Lets dissect this article and see what it means shall we…
It says $318.8 million Series C funding. The reality is that they have nothing, absolutely zilch. So this is basically seed funding over and above their earlier seed funding of $67.6 million. From where we stand they are going to raise another $500m before shutting shop.
They claim “XtalPi’s platform hosts over a hundred prediction algorithms” “We believe AI holds the answer to solving pharma’s productivity challenge”
Let me tell you something you might not want to believe because you believe in fluff being peddled around. Deep Learning at its core is basically function approximation and function fitting using data. Its like regression.
Automatski has proven in the past that Deep Learning CANNOT offer solutions to scientific problems that require accurate solutions in the complexity spaces of 2^1000 or 2^100,000. Why? Because it is based on function approximation and curve fitting with existing data. Thats why! And even if it hits the target with an accurate solution for small problems or a percentage of problems on “test” data. It cannot offer any assurance of reliability. And will never work in the production environment.
If any of the Deep Learning nonsense actually had even the remotest chance of working in Drug Discovery. Everyone with billions of dollars would have trained a multi-billion parameter deep neural network and would have done it. You know why nobody could do it? Yes! Because it doesn’t work.
So when they claim they have 100’s of A.I. algorithms on their platform. They basically have nothing. Or else they would have been valued at a trillion dollars. If they had even one, yes thats right, even one algorithm that worked in the scenarios of accurate scientific solutions. They would have been valued at trillions of dollars. Which Automatski has 100’s of solutions of millennium problems and thats why Automatski is valued at $275 trillion in 2019 money.
“Across several prominent global cloud providers, XtalPi routinely builds a supercomputing cluster of over a million cores on-the-fly, in just a few hours, to enable dozens of drug discovery and design projects in parallel.”
They are using classical supercomputing on the cloud also for their solutions. But even super computers cannot solve problems of molecules with 1000’s of electrons and AFAIK viruses and diseases have molecules with 10’s of 1000’s of electrons. The worlds largest supercomputers stop at ~100 electron structures. And using approximations which destroy the very purpose of the simulation there are research papers which report calculations of ~900-1000 electron structures. I bet you didn’t know that, did you?
So basically using Classical Supercomputers for Pharma & Medicine (aka Drug Discovery) is basically a whole lot of “GAS”
““Quantum physics and quantum chemistry methods form the core of XtalPi’s technology platform, which involves first-principle calculations whose accuracy is crucial for depicting intramolecular simulations in crystal structure prediction calculations and free energy simulations.”
There is just a teeny weeny bit of a problem with this statement. If you think carefully. The problem with Quantum Physics is that it requires a Quantum Computer (unless you are using O( 2^(n3) ) classical methods like Monte Carlo or other sampling based methods. Which obviously are useless). And nobody in the world has a Quantum Computer worth anything. The best anyone, and I mean “anyone” outside Automatski has is ~100 qubit erroneous junk. We know this, because, of various scientific reasons about the near impossibility of building hardware quantum computers. And because Automatski is the only, I repeat “The Only” Company which has a billion qubit infinite precision quantum computers since 2014. And we can use it for billion qubit planks scale simulations of Quantum Gravity. 10,000-50,000 Qubit Hamiltonian Quantum Physics and Chemistry for us is like Easy Peasy, been there done that. We are sitting on our tech just because we have not been able to fund the creation of a datacenter to run our machines at infinite precision. And offer it to our customers publicly.
If the claims in this article were true, they would have created a Coronavirus Vaccine and the cure to 8000 diseases. And they would be valued at trillions of dollars. If investors buy their gas then its upto them. Its a free world after all.
Honest knowledgeable people can try to explain the truth and show the way. But no one can fix stupid.