Are you ready to use a computer that contains living brain neurons melded onto silicon chips that ... [+] gives you advanced AI?gettyIn todays column, I analyze the recently announced bio-computer that contains a melding of everyday silicon chips with living brain neurons and has caused quite a stir. The sensational aspect is that the computer includes actual neurons working hand-in-hand with AI-based artificial neurons. Its a powerful one-two punch that highlights the future of AI.The added twist, as if one were needed, is that such a brain-in-a-box would seemingly be exorbitantly expensive and cost a proverbial arm and a leg to purchase. Nope, not so. This beauty of eyebrow-raising leading-edge high-tech is being priced at a modest $35,000 and the vendor pledges deliveries to start in three months (its just the cost of a brand-new VW Jetta, Chevy Colorado, Mazda 3 Turbo, or Ford Ranger, but a lot more futuristically practical).Break the piggy bank and find any spare change in your sock drawer if you are keen on owning and using a consumer-available early-on working product emblematic of neuromorphic computing.Lets talk about it.This analysis of an innovative AI breakthrough is part of my ongoing Forbes column coverage on the latest in AI, including identifying and explaining various impactful AI complexities (see the link here).The Nature Of Neurons And Neural NetworksBefore we jump into the particulars about this new bio-computer, it might be helpful to make sure we are on the same page when it comes to the nature of neurons.Im sure that you already know that the human brain contains neurons. They are the crux of how we compose thoughts and perform thinking. We have a lot of neurons, numbering around 86 billion or so. Its not just the number of neurons that makes things work. The neurons are variously interconnected and send signals back and forth. The estimated number of such neuronal connections is about 100 trillion and the connections themselves are referred to as synapses. This is a simplification but a generally apt way to depict the human brain.When you think about things, such as how your day is going, the neurons transmit on a biochemical basis and send a form of electrophysiological signals throughout portions of the collective set in the brain. Since the neurons are generally interconnected, it is fair to refer to this as a type of network, a neural network (NN).A catchy way to refer to the biological neural network in our noggin is to say that it is our wetware. Thats a witty way to recast the words of hardware and software to describe the living neurons and the biological neural network that does our thinking.One of the greatest mysteries facing humanity is how the activation or firing of our neurons relates to our ability to think and somehow gives rise to a sense of sentience and consciousness. Active research is trying earnestly to crack the code, see my coverage at the link here. Maybe in our lifetimes, we will figure this out.I sure hope so.AI And Artificial Neural NetworksSwitching gear, consider how modern-day AI works such as the advent of generative AI and large language models (LLMs).Via a computer, we model or simulate a kind of artificially faked neural network by mathematically and computationally using data and numbers to represent a semblance of neurons. This is not on par with real neurons. It is a far cry from how real neurons work. The computer version of a neural network is a limited mimicry and not actually how living neurons are.I usually go out of my way to refer to AI as employing an artificial neural network (ANN) and artificial neurons. The reason I emphasize that naming is that too many people seem to fall into the mental trap that the AI-based ANN is working on par as does a wetware neural network. It is a misleading assumption that unfortunately, the AI industry tends to sneakily allow to occur, doing so by referring to artificial neural networks as being neural networks (dropping out the word artificial).This is confusing and confounding.Anyway, the overall idea is that maybe by using computational neurons working in a faked or artificial neural network, we will get close to having AI that thinks. Nobody can say for sure whether this will be the case. It could be that we might need to find some other means to get computational AI to be on par with human thinking. For more of the nitty-gritty on how ANNs are at the core of generative AI and LLMs, see my explanation at the link here.One nice benefit of this existing path is that we can at least use ANNs to try and potentially learn about the real NNs that exist. Despite the ANN being a contrived variation, it still offers insights to some degree about what might be going on with actual NNs (see my analysis at the link here).Connecting To Human BrainsSuppose that we want to connect to a human brain and leverage the living neural network that a person has in their head.How might we do so?There has been quite a lot of progress lately in inventing brain-computer interfaces (BCI), also known as brain-machine interfaces (BMI). Ive covered the Neuralink company of Elon Musk and noted their various advances in connecting to human brains, see for example my discussion at the link here. Other companies are doing likewise.The notion of BCI or BMI is that we can access a living brain by having a type of portal, similar to having a USB port on a computer or smartphone. We could opt to send electrical or electrophysiological signals into the human brain, get the brain to activate, and then read the signals that arise.Via the portal, we could just read what signals are happening if thats all we wanted to detect. The portal also allows for sending in signals along with reading the signals.Id like to repeat my earlier remark and note that we still do not know how the living neurons and the wetware neural network give rise to human thought all told. Right now, we are only able to determine the most rudimentary aspects. Kudos to those pushing ahead to figure this out.Brain-in-a-box As Ultimate GoalHuman brains live inside peoples heads.I realize thats an obvious statement. Sci-fi movies have often sought to portray that we could remove a persons brain and have it working in a vat. The brain would seemingly still function as it did while residing in the head of the person who kindly gave up their brain (the movie plotlines usually entail whether the person did so voluntarily or under duress).No such possibility exists currently.If we cant get a human brain to exist outside of a human head as a fully functioning mind, maybe we could take a different course of action toward achieving a said-to-be brain-in-a-box.One approach would be to ramp up artificial neural networks to the level of being an exact replica of a human brain. A computer that had this ANN would essentially be a brain-in-a-box. An intriguing angle is whether we could then copy a human brain into the computer-based version and likewise copy the contents of the ANN into a human brain. See my mindful conjecture on this heady topic at the link here.Regrettably, we arent close to any of that.Sad face.Melding The Real With The ArtificialLets think outside the box. Theres a fabled idiom that if the mountain wont come to Mohammad, then Mohammad might need to go to the mountain.Heres how that applies.What if we could place living brain neurons onto silicon chips, so that the computer with its artificial neurons could be melded with living neurons, residing intimately together?Thats a clever and quite interesting first step.It seems to be the best of both worlds.At times, the computer could lean into the living neurons to help solve problems and do some form of thinking. The AI with its artificial neurons could be communicating with the living neurons. Signals sent to the living neurons would activate them, they would respond with signals, and those signals would feed into the computer and then into the artificial neural network. The ANN then performs various problem-solving and answering of questions.Round and round this goes.It is a twofer, a wetware living neural network that works collegially with an artificial neural network.Nice.A Somewhat Scary PropositionYour immediate concern might be where in the heck will they get the living brain neurons from. They cant just be scooped out of a persons head (well, you get what I mean). No worries. They can be grown in a lab. They are synthetic living neurons. They were not inside a persons brain.I trust that provides some relief. Nobodys living neurons are being hijacked out of their head when they are least suspecting it (though, we dont know what the future might hold).This reminds me of the urban legend involving a person who goes to a bar, gets drugged, wakes up, and sees that an internal organ such as a liver or spleen has been removed. The same tale can be enhanced to suggest it is the brain that gets removed, but that wouldnt be much of a story since the person would no longer be alive (until, someday, when we perfect brain transplants).Getting back to the neurons grown in a lab, there is an ongoing debate about the ethics of this practice.The upside is that those synthetic neurons can be used in a bio-computer, allowing progress in melding a semblance of wetware NN with ANN. No human being is harmed or disturbed. The downside is that some contend that growing neurons is not a proper thing to do morally, especially if solely to blend with a computer. Some might be open to going the synthetic neurons route if somehow those grown neurons were implanted in a persons head for reasonable and safe medical purposes.Threshold Not Yet ReachedThe number of synthetic neurons that are being melded with silicon chips is quite low right now. I mention this to clarify that the compilation of wetware neurons is not rising to the level of sentience or consciousness, as far as we can tell. They are toylike since it is such a minimal amount.A challenging ethical question reappears once the number of synthetic neurons starts approaching the amount that a spark of human thought or thinking truly might arise. In that case, you can certainly envision why some would be upset at this approach. Creating a new brain, as it were, synthetically and that no human has ever had in their head is unnerving and disconcerting.What rights should such a synthetic brain have?Thats a can of worms that society has not yet dealt with.You might find of keen interest that the launching point for a compilation of synthetic neurons when it transforms into a kind of human-like consciousness or state of thinking is referred to as the minimal viable brain (MVB). Techies might recognize that this verbiage is akin to a start-up company that makes a minimal viable product (MVP), showcasing what product is under development and represents the minimum requirements for viability, see my coverage at the link here.Where does MVB first arise?No one can say for sure.Synthetic Neurons And EmbodimentIve got another provocative question for you to contemplate.It goes like this. Some believe that human brains are reliant on our senses to adequately develop into a full form of human thought. In essence, because you utilize your eyes, ears, sense of touch, limbs, and other biological elements, the brain develops based on those inputs.A synthetically derived set of living neurons presumably has no such bodily embodiment. Ergo, some assert that those neurons arent going to be good at doing anything of substance. They lack having been embedded inside a human body. Without this, they arent going to suitably develop and reach a state of sentience.Can synthetic neurons that never have experienced embodiment be of avid use or will they inevitably be of little or near-zero use?Right now, no one knows the answer.Grab yourself a glass of fine wine and mull that over.Energy Efficiency Off The ChartsThere is something immensely astounding about the human brain that you might not have especially considered.Our brains are incredibly energy efficient for what they accomplish.An adult human brain consumes around 12 watts, while a standard light bulb uses about 60 watts. A light bulb merely emits light and gives off excess heat. A brain thinks. All that thinking is taking place with a pittance of energy. Admittedly, it is still an energy hog when it comes to the functioning of the human body, but on a relative basis, it is energy efficient, as I will elaborate on next.The reason this is significant is that you probably know that the AI industry is currently using up a tremendous amount of energy while running present-day AI systems. A rush to build data centers with thousands upon thousands of servers is underway. The energy demand for these data centers is through the roof.The bottom line is this. Those synthetic neurons residing on silicon could be a big energy saver. The AI in the bio-computer could calculate that the artificial neural network would consume some Z amount of energy, while the synthetic neurons could do the same processing for a lesser amount than Z.In that case, let the synthetic neurons take the workload.Some underscore that this could be a notable boost to sustainable computing.Where This Is HeadingThe melding of living neurons onto silicon chips requires rethinking how computer processing is going to best be arranged. When should the onboard wetware neural network be used? When should the AI artificial neural network be used? Whats the proper balance of the two means of processing requests and having the computer solve problems?Great questions.New ways to describe these arrangements are beginning to emerge.I earlier mentioned that this is a movement toward a brain-in-a-box. That wording though doesnt aptly depict the current situation and thus some insist it is a misnomer. Ive noted that the relatively scant number of onboard wetware neurons does not yet amount to a brain and referring to this as a brain-in-a-box is premature.Another piece of terminology is to say this is neuromorphic computing. That is yet another wording that creates some heartburn. A computer that is purpose-built to run artificial neural networks is commonly classified as neuromorphic computing. If you had real neurons, the phrase doesnt capture that aspect that the computer is more than ANNs alone.One phrase that might catch hold is to refer to this as synthetic biological intelligence (SBI). Plenty of other new phrases are popping up daily. For example, perhaps a bio-computer of this kind could be a wetware-as-a-service. Thats catchy. Try this next one for size, we could say that we have a biological neural network server stack. And that the bio-computer has in vitro living neurons and in silico artificial neurons.If you like, come up with your own phrasing and see if it will gain traction.Good luck.The Upcoming Bio-ComputerThe company that is making and selling the new bio-computer is Cortical Labs, based in Australia and the computer is coined as the CL1. Their tagline for the CL1 is that it is "the worlds first code deployable biological computer.Since the bio-computer contains living neurons, the system comes with an environmental containment component that is intended to keep the neurons alive for up to 6 months. This provides a self-contained life support capability. No need to have you contend with trying to nourish or ensure that the living neurons remain alive. The bio-computer takes care of that for you. A touchscreen displays the status of the environmental containment and the synthetic neurons.Other than that, the rest of the computer is akin to a conventional computer system, including USB ports, a built-in camera, and so on. Just make sure to tell your kids not to mess with the bio-computer since they could potentially accidentally upset things. It is probably best to keep your rambunctious dog or curious cat away from the bio-computer too, lest your beloved pets inadvertently knock the bio-computer off a table. Etc.The firm and its researchers have worked on this setup for several years. They announced in 2022 that they were able to have an early version that played the popular video game, Pong. In a published research paper in 2023, entitled The Technology, Opportunities, And Challenges Of Synthetic Biological Intelligence by Brett J. Kagan, Christopher Gyngell, Tamra Lysaght, Victor Cole, Tsutomu Sawai, and Julian Savulescu, Biotechnology Advances, 2023, they made these salient points (excerpts):Advancements in hardware, software, and synthetic biology (wetware) have resulted in new methods for interacting with in vitro biological neural systems.The most advanced of these have sought to embody these neural systems into simulated environments to elicit dynamic goal-directed behavior, referred to as Synthetic Biological Intelligence (SBI).SBI systems can be broadly defined as the result of intentionally synthesizing a combination of biological and silicon substrates in vitro for the purpose of goal-directed or otherwise intelligent behavior.SBI is distinct from brain-computer interface (BCI) and similar approaches as it does not involve whole organisms, using only specific biological material, usually neural tissue derived typically through synthetic biology processes, as a biomimetic material within the larger system.They provide a specialized operating system, their Biological Intelligence Operating System (biOS), which crafts a kind of simulated world for the living neurons, sending signals to them and getting signals from them. The impulses of the neurons essentially impact this simulated world.The Future Is In This DirectionWhats your reaction to a computer that encompasses living neurons and can be kept at home or in the office?Shocking and disturbing might be your initial reaction. Or perhaps, instead, altogether fascinating and exciting.I bring this up because these kinds of setups are usually only available at a high cost and are in high-tech labs of sizable companies. Consumers can now enjoy the same capability. Sure, $35,000 is a steep price for a home-based computer, so the odds are that your neighborhood isnt going to be suddenly inundated with them. The price though is low enough that tech-related start-ups might acquire them and find uses that are not yet already envisioned.I would wager that this is going to be an avid direction for furthering advances in AI. Think of the remarkable possibilities. A lot of issues though present themselves, especially scale. Can the complexity and volume of the synthetic neurons be increased to a level that makes them more effectively utilized?If the size and complexity turn out to be an insurmountable bottleneck, the capability would indubitably remain at a low ebb. The interest and usage would be narrow, and somewhat imprudent other than for basic research pursuits.One final thought for now.It seems that we are daily bombarded with assertions that we are all living in a Matrix-like world. We are merely simulations. None of us is real.Do you think those synesthetic neurons are thinking the same thing? Maybe one neuron is saying to the other, hey you, do you have a feeling where you are not sure if youre awake or dreaming?Thats the first question Ill ask if I end up trying out the new bio-computer. I promise to let you know what the living neurons have to say.