Data-Sleek founder and CEO, Franck Leveneur made a guest appearance alongside Couchbase CMO, John Kreisa on DM Radio’s Really Real-Time Data hosted by Eric Kavanagh. During the interview, Leveneur, Kavanagh, and Kreisa discussed the importance of data management, the role of real-time data, and the future of AI in business.
This segment delves into the importance of in-memory architectures for real-time data processing and discusses the future of AI in business decision-making. Kreisa highlights the role of in-memory architectures in enabling high-speed queries and analysis. He cites use cases like user stores and session stores in gaming and social media applications.
The conversation then shifts to AI’s potential to augment human decision-making processes, particularly in language models like ChatGPT. Leveneur envisions a future where AI becomes more humanistic by incorporating emotions and multi-model interactions. He also explores the idea of these AI models serving as virtual assistants for various business functions.
The guests discuss the need for organizations to establish dedicated roles or teams to integrate AI into their operations strategically. Titles such as “Chief AI Officer” and the concept of an “Office of the President” are two potential changes to the traditional business hierarchy. They acknowledge the potential risks and ethical considerations surrounding AI’s increasing abilities and the need for proper governance.
Overall, the segment explores the synergy between traditional data management architectures and emerging AI technologies. It emphasizes the importance of these advancements while addressing the associated challenges and complexities.
Key Takeaways
- AI is increasingly becoming a part of enterprise toolkits that can troubleshoot and do design work.
- Integrating AI into the workflow will inevitably increase and advance all facets of business, including technology itself.
- Mass adoption of AI in the workplace makes way for new roles and changing traditional hierarchies.
Really Real-Time Data Interview: Segment 3
Host: Eric Kavanagh
Guests: Franck Leveneur (Founder, Data-Sleek), John Kreisa (CMO, Couchbase)
Broadcasted May 23, 2024
Find the full podcast on DM-Radio.Biz Here.
Read Part 1 of this interview:
Read Part 2 of this interview:
Read Part 4 of this interview:
The Importance of In-Memory Architectures
Eric Kavanagh: All right folks. Back here on DM radio talking all things data, real-time data. We’ve got John Kreisa from Couchbase and Franck Leveneur from Data-Sleek. I do want to talk about the power of in-memory. Spinning disc is still out there. There are lots of spinning discs.
Boy this solid state stuff is getting much much better these days. Of course in-memory architectures are designed to be very, very fast. Now, you can only fit so much in-memory, so you have to watch out for that.
You don’t want to overload the system and want some problems. John, I’ll throw it over to you to talk about the importance of in-memory. John, give us some use cases for where that stuff really serves a good purpose.
Use Cases for In-Memory Architectures
John Kreisa: Yeah absolutely. I mean in-memory architectures support really, really high-speed query and analysis functions. Principally, they can speed up a traditional app by having a cache.
If it’s your traditional architecture, as is in the case with Couchbase, we are a memory-first architecture. Why that matters is for specific use cases as you said. If we look at some of our customers, we do user stores. As a user store, that’s a very common use case for gaming and for other kinds of applications that need to serve very, very quickly.
LinkedIn wrote a blog not that long ago about how they’re using Couchbase to accelerate all of the logins into LinkedIn. You can imagine how many times LinkedIn gets logged in on a daily, per minute basis. They’ve incorporated Couchbase into their architecture to give them that speed. That’s really important.
We’re commonly used in gaming to create user stores and session stores. It’s the same thing. The beauty of the Couchbase architecture is that we can both serve as a store and a system of record as an operational database. They kind of get the twofer in that way. It’s that in-memory architecture that enables those kinds of applications and many, many more. That’s a good start on examples.
The Future of Humanistic AI in Business Decision-Making
Eric Kavanagh: Yeah those are excellent examples. Maybe Franck, I’ll throw it over to you. The environment is getting so gosh darn complex. I’ll tell you another–once again weave in some of this AI stuff.
I was talking to a guy from a very, very big manufacturing company here in the US. They have all sorts of obviously very, very intensive use cases and they use the large language model. They asked it a question about something important and it came back with an answer, but it only had a 59% score of confidence. They’re like, ‘well obviously we can’t use that.’
A member of his team said, “Well, let’s ask it why it’s only a 59% confidence score.” They did, and it said “Oh, because these data values didn’t come through correctly and this was wrong and this didn’t show up, etc., etc.” And they’re like, ‘Oh, now we know what to do.’
It’s kind of amazing what this AI can figure out even in very complex environments. It’s another part of the toolkit to do troubleshooting and to do design work. What do you think about that Franck?
The Evolution of AI, Innovation, and Ethical Considerations
Franck Leveneur: Yeah definitely. You know, it’s the knowledge it has. It’s amazing. Like we said at the beginning of the show, it’s just the beginning. It keeps increasing. I saw a few videos of ChatGPT 4 and now they are incorporating visual sound, and they have two AI bots talking to each other.
You can use your camera on the phone. It can look at you and see your hair. Like if you tell it, “I have an interview coming up,” it looks at you and says, “Well, you should redo your hair a little bit.” The guy puts a hat on and it says “Well, I would not recommend wearing a hat for your interview. And the AI voice is laughing.
AI Models Have Feelings, Too?
Now, they are starting to add emotions, like sarcasm. It’s interesting because I guess they’re making the AI more human. It’s going to be more friendly, and people are going to feel connected to it.
That’s the danger because we’re already isolated. We have iPhones. Now if we have a body AI where we can ask any question we want, anytime, we’re going to be connected to that AI all the time. In terms of the business world, I think that’s where it’s going to go.
Innovation Across Multiple Disciplines
Tapping is great but talking is faster. You can ask questions and then it is going to respond to you. This has been in movies for a long time. Minority Report, where they move things around on the screen.
If you look at Marvel Iron Man, where Tony Stark talked to Jarvis, helping him build something. So I think this is where we’re going to go. We’re going to ask a question to AI, something you’re not sure about. Like, ‘Hey, can you analyze this?’, ‘can you analyze that?’, ‘what if I try this?’, ‘what if I try that?’, ‘what do you think about this idea?’ This will be in multiple disciplines: marketing, sales, support, and technology.
Superpower AI Draws Ethical Considerations
I think it’s probably going exponential because this is also going to accelerate technology. We were already going very fast, but now AI is just going to be some superpower that allows us to even create faster, better coding.
The code that ChatGPT creates is just amazing. I mean, you can just copy and paste it. It can also understand your code. You can give it to it, and you can read it. It’s fascinating but also dangerous. We have to be careful where it can go.
If you look at the internet and social media, what happened right there, where a lot of issues with AI are going on, could be the same. Don’t get access to military bases of AI. For example, what if something happens? What if something talks to AI, and he thinks it’s a simulation?
I think there are a lot of protections that need to be put in place. Maybe in business, some people have more money, so they can build stuff that can compete against smaller businesses and affect them. I don’t know, it’s a huge scope right that we have to look at.
Evolving C-Suite Roles to Include AI
Eric Kavanagh: Yeah I almost think, and John, I throw this over to you. Basically, every organization–I’ve thought for a long time and evangelize this concept–you need to have an, I call it, like an information strategy group.
Now you do have things like centers of excellence for example. I was talking to a gentleman from Atlan yesterday. They have something very cool, called the “Office of the President.” I think that is what they call it.
There are a bunch of people who work for the president of the company, which is very interesting. I mean obviously you have direct reports, like you’re head of marketing, and your head of sales, and finance, and so forth. But to have this Office of the President, I find a very compelling concept in an organizational hierarchy.
I think that now more than ever, companies need some person, or group of people who are really looking at these technologies, and understanding them, and working with them, and figuring out how they’re going to patch into existing information architectures. Whether that be for real-time data or stochastic purposes, or whatever.
But you need someone. That used to be the enterprise architect. That was the person who kind of understood how all the pieces come together. This person would fit in this group. What do you think about that idea?
The Emergence of Chief AI Officer
John Kreisa: It’s a good question. I’ll say even just from Couchbase’s standpoint, our CIO is the person who’s leading the charge on cross use of AI within Couchbase. I think for our customers, it does need to be somebody who probably sits above a lot of different functions because of the holistic nature of that. The fact that it’s going to touch customer experiences, it’s going to touch your engineering team, it’s going to touch your marketing. To your point, it really is going to holistically impact the entire organization.
Perhaps, you know an office of the CEO is right or maybe CIO at some orgs. Again, it depends a little bit on how they’re set up a bit. It does need to be–like we all saw the emergence of Chief Data Officers. There will be an emergence of Chief AI Officer. That’s what I’m speculating, but maybe it might be.
The Evolving Landscape of AI Cost and Capabilities
Eric Kavanagh: Well because it impacts everything. When I look at these foundational models, not just the large language models, there are these, what they call small language models, now, and other foundational models. I have to think that there are still other shoes left to drop.
You think about the cost structures. It doesn’t cost very much to use this stuff. Well, that’s because they’re bringing you in, right? I mean that other shoes are going to drop someday, and your enterprise access to ChatGPT will go from what it is now, $10-20-30 bucks a month, to more like $200-300 or something.
I think you’ll start to see some stratification of the different layers and what you get for all that. Clearly, they have captured the imagination of the business side. Everyone in it is kind of going, ‘Oh man what’s going on here?’ And for low-hanging fruit, it’s very, very powerful. I mean it’s not a real-time database engine, but it’s like, what can’t these things do?
At a certain point, you got to start asking yourself these questions. It’s like geez, yeah we have to watch out for the deterministic stuff, that’s for sure. I found they’re actually very good at ETL, for example. Good at doing complex ETL jobs. When you have this Swiss army knife that is so powerful, and right now is pretty cheap, it’s like, man, you better be working with this stuff on a regular basis.
Yeah folks, our third break is coming up here in just a second. If you want to be in the show send me an email info@DM-radio.Biz. That comes right to me. We’ll be back with the podcast bonus segment you’re listening to DM radio.
This interview has been edited lightly for clarity.
Find the full podcast on DM-Radio.Biz Here.
Jump to Part 1 of this interview:
Jump to Part 2 of this interview:
Jump to Part 4 of this interview:
Want to host Data-Sleek on your next podcast? Contact us for our speaker sheet and to set up an appointment.
