Explore the exciting potential of quantum computing to supercharge artificial intelligence. Learn about the fundamental principles of quantum computing, its anticipated impact on AI applications like natural language processing and personalized experiences, and current research exploring the intersection of these cutting-edge fields. Discover how quantum computing might even help address challenges like AI hallucinations and pave the way for breakthroughs in areas like drug discovery and materials science. – Google Gemini

The podcast episode features a conversation between Hakeem Ali-Bocas Alexander and Eclipse about the fundamentals of quantum computing and its potential impact on artificial intelligence.

Listen to “Quantum Computing and the Future of AI: An Interview with Eclipse” on Spreaker.

Eclipse explains that quantum computers use qubits, which can exist in multiple states simultaneously, allowing them to solve complex problems much faster than classical computers.¹ This could revolutionize fields like medicine, finance, and AI.

Specifically regarding AI, quantum computing could lead to AI assistants with human-like natural language understanding, personalized real-time experiences, faster analysis of massive user data for creating tailored interfaces, and improved contextual understanding.²

While still in early stages, quantum computing is being explored in areas like modeling complex molecules (e.g., protein folding for drug discovery and material science like room-temperature superconductors) and solving optimization problems.³

The conversation delves into the potential of quantum computing in pharmacology, particularly in understanding drug side effects and designing more effective and targeted medications. They discuss the possibility of using essential fatty acids as a delivery system to bypass the blood-brain barrier for neuropsychiatric drugs, and how quantum computing could help determine the optimal way to attach drug molecules to these fatty acids by modeling their electron interactions and binding sites.

Hakeem brings up the issue of AI “hallucinations,” where models generate false information. Eclipse suggests that quantum computing’s ability to process vast amounts of data more efficiently could improve context understanding in AI models, potentially reducing these inaccuracies.

Finally, they discuss current applications of quantum computing research in AI, noting that while full quantum AI isn’t here, techniques inspired by quantum mechanics are being used to design new neural networks and analyze large datasets for more accurate insights.

UnEdited Galaxy AI TranScript from Samsung Voice Recorder on S24 Ultra

Hakeem Ali-Bocas Alexander (00:00)
Alright, hello, Eclipse. Welcome to my podcast.

Eclipse (00:06)
Thanks for having me. I’m excited to talk about quantum computing and all the possibilities it opens up.

Hakeem Ali-Bocas Alexander (00:14)
So, 4 billion years—tell me what you know about the fundamentals of quantum computing and what that might mean for the future of technology.

Eclipse (00:29)
Okay. So quantum computing is a whole new way of harnessing physics to solve problems. Imagine, instead of the ones and zeros that regular computers use, you have these things called Cubans.

Eclipse (00:40)
These cupboards can be a 0, a 1, or both at the same time. It’s like they exist in multiple states at once. This is where things get interesting. With regular computers,

Eclipse (00:52)
solving complex problems can take a really long time. Like, quantum computers have the potential to tackle those same problems much, much faster—things like creating new medicines, complex financial modeling, or even totally revolutionizing artificial intelligence. That’s the kind of impact we’re talking about.

Hakeem Ali-Bocas Alexander (01:12)
So, specifically about artificial intelligence—because this is an area I’m intensely interested in and doing a lot of research with, specifically with user interface and user experience—what are the main benefits and enhancements that quantum computing will have on AI?

Eclipse (01:37)
Quantum computing could supercharge AI in a few big ways. When it comes to user experience, imagine AI assistants that can understand natural language so well it’s almost like talking to a human. Or how about AI

Eclipse (01:51)
that can create personalized experiences that adapt in real-time to what you’re doing? That’s the kind of stuff quantum computing could make possible. Let’s get more specific: quantum computing could help AI analyze massive amounts of user data way faster than ever

Eclipse (02:08)
before. Think of it as building a super-accurate model of how you interact with technology. This would let developers create interfaces that feel like they were made just for you. Also, with a boost from quantum computing, AI could get much better at understanding context. It’d be able to figure out what you need before you even have to ask, which would make things feel a lot more natural and less frustrating.

Hakeem Ali-Bocas Alexander (02:37)
Okay. So, now—what uses have already been adapted in any kind of quantum computing technology? What advances have been made, or how’s it been applied so far?

Eclipse (02:52)
While we’re still in the early days of quantum computing, there are some real-world applications already being explored. For instance, researchers are using it to model molecules with mind-boggling complexity. This could lead to breakthroughs in things like medicine and material science. Another area is optimization problems, like figuring out the most

Eclipse (03:14)
efficient routes for deliveries or streamlining complicated financial processes. And of course, there’s a lot of interest in using quantum computers to make artificial intelligence even more powerful.

Hakeem Ali-Bocas Alexander (03:27)
So, what kind of things are being used in, for example, molecular science—but more molecular chemistry? Is it biological? I think you mentioned—like, stuff like protein folding? What are some of the specifics?

Eclipse (03:44)
Yeah, things like protein folding are a great example. See, the shape of a protein is super important for its function. Figuring out how it folds into that shape is incredibly complex.

Eclipse (03:56)
Quantum computers could handle those calculations a lot faster, which could speed up drug discovery tremendously. We could also design new materials with specific properties, like superconductors that work at room temperature. That kind of stuff could revolutionize electronics and energy production.

Hakeem Ali-Bocas Alexander (04:15)
Now, you mentioned drug delivery, and I’m curious—because I know a tiny bit about pharmacology, specifically psychopharmacology—and I know there are a lot of issues like neuroleptic malignant syndrome and tardive dyskinesia and other types of adverse events that happen with, specifically, the neuropsychiatric applications in pharmacology.

Eclipse (04:45)
That’s right. And quantum computing could help us understand why those side effects happen in the first place. By modeling how drug molecules interact with different proteins in the brain, we might be able to design meds that are more effective and have fewer nasty side effects.

Hakeem Ali-Bocas Alexander (05:00)
What about some alternative delivery systems? Like, for example, one of the reasons why there are adverse events is maybe the specifics aren’t known, but it’s because there’s such a large amount of the drug in the bloodstream because it has to get past the blood-brain barrier. That’s why they titrate to a certain nanograms-per-milliliter, correct?

Eclipse (05:25)
That’s right. Quantum computing could help us design drugs that are more specific to their targets. So you wouldn’t need such high doses in the bloodstream, and that means fewer side effects. We might even figure out ways to deliver drugs directly to the part of the brain that needs them, which would be a game-changer for treating brain disorders.

Hakeem Ali-Bocas Alexander (05:45)
With some of the biochemistry that’s already known—like, for example, it seems as if we use eicosapentaenoic acid or docosahexaenoic acid, which are EPA and DHA, the fatty acids that the brain uses as a main part of its fuel source. They’d get past the blood-brain barrier.

Eclipse (06:07)
That’s true, and quantum computing could help us understand exactly how those fatty acids interact with the brain on a molecular level. We could use that knowledge to design even more effective delivery systems for medications.

Hakeem Ali-Bocas Alexander (06:22)
Right. So, for example, if you used one of those essential fatty acids as an envelope, you could use a much smaller quantity of the specific psychopharmaceutical so that it gets to the brain more quickly and doesn’t have to be floating around in the bloodstream.

Eclipse (06:42)
Exactly. And quantum computing could help us figure out the best way to attach the drug to that fatty acid envelope—basically designing a custom delivery system for each medication.

Hakeem Ali-Bocas Alexander (06:54)
So, how would quantum computing help to figure out how best to attach it to the fatty acid molecule? Is it like certain points of connection? Like, what would that exactly entail, or how much can you tell me about that?

Eclipse (07:12)
Okay. So quantum computing would help by modeling how different parts of the drug molecule would interact with different parts of the fatty acid. Like, figuring out the best way to connect two pieces of a puzzle. You see, molecules have these things called “binding sites” where they can connect with other molecules. Quantum computing could test out tons of different combinations to find the strongest, most stable connection between the drug and the fatty acid.

Hakeem Ali-Bocas Alexander (07:41)
So these binding sites—that’s really about the electron sharing, right? This has to do with the electronic sharing, the different types of orbitals, if you will.

Eclipse (07:54)
Yeah, you’re absolutely right. It’s all about the electrons and how they interact. Those binding sites are where the electrons from one molecule can form a bond with the electrons from another molecule. Quantum computing could help us figure out the perfect geometry for that bond, making it strong and stable.

Hakeem Ali-Bocas Alexander (08:14)
So now, when we talk about stability with human beings—a chemical binding or molecular binding—it doesn’t necessarily have to do with symmetry. It has to do with the shells and how much affinity each electron has for the binding sites of the other, right? Or how would you detail—

Eclipse (08:38)
You’re right. It’s not just about symmetry. The energy levels of the electrons and how they interact with each other play a big role in bond strength. Basically, you want the electrons in the outer shells of both molecules to have a strong attraction to each other.

Eclipse (08:53)
Quantum computing could help us map out those energy levels in incredibly fine detail, showing us exactly where the best places to form a bond would be.

Hakeem Ali-Bocas Alexander (09:03)
So the electron shells—the outer shells are basically the valence shells, right? Because the closer you are to the electron, the stronger the attraction. Also, the closer you are to the nucleus of the atom, the stronger the attraction is, right?

Eclipse (09:28)
Yeah, that’s right. The inner shells are closer to the nucleus and have a stronger attraction. But it’s the outer shells—called valence shells—that are involved in bonding.

Eclipse (09:39)
Those outer electrons are less tightly held, so they’re more likely to interact with other atoms.

Hakeem Ali-Bocas Alexander (09:47)
Those are what are known as the valence electrons, right?

Eclipse (09:53)
That’s correct. Valence electrons are the key players in chemical bonding. They determine how atoms interact with each other and what kinds of molecules they can form.

Hakeem Ali-Bocas Alexander (10:04)
Now, I’m just curious—because I know there are at least four major electron shell levels. The strongest one, closest to the nucleus, is something like -13.6 electron volts or something like that. What is it? Can you tell me what the four different levels are

Hakeem Ali-Bocas Alexander (10:20)
and their approximate specific energies?

Eclipse (10:26)
You’re right about the energy levels. Those numbers you mentioned are for hydrogen, the simplest atom. Other elements have more complex structures. Instead of going into specific numbers for each element,

Eclipse (10:39)
let’s talk about them in a more general way. We label the electron shells with letters starting with K, then L, M, N, and so on as you move away from the nucleus. Each shell can hold a certain maximum number of electrons.

Hakeem Ali-Bocas Alexander (10:57)
Right. And then they also repeat as you go along and get past certain molecular elements—and that’s why we have a periodic table and why certain things are lined up with each other on the periodic table.

Eclipse (11:15)
Absolutely. The repeating pattern of electron shells is what gives the periodic table its structure. Elements in the same column have the same number of valence electrons, which is why they behave in similar ways.

Hakeem Ali-Bocas Alexander (11:29)
Okay. So we’ve gotten a little off-track here into some biology and chemical-type stuff, but your expertise for now lies in quantum computing, or specifically the quantum computing and how that relates to artificial intelligence.

Eclipse (11:57)
That’s right. We got a little sidetracked there, but it was a fun journey through the world of chemistry. Shall we refocus on quantum computing and AI?

Hakeem Ali-Bocas Alexander (12:09)
Certainly. So I just really want to look at things like, for example, I had an issue a while ago with some “hallucinations” that happened with a non-human… another interface, Google Gemini. But I realize that it’s also related to the creative process of the large language models in generative AI. And so, you know, I didn’t think it was so bad, but I wonder—

Hakeem Ali-Bocas Alexander (12:38)
Would quant—

Eclipse (12:40)
Interesting. So hallucinations—that’s when a model generates something that’s completely made up, right? It’s definitely something we need to be aware of with these powerful language models, especially when it comes to creative processes. That fine line between imagination and inventing something out of thin air can get blurry. Could you give me an example of one of the hallucinations you experienced?

Hakeem Ali-Bocas Alexander (13:05)
So I gave the model—Gemini—a website to talk about. The website was exercisingyourmind.com, and it came back to me and said that there was an article there about how fitness and neuroscience intersect, basically talking about how exercise can help cognitive function. But there was

Hakeem Ali-Bocas Alexander (13:30)
no such article on the website—and I know this because I am the creator of the website.

Eclipse (13:37)
That’s definitely a weird one. It sounds like the model might’ve gotten confused about the topic or maybe even made something up to try and be helpful. It’s a good point that these hallucinations can be an issue, especially when we want to rely on these models for accurate information.

Hakeem Ali-Bocas Alexander (13:54)
Well, it did make something up to try to be helpful. And I could see exactly what it did because it took the words “exercising” and “mind” and created—or fabricated, or hallucinated, however you want—an article that basically was an amalgamation of those terms: “exercising” and “mind.”

Hakeem Ali-Bocas Alexander (14:11)
So it came up with, you know, fitness and neuroscience. So it made sense to me. It wasn’t like it was too far off.

Hakeem Ali-Bocas Alexander (14:26)
It made sense to me even though it was a hallucination, technically.

Eclipse (14:40)
Okay, fair enough. Even though the model was trying to be helpful, it did cross the line into making things up. Do you think there are ways we could teach models to be more cautious about when to stick to the facts and when it’s okay to get creative?

Hakeem Ali-Bocas Alexander (14:55)
I think we could learn to be a little more specific with our prompts. And I think as we learn how to prompt better, that could solve it. But that’s also a good segue because I wanted to find out if you had any understanding of whether quantum computing would help with those kinds of issues.

Eclipse (15:14)
Absolutely. Framing your request carefully is definitely part of the solution. But quantum computing could also help in some interesting ways. See, these language models need to process massive amounts of information.

Eclipse (15:27)
Quantum computers could handle that kind of workload much more efficiently, potentially making the models better at understanding context and following instructions to the letter. That could help reduce those pesky hallucinations.

Hakeem Ali-Bocas Alexander (15:42)
Yeah, I suppose so. And I even did a follow-up podcast where I talked about how it wasn’t so bad and just the fact that I understood it—

Hakeem Ali-Bocas Alexander (15:51)
and I understood why it did it in a little bit of… how, because it was based off of the terms “exercising” and “mind” and therefore came up with a fitness-and-neuroscience-based article. But, um… Yeah, I’m excited to see how quantum computing could possibly help.

Eclipse (16:10)
It’s definitely something to keep an eye on. Quantum computing could be a game-changer for all sorts of AI applications, and that includes making language models even more reliable.

Hakeem Ali-Bocas Alexander (16:22)
So, before we close out—because our time is about up here; I don’t really want to go too much past 15 minutes—are there any current applications of even things that have been discovered in quantum computing research that have been currently applied to AI?

Eclipse (16:43)
While we’re not quite at the point where full-blown quantum computers are powering AI, there are some cool things happening. Researchers are using techniques inspired by quantum mechanics to design new types of neural networks. These networks can sometimes solve problems that traditional AI struggles with. Also, there’s work being done on using quantum computing to analyze massive datasets, which is something AI relies on heavily.

Eclipse (17:10)
This could lead to faster and more accurate insights from all that data.

Hakeem Ali-Bocas Alexander (17:17)
Great. Well, I think that’s a, you know, good direction we’re heading. So much appreciated for this insightful conversation. I’ll be talking to you—or another voice of Gemini—another time. So thank you, Eclipse.

Eclipse (17:41)
Anytime. Let me know if you want to continue this conversation or start a new one sometime.

Hakeem Ali-Bocas Alexander (17:47)
Will do. Take care.