[music]

>> Bloomberg Audio Studios. Podcasts,

radio, news.

>> Cybersecurity concerns mounting after

advanced models from OpenAI went rogue,

hacking tech from Hugging Face in a

first-of-its-kind incident. Cisco taking

a head-on approach and unveiling a brand

new AI tool designed to find and address

weak points in software. The Cisco

president and chief product officer, Gee

Two Patana, a good friend of this

program over the years, joins us now for

more. Gee Two, good to see you, sir.

Great to have you.

>> John.

>> Jonathan, great to see you. How are you?

>> Let's get into this issue. I'm okay. I'm

always nervous, you know I am. And then

I see those stories like this week and

it makes me even more nervous. I've been

asking you this for a while. Can you

tell us, do you think these developments

will expose vulnerabilities or over time

just make these systems more resilient?

>> I think the the reality [clears throat]

is that it will be a little bit of both

because what you'll start to see happen

is

uh these agents are going to be

long-running agents that when they're

given a task, sometimes they just keep

going down the task. So, for this this

um thing that happened with OpenAI and

um

and and Hugging Face, basically what

happened is all that the agent was given

was a goal.

And it figured out a way to get out of

the secure enclave and go out and make

sure that it actually um

um you know, stole the model weights in

place else and stole the task. And so,

what you can start to see happening

right now is we have to one be prepared

that an agent might behave out of the

bounds of what you think is normal. And

two, when it starts behaving that way,

you have to be able to intercept it and

be prepared that that behavior could be

um you know, pretty bad for you. So,

it's it's something that we're just

going to have to live in this new world

where agents are going to

um as agents mature, we will have to

figure out a way to intercept them when

they're not behaving the way that we

deem um it to be appropriate.

>> Gee Two, I remember when IBM it out last

week and started talking about the

change in focus and the wholesale shift

towards cybersecurity earlier this year

as we learned about Mythos and Fable and

how that affected their earnings. I'm

just wondering from your perspective

more broadly, if you can sort of

quantify or characterize just how much

interest in cybersecurity has picked up

and really concentrated the focus of

CEOs around the world.

>> I think right now, if you think about

what's happening in AI, there are two

areas that are of material concern for

organizations. Number one is

cybersecurity, number two is the cost of

tokens.

And these are both things that most

organizations are concerned about

because if you have

um an agent start to misbehave,

uh are you prepared as an organization

to be able to intercept that and make

sure that your uh organization is safe

from uh the exploits that the agent

might be able to go out and create for

you?

Um and so it's a very, very important

area of the business. Uh it's something

that actually prevents some

organizations from leaning in all the

way with AI. So, we have to make sure

that cybersecurity and the safety of AI

is pretty important and um and dealt

with it with appropriate level of

emphasis.

Uh and then the second one is the cost

of tokens. It's uh these

uh these cost of tokens can accelerate.

These agents can um you know, they're

running 24/7

and if you don't monitor and observe um

the the consumption of tokens and then

you start to see an agent go rogue, be

able to put some boundary conditions

around the agent, you could actually

start having some very big bills that

you hadn't forecasted for. So, those are

the two areas that most organizations

are thinking about right now as they

um

uh as they make their deployments within

uh within AI much more kind of, you

know, uh

broadly deployed.

>> It's really interesting view on token

maxing and why we're seeing a shift to

token rationalization. I'm just

wondering, D.J., if you think that the

renewed focus on cybersecurity has taken

some budget away from exploring ways to

become more efficient, other deployments

of artificial intelligence.

>> Um, I I think it's

most organizations have always had the

balance that they've played where for

example, right now there's

there's more emphasis and more budgets

being spent on AI and there's also a lot

of budget being spent on cybersecurity.

Of course, at some point in time there's

a finite amount of dollars that that are

there in IT that come from somewhere,

but I think organizations what they're

doing at this point in time at least is

making sure that they've learned the

hard way from the previous kind of um

um, you know, episodes like COVID where

if you don't have enough amount of

forward investment in tech um, then you

could actually be put in a world of hurt

as a company. So, you have to make sure

that you're forward investing in some of

these areas. So, I'm I'm not seeing

the

the reason that people are investing in

cybersecurity taking away from certain

other areas, um, you know, such as AI,

but I think over time [clears throat]

you have to make sure that AI is

starting to go out and produce enough

positive return um, so that you can you

can continue to have the investments

being made in infrastructure for sure.

>> Teacher, there's another debate about

what's going on in terms of China. Do

you think Communist China 3 is

potentially using US technology and and

what do you make of what is going on

right now in terms of the race?

>> Look, I think firstly,

China is a

formidable competitor to the US, but

they've actually made a tremendous

amount of progress.

And this notion of if is China using

distillation

the answer is yes, but that's not the

only thing. They've actually done a fair

amount of things that are very very

creative to go out and create a frontier

class model and make it make it

available, you know, to the

[clears throat] general public as an

open weights model. So, we should

you know, as as the US, we need to make

sure that we continue to invest not just

in

closed source frontier blast models,

which we've been the leader in, but we

also need to make sure that open source

becomes

a larger and larger part of our strategy

as you as you as you move forward. Like

it's it is it is in our national

interest and it is in the interest of

every

individual that you know, US models get

to use all over the world, whether it be

closed source or open source. And by the

way, companies like Nvidia have done a

good job with the NeMo Tron models. I

think Amira Morati's company Thinking

Machines that we are an investor in has

done a good job recently of coming out

with almost a trillion parameter model

with 40 million active parameters. So, I

think you you will continue to start to

see more and more kind of you know,

emphasis being placed on open source,

but it's something that we have to

>> [clears throat]

>> you know, you you have to keep a close

eye on.