🥊Knocking Climate Change Out
Energy Generation x Quantum Computing
Everyday we take in 10,320 Quintillion joules of energy from the sun.
If we were to leverage 0.1% of that energy, that would be enough energy to power the globe for 2 years😦
But we aren’t doing this 👉🏾 Solar Panels are only 15% efficient
The maximum efficiency being 33%, which is because of the material they are made up of:
So no matter how much energy it takes in — it will only be able to convert 33% of that into usable electricity.
By designing & simulating new materials that solar panels could use instead, we could make solar panels significantly more efficient.
We’ve use AI algorithms, & generative design to simulate aircrafts that are cheaper or faster — so shouldn’t we able to do something similar for materials❓
Because materials are significantly more complex — taking billions of dollars and insane amounts of computational time to generate a new one.
Materials are made up of particles that follow the fundamental laws of quantum mechanics — and so using a quantum device>classical computers would be ideal.
Quantum computers leverage quantum mechanics to solve incredibly difficult problems.
For example, if we took a maze & gave it to a classical computer to solve — it would go through every possible path to find the correct one.
But a quantum computer would take the problem, and be in a superimposed state of all the possible paths in the maze, finding the solution significantly faster.
🔑Using quantum computing, we can simulate new materials to create more efficient solar cells.
This can be done using an algorithm called a variational quantum eigensolver [VQE]
- Simulates energy levels of a molecule
- Passes it to a classical computer
- Solves an optimization problem where it optimizes that molecule [based on our intended parameter — those related to energy efficiency]
- Outputs optimum molecule/material for our purpose
Energy Storage x Artificial Intelligence
Lithium ion battery pack research to make batteries cheaper has stayed stagnant.
Just like before, the name of the game is economics.
If it is cheaper to use energy storage methods (batteries) that don’t use fossil fuels, only then will it be used.
$20/kWh — is where we need to get to, to make that happen.
- right now we’re at $160/kWh
- predicted to be at $94/kWh in 2024
- predicted to be at $62/kWh by 2030
Still 3x more expensive than where we need to be.
Battery research is slow. Battery research is slow because of the process that we are using.
- Research for different compounds
- Synthesize them to experiment on them
- Experiment to see how conductive they are
This same process as was used in the 1900s.
BUT we now have more access to data, greater computational powers, advanced ML — yet still using the same approach❓
Using ML we can test 1000s of compounds in just minutes.
Leveraging neural networks — 2 NNs
- Structural data is taken in (like average lithium anion distance in different electrolyte compounds) and used that to predict the conductivity of never before seen chemical compounds.
2. By using another ML model which takes in elemental information (ionization energy, electron affinity) of the component parts of electrolytes we can predict how chemically stable the compounds will be.
NN filtering compounds based on conductivity + NN filtering compounds based on chemical stability = go through 1000s of compounds/min and accelerating battery research to make them cheaper.
Food Production x Cellular Agriculture
For every 100 pounds of food we feed a cow, the cow retains only retains 3 pounds of food for us.
3% efficient technology
In the next 2 minutes, you’ll learn about how the current biological approach taken to produce meat can be disrupted using technology to produce the SAME MEAT but with
- 98% less water
- 99% less land
- 91% less greenhouse gas
This starts with myoblasts — they have the potential to become muscle tissue but they are not muscle tissue yet.
We get more of these cells, & they proliferate because they have an extracellular matrix.
Because cells are anchorage dependent, they need a sort of protein structure (e.g. collagen) that helps them communicate & continue proliferating.
Then the ‘fibrinogen growth factor’ or FGF comes into play which convinces these cells to continue growing, until we run out of FGF and then the process switches.
We start fusing these cells together — through a process called myogenesis — creating the formation of myotubes. Myotubes are the basis of muscle tissue.
Many myotubes together = tissue.
put through biological incubator [cow — gives it the FGF, proteins involved in myogenesis, extracellular matrix]
Cows are the food machines.
But they aren’t optimized to do this.
Cows are meant to reproduce & survive, which means they can do this process but they’re not good at it.
97% of what we feed them doesn’t go towards sustaining this process.
So if cows are machines — why don’t we create alternative machines that are actually optimized to do this = cultured meat.
- factors — FGF, myogenesis factors, nutrients (carbohydrates, lipids…)
And a cultured meat system provides them both of these things.
- acquire stem cells
- use medium (which has all the factors they need)
- give them the right structure (using gelatin, insect scaffolds…)
- and get the…
Traditionally, if you wanted a burger you’d have to kill a cow to get your burger.
But with cultured meat all we need is a tiny portion of a cow’s stem cells [alive & painless process], culture them using the steps above and then we get our hamburger which is and tastes like any other, with our cow still alive & healthy.