Dear Friends – Please find a short clip from the Atomic Podcast with Ed Heinbockel co-founder, president, and CEO of SavantX. https://lnkd.in/gAD_m47y
Our conversation explored applied Quantum and AI methodologies for solving real world problems. I rarely come across a leader who has used and developed Quantum methodologies with real-world results.
🎤 The full podcast discussion can be found here: https://lnkd.in/g7SE93zB
➡ We discuss gate-based quantum and quantum annealing. SavantX is using quantum annealing for solving complex compute issues and gaining efficiencies not possible with traditional approaches.
➡ I have heard quantum will revolutionize, quantum is not ready for prime time, quantum is the future, quantum is a pipedream and quantum is too finicky to be a stable source of compute power.
➡ Ed and his team break through this hype and confusion to apply D-Wave quantum annealing ground-level up using AI and hardware design to get to the heart of port container handling, AI safety and validation, advanced simulation.
I hope you enjoy our discussion as much as I did!
Thank you, Ed, for your time and insights!
Thank you, KTTPMedia Media for your ongoing excellent work for the Atomic Podcast! https://lnkd.in/gsjPTp7Z ➡
I can get some, you know, one hundreds and I can build out some really great servers and some compute, GPU compute or I can go to D wave annealing process. Why would I go one way or the other? And you're probably going to say it's the problem we're trying to solve. I understand, but. In terms of actual compute horsepower and the ability to to get through and sort of supercharge a parallel process, what is it that? One gives versus the other. Yeah, that's a great question. So let me answer it by saying that I'll give you a use case. So at Port of LA, where we implemented our first quantum process in production, two shifts a day, six days a week kind of thing. What we were doing was one process out of many that our hone engine was, was was basically making calls to. So think about a port, is it moving containers around this one, 1.5 million containers a year. And the idea is you need to efficiently. Offload them from a ship, you need to put them in the yard, you need to do it in a thoughtful manner. So you minimize it all gets down to minimize the number of times you have to touch and move a container because every time you move a container, it's expensive. So if you can get a container in and out of there, which is 2, maybe 3 touches as opposed to five or six, it's a huge savings. And so one discrete process. We had a pull down menu of processes that we identified as good candidates for quantum and we took the the one that we thought would, you know, we could. Would be most useful initially and we took appointing trucks would come in and then which truck would go to which crane to be loaded and these big going to call RTG rubber tire gantry cranes are moving back and forth across these rows trucks would come in. And then we would look at we you know, you know, here's a, here's an example to answer your question directly. So if you have 30 trucks in queue, that's a 30 factorial problem. If you're going to run that on a supercomputer HPC, it literally would take you do run the numbers, it would take you weeks or months. And we were able to run it on the D wave. Full cycle time of 15 seconds. That's a really good example of needing a compute done in a in a in a sufficient amount of time, so it actually has utility in the real world.
Great interview with Ed Heinbockel!