What’s Next is a monthly conversation between SDxCentral CEO Matt Palmer and a senior-level executive from the technology industry. Each month, Matt has an informal but in-depth video chat with a fellow thought leader to uncover what the future holds for the enterprise IT and telecom markets – the hook is each guest is a long-term acquaintance of Matt’s, so expect a lively conversation.

This month, SDxCentral CEO Matt Palmer spoke with Vijoy Pandey, SVP of Cisco’s Outshift group, which specializes in new and emerging tech. Cisco has named Pandey their “technology visionary for a software-focused AI-first digital future,” which is a long but fitting title for a person who holds more than 80 patents in cloud and artificial intelligence/machine learning (AI/ML), among others, and who was head of engineering for Google prior to Cisco.

Editor's note: Below is a sample of the wide-ranging conversation between Pandey and Palmer, edited for length. To hear the full conversation, be sure to watch the video.

On modern software as a distributed application

Matt Palmer: Vijoy, thanks for joining us. It's been so long, and I'm thrilled to have a chance to chat with you a little bit about apps and services.

Vijoy Pandey: If you think about all modern software, all modern applications are really distributed applications. They're not monoliths anymore. They're not running on bare metal or VMs. There might be open-source components that you might pull in. You might be developing some components on your own internal development, other APIs. And increasingly, you're picking APIs from a whole host of SaaS (software-as-a-service) providers that provide specific capabilities. You bring all of those things together, even mobile APIs, all these things together to build your modern distributed application.

And that's one of the problems that has been exciting me quite a bit over the past few years, because in this kind of a distributed approach the lines get blurred, right? Whose responsibility is it to secure your application? Whose responsibility is it to manage trust behind customer data? So it's been a really exciting time to look at this space.

On generative AI (genAI)

Palmer: I'd like to double-click a little bit more on the generative AI, and would love to get what your take is on the productivity and the product in terms of datasets and, whether you own them or not, how are you able to use that to train the models and use that today?

Pandey: That's a question that everyone is struggling with. First and foremost, there's this massive hype cycle happening. But there is truth behind it. Yes, there is hype. But there is reality, and there is going to be a step function change because of generative AI in our lives. Starting with NLP (natural language processing (NLP)) being the de facto UX. But then also, there are whole new paradigms that are coming about because of generative AI and the whole industries that are going to get disrupted because of generative AI as well.

If you think about an enterprise, Cisco being one, but everybody that I talked to – I've talked to more than 40 customers at this point – is struggling with this, which is both product and productivity. And you're looking at foundation models. Whether it's GPT-4. For whether it's hosted by Azure or it's hosted by OpenAI. You're looking at open-source models, which are getting better and better over time; and again, those could be hosted on the cloud or on prem. So then, a whole slew of these foundation models — private tenant, public tenant, open source, hosted or on prem.

But you're also looking at a wave of customization, because that's where differentiation is going to happen. You might use a bunch of foundation models for general-purpose use cases, the productivity use cases. But the customization is where the value is, and then the applications and the prompts are where the value is. And how you integrate that with your application and chain them together is where the value is.

So you're looking at this landscape almost no differently than the multicloud landscape, but like 100 X. And so you look at this and think, how do I manage this complexity? And how can I prevent data leakage, privacy, protection, everything around responsible AI, everything around security and data normalization and cost management. So how do you deal with all this is something that everybody's struggling with.

On the importance of good data

Palmer: What advice do you have for our audience about the key to getting started and having good data? How should they look at getting started from where they're at today?

Pandey: We all belong to organizations where data is scattered. They’re data puddles. And so there are data puddles that are everywhere, and government regulations are not doing us any favors by making sure that those puddles remain puddles. And so the way to think about this is, how do you build an architecture, build guidelines, so they can all be pulled together to drive insights?

There are networking datasets. And then there are security datasets. And how can you pull them together to drive something that is bigger than just networking and security? It starts with clean architectures, cleanliness of data, making sure that you can normalize it the right way.

But then there are other pieces that you need to layer above it. And one of the things that Cisco is really passionate about — given our mission statement is to power an inclusive future for all — we built a responsible AI framework.

But more than that, we need to protect our customers’ data and customers’ IP, because we are in our customers’ environments. And so is our community. I mean, if you think about all of our customers, they are in their customers’ environments, and we need to be really, really responsible.

So that’s the gamut. Making sure that they can talk to each other so that we can drive some value above that. And that can happen by building some new use cases first and then going after boiling the ocean. And then the responsibility piece, which is, to us, really important.

Watch the video for the rest of the conversation between these old friends, who also happen to be tech visionaries.