Pretty soon, many of us are going to have little to no reference for what things were like before LLMs could talk to you, generate pictures and stories and other art forms, act like therapists, blow through Turing tests, and doing just all of what they’ve been doing to creep us out over the last five years. The uncanny valley, the last little tells of non-human creative origin, will vanish. We will be living side by side with another species, one that has not been evolved in the biological way, but rather, engineered in silicon.

The philosophical ramifications are intense. We deal with this every day in the MIT community, as our best and brightest work on the problems of the day. I work on it, as an event organizer, a partner at Link Ventures, and in collaboration with colleagues like Daniela Rus and Ramesh Raskar, putting together responses, as it were, to the breakneck pace of technological change.

But my other point was: soon, a new generation will have grown up with all of this. And that really makes a difference.

Those of us who are not “AI-native,” in that sense, will have an uphill battle to fight. Those of us old enough to have seen the origin of the Internet, the personal computer, digital color, and cloud computing, all of them big changes in themselves, thought we were presiding over the great era of change. We weren’t even close, because what Back to the Future and other imaginings failed to prepare us for was intelligence out of computing, ex nihilo in a formidable sense, that, as Bill Gates says, was put together in a lab, and developed further on its own.

It seems evident to me that in the years to come, all of us, even us older people, will need to learn to become AI-native. What does that mean?

I moderated a talk at the Next Endeavor event Sept. 14-15, in Mountain View, California, which Imagination in Action and partners put on (disclaimer: I help to run these events) between Denise Lee of Cisco and Joe Cordeira of UBS, a wealth management firm. The three of us explored what it means to be AI-native, for the time to come.

Both of these experts talked about how they use agents.

“I actually have agents working at night,” Lee said. “So in the morning, I have a summary of what I need to do to start my day. I’m very thoughtful about the prompts and what goes into that, so that it’s not giving me generic information, and it’s constantly learning. My agent is very specific to me, and that just helps me throughout every task for every day.”

“I have quite a few agents,” Cordeira said, “and I’m a firm believer in the end-user experience of using agentic AI, and really bringing it all together, and having one orchestrator agent. So I use AI extensively at home, and it helps me accelerate what I do at work as well. So, (I’m) a big AI optimist, I would say.”

With that in hand, we got into the meaning of AI-native.

“We all have this vocabulary and taxonomy that is constantly changing, and I feel like every two or three weeks, there are new words that are added to this vocabulary,” Lee said. “To me, AI-native is almost synonymous with what Silicon Valley has in the water, right? It’s that ideal of: we want to go build, and we want to go change the world. How do you embrace it and figure it out, and do it for good? And that’s easy for someone to say who’s a fourth-generation Silicon Valley person. But every time I travel, I am reminded that (AI) is not native or organic to a lot of other conversations.”

She gave this further defining thought:

“AI-native is always innovating, always thinking about what’s next, and doing it for the better good.”

“If we were to build this organization, this business, from scratch,” Cordeira said, “with the capabilities of AI today, how would we build it? What would it look like? We are definitely not, and I think everyone knows this by now, taking existing processes, existing, you know, paradigms and just slapping AI on them. We have to really think through, front to back, how something works, and reimagine it for the era of AI.”

He had this to say about the human role:

“The human element is really important. How do we blend AI-native with that human element?”

A timely question, to be sure.

Roadblocks to AI Nativity

I asked those in the other pair of seats about the challenges inherent in the process.

“I think at the end of the day, there’s a lot a lot of challenges,” Cordeira said, adding that his company has been in business a long time, long before AI was part of the fabric of modern business. “I’ll go into some of the details, but the biggest challenge is just ‘change leadership,’ driving transformation. If you think about firms like ours that have been in business for 160 years, you know we’ve been successful at doing something a certain way, and that’s afforded us a lot of economic (gain) over the years. But to do that … you find what you’re really good at, and you scale it, and you harden that system. Hardening that system means you’re building technology in a certain way. Your processes are a certain way. You’re hiring certain types of people, right? And you know, we built the business in that way, and now we’re in this new era where we need to reimagine it. And obviously all of those systems are the things that need to be transformed and changed, in some cases, dismantled and rebuilt. It is really the classic innovator’s dilemma here.”

That certainly covers a lot of challenge. Cordeira talked further about sourcing data and tools, and Lee brought up working in legacy environments. Some of the challenges, she noted, are financial:

“You’ve got to make your numbers quarter to quarter,” she said. “Most leaders are looking at: are we going to make our numbers this quarter? As startups, I know that doesn’t happen overnight. That doesn’t happen in a quarter, or even two.”

She noted that she’s been working on a startup inside of Cisco for the better part of four and a half years.

“It’s called Energy Networking Systems,” she said. “Prior to that, it was the engineering sustainability office. It has never been a large team, maybe a two-pizza-box team, as we call them … I was working and managing through the matrix of the organization to get things done – (that’s) where that startup mentality, that sort of AI native mentality plays in.”

Another major challenge, she suggested, is around human pushback: to data centers, to a greater role for AI, to the infrastructure that’s needed for advancement.

“You can’t really go too far without seeing a headline around this,” she said. “It doesn’t matter what part of the stack you’re in, if you’re only (getting) the API software layer. There’s nothing to build on if those data centers aren’t going up … the enterprise has a smorgasbord of challenges and limitations, but it also has a ton of opportunity to go faster.”

To my next question, Lee humored me with a very short answer.

My question: what is the thing that most executives do not understand about trying to make an AI-native company?

“The old legacy way doesn’t work,” Lee said.

Cordeira had a longer answer.

“I don’t think this is a technology problem,” he said. “It’s really not. It’s a people problem - at the end of the day, we’re going to have better and better technology, better agents, better systems around it. The limiting factor is definitely going to be the human side.”

I thought this bit by Cordeira was interesting. He talked about the history of UBS building around AI.

“We honestly thought: if we build it, they will come,” he said. “And we built really, really, really good stuff. I mean, truly game-changing AI capabilities. And we communicated it in a scalable way.”

The problem, he noted, was that financial advisors are inundated with emails and offers.

“(They) receive literally hundreds of emails a day,” he said. “And while we’ve built something really great, they’ve been running their practice for 30 years the same way. And that’s not enough. That’s not enough to get them to change their ways. So we’ve actually had most success with hand-to-hand combat. We need to sit down with them. We need people in the field. We need more humans, right, to help the humans use the AI, and that that has been a successful tactic.”

In addition, Lee talked about energy, and Cordeira talked about regulation.

“I think most of us in what we traditionally call first-world countries take energy for granted,” she said. That’s ironic, since at press time, there’s a threat of grid problems in the mid-Atlantic.

“How many people in this room understand how your local utility commission increases your rates?” Lee asked. “How many people understand, you know, what happens when Google, Microsoft, Meta want to build a data center in your backyard? What does that do? How many people understand what that does to your water? These are things that just from an energy perspective, we all need it, and we all take it for granted.”

“It does slow down progress,” said Cordeira of regulation. “And it’s actually really important. It’s a core principle of how UBS operates. And again, it’s the reason why we’ve been in business for 160 years, because we are safe, right? And we’re regulated, and obviously we take those obligations very seriously.”

Here’s the part where we discussed new humans coming along and inherently embracing the tech.

“AI-native employees are like sponges,” Lee said, “and they are consuming, and they’re learning, but they can’t do it on their own. And they need the (prior) generation … to help them understand the critical thinking involved for the specific jobs. … I think every job is going to be completely transformed with AI. I think that’s every job.”

Already, she said, executive assistants are working with agents, and it’s transforming how they do their work.

“Having a human-driven executive assistant is still going to be necessary for some of the work,” she said. “I cannot do everything I do without a small team of humans working things behind the scenes. There’s critical thinking that is involved.”

With optimism, she described some serendipitous outcomes.

“This is such an exciting time in history,” she said of the new generation. “How do we maximize their expertise and their skill sets? AI is helping with that as well.”

“Everyone should think of themselves, in the future, as a process engineer, not just a processor,” Cordeira added. “And then, when the agentic future is fully here, and you’re one that’s managing agents that are executing a process, that role makes a lot more sense.”

Here’s how Cordeira closed the segment.

“There’s a huge room in this world for experiences,” he said. “AI is going to help us plan them more efficiently, but I think this will continue to have tremendous value to society.”

Think about all of this in the context of where we are today, and where we will go tomorrow.