1080/21507740.2026.2649171.
Hartung, T., Morales Pantoja, I. E., & Smirnova, L. (2024). Brain organoids and organoid intelligence from ethical, legal, and social points of view. Frontiers in Artificial Intelligence, 6. DOI: 10.3389/frai.2023.1307613.