Even those AI systems as sophisticated as Chad (thats what Ive come to call him) or Claude, depend on the same silicon-based hardware that has been the bedrock of computing since the 1950s.
But what if computers could be molded from living biological matter?
Some researchers in academia and the commercial sector, wary of AIs ballooning demands for data storage and energy, are focusing their enerygy (pun ) on a scientific field known as biocomputing.
This uses real or synthetic biology, such as miniature clusters of lab-grown cells called organoids or electro-active polymer hydrogels (I call them bipolar gels), to build a new computer architecture.
Before we start!
If you like this topic and you want to support me:
- Comment on the article; it will really help spread the word 📢
- Connect with me on Linkedin 🔗
- Subscribe to TechTonic Shifts to get your daily dose of tech 💉
One of those biocomputing pioneers is the Swiss company FinalSpark, which earlier this year debuted its Neuroplatform. That is a computer platform powered by human-brain organoids, which scientists can rent over the Internet for $500 a month.
In an interview with Scientific American, FinalSpark’s Fred Jordan (he is one of the co-founders of FinalSpark), is really excited about how unique their platform is. He says its the first neuronal computing platform thats available to the public like this.
FinalSpark was able to fund this project thanks to the success of their earlier startup, and their goal is to make AI that uses 100,000 times less energy than what we have now.
And that is a pretty big deal!
If only it wasnt alive and darn slow.
Heres how the Neurotic platform works
A little pun…. forgive me
The platform has got these processing units that host four spherical brain organoids each.
A brain organoid is a miniature, and very simplified version of a human brain which is grown in a lab from stem cells. These organoids are three-dimensional, and multicellular structures that mimic some of the features and functions of a developing brain, like the way that the brain is organized, the cell types, and even some aspects of brain activity.
These organoids are teensy-weensy tiny little buggers, only about 0.5 millimeters wide, and they are connected to eight electrodes that stimulate the neurons and link them up to regular computer networks.
To make the neurons learn, they expose them to dopamine, which is the chemical responsible for things like the brains natural reward system.
If everything goes according to plan, these little buddies could end up working just like silicon-based AI and slave away as processing units similar to CPUs.
One of the nifty things about the Neuroplatform is that you can actually watch the organoids in action, 24/7, via live-stream.
But dont bother.. nothing really fancy happens there anyway.
The biggest challenge right now is to figure out how to train the neurons to do what they want them to do. But despite that, a bunch of universities – 34 of them, in fact – are already interested in FinalSparks biocomputers.
The University of Michigan, for example, is one of their early users. And they are looking into how they can change the activity of the organoids using electrical and chemical prompts. If they can crack that code, it could lead to a programming language thats specifically designed for organoids.
A prompt to program your future baby?
Its getting crazier and crazier.. please remind me to talk about some ethical stuff later on.
Briding the gap between biological and artificial intelligence
Meanwhile, over at Lancaster University Leipzig, they’re exploring how to integrate organoids into different AI learning models.
Integrating brain organoids into AI learning models is quite interesting because it could possibly bridge the gap between biological and artificial intelligence. And that will combine the flexible (and adaptive) way that living neurons learn with the speed and scalability of AI.
As said earlier, those Organoids are composed of human neurons and can naturally form neural networks and learn from stimuli in ways similar to the human brain.
As a TTS reader, you already know that our current AI models rely on pre-programmed algorithms and large datasets. But organoids have the potential for self-organization and real-time learning!
Now that could definitely lead to more potentially new forms of AI.
By combining biological intelligence with AI, we might create systems that learn and adapt in more human-like ways, which could then be valuable in applications where traditional AI struggles.
Think for instance of understanding complex, unstructured environments or learning from sparse data. And that is something Embodied AI (AI in a robot) could benefit from, because it could be thrown into a novel situation where it has no reference data.
As always, there are hurdles
Now, as promising as organoid computing is, there are some pretty big hurdles to overcome.
For one thing, theres no standard way to manufacture these systems yet. And the organoids themselves dont live very long. They live on average, only about 100 days. But hey, thats actually a big improvement from when they first started, when the organoids only lasted a few hours!
Jordan mentions that FinalSpark has been working hard to refine their organoid production process, and right now, theyve got somewhere between 2,000 and 3,000 organoids in their facility.
Other types of bio-computing systems
FinalSpark isnt the only player in the biocomputing game.
Theres this guy named Ángel Goñi-Moreno from Spains’ National Center for Biotechnology who has been studying something called cellular computing. That is where you use living cells that have been modified to do computational tasks like memory and logic gates.
Memory and logic gates are fundamental components of any computing system. So naturally you would think of building this into your cellular computing system as well.
He thinks biocomputers could actually outperform silicon-based ones when it comes to environmental sensing – he calls it cellular supremacy.
Cool name, he obiously drew a little inspiration from Quantum Supremacy.
Imagine being able to put a bacterial computer in a polluted lake and have it give you detailed data on the environmental conditions. Now that is a thing which traditional computers just cannot do!
An example of a cellular computer: The incredibly smart Sheldon J. Plankton from SpongeBob.
Fungal computing
And then theres Andrew Adamatzky from the University of the West of England, who is all about fungal computing.
And no, before you ask, you wise ass, he isn’t letting your toes with athletes foot do the math.
He has found out that mycelia (those are networks of fungal strands), have electrical properties that are similar to neurons.
His goal is to create a fungal computing system that can learn and recognize patterns.
Adamatzky thinks fungal computing has some advantages over brain organoids, like being easier to cultivate, costing less, and not raising as many ethical concerns.
And that I am in total agreeance with !
*** Hot off the presses ***
This news came in yesterday: Scientists grew a mushroom into a robot to act as its brain
The researcher behind this work is Rob Shepherd from Cornell University, and his research focuses on developing biohybrid robots that integrate living organisms with robotics.
Shepherds team at Cornell created robots controlled by fungal mycelium, the network of thread-like structures that fungi use to communicate through electrical signals. They embedded the mycelium directly into the robots electronics.
This way, the researchers made these robots convert fungal signals into movement.
So actually this led the fungus to function like a brain that responds to environmental changes, such as light.What was the name of that series again, where a mycelial network infected humans which turned them into living … monsters.
The Last of US !
Thank you !
Lets continue…
They developed two robot types: one with wheels and another spider-shaped with soft, flexible legs. In tests, these robots moved according to the natural signals from the fungus, altered their behavior in response to ultraviolet light, and could also be controlled manually.
This fusion of organic and synthetic elements requires expertise across engineering, biology, and signal processing.
The team sees potential uses in agriculture, where these ShroomBots could monitor soil conditions and decide, for example, when to add fertilizer to reduce environmental impact.
Them fungi could process various stimuli like light, heat, and chemicals, and that could result in more autonomous, adaptable robots. It could also whipe out the human race, if climate change or Benu doesnt beat them to the punch.
More strange types of bio-computing
Okay, so weve covered a lot of ground here, but theres even more to explore in the world of biocomputing!
Lets talk about DNA computing first
This is where researchers use the unique properties of DNA molecules for data storage and processing.
The first guy to show this was possible was Leonard Adleman from the University of Southern California, way back in 1994. He used DNA strands to solve a Hamiltonian path problem, which is pretty cool.
The Hamiltonian path problem is all about finding a path through a graph (a set of points connected by lines) that visits each point exactly once. Its a bit like how optimization algorithms solve the salesmans problem.
But not important for the story.
Lets continue.
If you have had done your homework as you should in biology class, you know that DNA stores data is in sequences of nucleotides (the base pare combos ATCG), and through biochemical reactions it can perform operations similar to traditional logic gates.
One of the big advantages of DNA computing is the potential for parallel computing on a massive scale. You can process millions of DNA strands at the same time in a test tube. Now, that makes DNA computing highly scalable for certain problems. But there are still a lot of hurdles, like high error rates and the difficulty of reading DNA outputs.
Now, Microsoft and the University of Washington have been making some big steps in DNA-based data storage. They have developed a system that can code digital information into DNA! That means you can store a huge amount of data in a really tiny space.
A gram of DNA can theoretically hold around 215 petabytes (215 million gigabytes) of data.
The system isnt quite ready for everyday use just yet, because its still pretty expensive to build and the read/write speeds are so slow. BUT, they are making progress reasonably fast.
DNA computing could end up being a revolutionary technology for specialized applications like combinatorial optimiation (a subject I was working on when I was 14 for some crazy reason), or big data analysis where exploring numerous possibilities at once is needed.
It could also be really useful in fields like drug discovery and genomics, where huge, huge datasets need to be processed in parallel to identify patterns or solutions without blowing the energy grid.
Next up, Bipolar Gels
Nah, they’re actually called electro-active polymer hydrogels, but I prefer my own term.
These gels are soft and flexible and can change shape in response to electrical signals.
What makes them interesting for creating bio-computers is that they can be engineered to respond to electrical stimuli in a way that is similar to biological neurons. This makes them a good candidate for creating interfaces between biological and electronic systems that are more bio-compatible.
This is the future :
There are researchers like Zhenan Bao at the Stanford University that are developing these hydrogels to create soft and stretchable circuits. The idea is that these materials could integrate with living tissues, to be used as scaffolding for nerve regeneration or brain-machine interfaces.
The cool thing about these hydrogels is that they can change their electrical properties based on their environment, which means they can provide a dynamic interface between neurons and electronic systems.
And over at MIT, Timothy Lus lab is combining electro-active hydrogels with genetically engineered cells to create circuits that respond to biological signals. Their goal is to develop living circuits that can operate in environments where traditional electronics would fail, like inside the you and I.
These circuits could create new forms of bio-computing, especially in medical applications where things like real-time monitoring and response are really important.
And then there are researcherers from the University of Toronto who are exploring hydrogels in soft robotics. They are combining them with bio-inspired actuators in the hopes of developing robots that can adapt to their environment.
Bio-what? A bio-inspired actuator is a device that tries to mimics the movement of a biological muscle or a tissues to get robots or mechanical systems in motion. An actuator converts energy (such as electrical, hydraulic, or pneumatic) into mechanical motion to move something. But under stress (like in deep sea environments) these things tend to break apart cause they are made from plastic or metal and use electric motors. But bio-inspired versions are designed so that they are flexible and soft (enter the hydrogel).
The biological musculature of the future:
Bacterial computing
This is where researchers use genetically modified bacteria to do some computational work.
The idea is to use the natural processes of bacteria, things like gene expression and response to environmental stimuli, to create logic gates and circuits.
Angel Goñi-Moreno at Spains National Center for Biotechnology is a key player in this field.
He and his team are designing genetic circuits in bacteria that can perform basic logical operations, like AND, OR, and NOT gates. They modify these bacteria in such a way that they are able to produce specific outputs in response to inputs. This way they develop bacteries that can perform complex computational tasks.
The University of California, led by Jeff Hasty, has also made some contributions to bacterial computing. Hasty’s lab has developed a synchronized genetic oscillator in E. coli.
Escherichia Coli (yes, the poop bacterie) is a model organism in biology. This oscillator can be used to create more complex circuits that could potentially be used in biocomputers.
A genetic oscillator is a biological circuit that produces rhythmic, periodic signals in living cells. That is similar to how an electronic oscillator generates repetitive signals in electronic circuits. These circuits can be designed to perform tasks, like cell behavior, or creating responses to environmental signals.
This could be applied in areas such as synthetic biology, or drug delivery, and of course environmental sensing (toxicity etc.).
How does this work, you might ask yourself?
A gene produces a protein that inhibits the expression of another gene, while that second genes product inhibits the first gene.
That is common in cells.
This creates a cycle where the expression levels of these genes oscillate over time.
And once a protein is made, it might take some time to accumulate to a level that effectively represses or activates another gene. This dely in degradation affect the timing when a gene is turned on or off – and that is the oscillators timing mechanism.
This is so brilliant, people !
Bio-computing research that is waaay out there
And it gets weirder and weirder folks…..
Like quantum bio-computing
This is a pretty speculative field and it wants to use quantum effects found in biological systems for computing purposes.
The idea is to explore biological processes like photosynthesis or enzyme catalysis and to find out if they can be used to perform quantum computations.
Lets go down the rabit hole….
Jim Al-Khalili et al. at the University of Surrey are investigating quantum coherence in biological systems.
Think of photosynthesis in plants.
They think these processes may use quantum superposition to achieve high-efficiency energy transfer. If that is true, it could inform the development of new quantum algorithms inspired by biological processes.
Quantum what to achieve what?
This is somewhat complex stuff, but lets give it a shot: Quantum superposition is a state where particles (think electron, quark, boson, yadayada) exist in multiple states at once (that is called superposition, which is the basis of a quantum computer). In the context of energy transfer, this would mean that energy could take multiple paths simultaneously, choosing the most efficient one to reach its destination.
If this holds true, understanding how these biological systems naturally use quantum mechanics could inspire new quantum algorithms that mimic these highly efficient processes.
Whilst over at MITSeth Lloyds team is looking into how quantum dots—tiny particles made from semiconductors—can be combined with biological molecules to build new types of quantum computers.
By merging these two different technologies, the goal is to create a hybrid system that uses both biological and quantum properties. This could lead to new ways of computing that arent possible with just quantum computing or biological systems alone.
An application of Quantum Dots
Researchers at the University of California, Berkeley, led by K. Birgitta Whaley, are studying quantum effects in enzyme catalysis. They want to understand how quantum tunneling might influence the speed and efficiency of biochemical reactions.
Quantum whatnow?
Quantum tunneling is a cool thing in the weird world of quantum mechanics where particles pass through a barrier that they would not be able to cross according to (classical) physics. If a particle does not have enough energy to overcome a barrier, there is a small chance that it can "tunnel" through it. And that is because of the wave-like nature of particles in quantum mechanics. Yes... wave like nature of particles. Not a typo
Insights from this research could inform the design of quantum bio-computers capable of ultra-fast computations. Think of building molecular scale sensors or quantum information storage (storing your info in Qbits).
In our computers, 1 Tb stores 8 trillion bits.
But Qbits can exist in multiple states at the same time (that is superposition), which means exponentially more info.
A 50 Qbit system represents 2^50 states (1 million billion) different combinations. And DNA computers can store 215 million gigabites, so that is significantly less even.
There is more…..If you havent dosed off by now, here is my last chance…
Because there is more novel research on the horizon.
New Frontiers in Robotics, Synthetic Biology, Biochips, and Cyborg Cells
Ooh, those are some really cool and cutting-edge areas of research in biocomputing and wetware!
Lets get down, and explore them a bit more.
Soft robotics and bio-hybrid machines.
This is all about combining living tissue with synthetic materials to create robots that can adapt to their environment. That is super useful in situations where traditional rigid robots just wouldnt cut it.
Zhenan Bao at Stanford is doing some really neat work in this area, developing soft, stretchable circuits that can integrate with living tissue. This could lead to robots with more natural movement, which is pretty exciting!
And over at Harvard, Robert Woods team is creating biohybrid actuators that combine living muscles with synthetic materials. Imagine how useful that could be for tasks like surgery or operating in unpredictable environments.
And the folks at the University of Tokyo are even exploring how bio-hybrid machines could be used in disaster response, making them adaptable for tasks like searching for survivors or removing debris.
Its amazing how these bio-hybrid systems could revolutionize robotics!
Next, lets talk about cell-free synthetic biology.
This approach is all about using biological components outside of living cells to perform functions like computation or sensing.
The cool thing about this is that it removes the complexities of cell growth and division, making it a bit simpler to work with.
Vincent Noireaux at the University of Minnesota is doing some really interesting work in this area, developing cell-free systems that use ribosomes and enzymes to create synthetic circuits.
These circuits could be used for tasks like protein synthesis, which is pretty amazing. And over at Harvard, Pamela Silvers lab is taking things a step further by merging cell-free systems with electronics. This could be really useful for real-time environmental monitoring, which is super important.
And the researchers at the University of Tokyo are even developing cell-free systems for rapid diagnostic tests, which could be a game-changer for point-of-care applications.
Now, lets move on to biocompatible biochips.
These devices are designed to interact seamlessly with biological systems, which opens up all sorts of new possibilities for monitoring, diagnosis, and therapy.
John Rogers at Northwestern is working on implantable biochips that could be used for long-term health monitoring and drug delivery. Just think of having a tiny chip inside your body that could keep track of your health and even deliver medications when needed!
And Timothy Lus lab at MIT is taking things even further by combining biochips with synthetic biology. This could lead to smart prosthetics and other bio-integrated devices that require real-time data processing.
The possibilities are endless!
Last but not least, lets talk about cyborg cells and synthetic organisms.
This is where things get funky!
These are cells that have been engineered with synthetic components to perform new functions. This is really blurring the line between living and non-living.
Rashid Bashir at the University of Illinois is doing some really cool work in this area. He is developing cyborg cells that can sense environmental changes or even deliver drugs.
Awwww… Our dearly beloved Sheldon J. Plankton, now turned into the Borg:
And over at MIT, Ron Weiss is programming synthetic organisms with genetic circuits that allow them to autonomously produce drugs or degrade pollutants. Thats right, tiny living machines that can clean up the environment!
Phew, that was a lot to cover!
Some ethical stuff
Speaking of ethical concerns, thanks for reminding me (NOT!).
Because that is a big issue in this field.
Because I think that using human neurons for non-medical things definitely need to raise some eyebrows, especially when it comes to the potential for consciousness in lab-grown brains.
Now, there is no evidence that organoids have developed consciousness yet, but FinalSpark is taking this seriously. They are consulting with philosophers and researchers to really dig into these ethical questions.
😥
But despite these concerns, Jordan is convinced that human neurons are the way to go for the most effective learning, which is why hes so committed to using them in biocomputing.
…
Whew, that was a lengthy, and deep dive into the research of bio-computing and wetware!
As you can see, each area is unique in their way to approach the future of computing, trying to blend biology with technology in quite nifty ways.
It is quite amazing to think about how researchers are using everything from DNA and proteins to bacteria and brain cells to create living computers.
And the potential is just mind-boggling: from data storage and environmental sensing to healthcare and even space exploration.
Signing off – Marco
Well, that’s a wrap for today. Tomorrow, I’ll have a fresh episode of TechTonic Shifts for you. If you enjoy my writing and want to support my work, feel free to buy me a coffee ♨️
Think a friend would enjoy this too? Share the newsletter and let them join the conversation. LinkedIn appreciates your likes by making my articles available to more readers.
Leave a Reply