The AI Capability Gap: Why Humans Are Falling Behind | James Garner

Full Episode Transcript

This is the full recorded conversation from The Human Protocol Podcast, published by The Human Protocol Group. © 2026 The Human Protocol Group. Please link to and attribute the original episode when referencing this conversation.

THE HUMAN PROTOCOL (00:01.611)
The real risk isn't what AI can do. It is what we are not ready for, right? This is the human protocol, the podcast about staying human while building the future. I am your host, Mykel Salomon. Today's conversation is about a gap that is growing faster than what most people realize: the capability gap. Not between companies, not between countries, but between what AI can do and what humans are ready to understand.

adopt and trust. My guest today is James Garner, head of AI and data, and working at the front forefront of AI agent orchestration and real world AI systems. James, my friend, welcome to the Human Protocol.

James Garner (00:48.088)
Thank you so much. It's a pleasure to be here.

THE HUMAN PROTOCOL (00:51.315)
It's a pleasure to have you on. And to before we kick things and we get down to business, James, I would like to start with maybe you telling us a little bit more about yourself. I'll let you go as personal as you want it to be, just for the audience to know who John who James is and what's keeping you awake at night in a way. What's what's you're passionate about lately?

James Garner (01:13.624)
Yeah, good question. So a bit of a strange background. My background is not in tech at all, actually. I am something called a quantity surveyor, which is a profession from the UK. Basically, the cost management of construction works is what I kind of did my degree in and spent 20 years, two decades delivering. And I was lucky enough to deliver all kinds of projects around Europe, mainly.

but mainly in the higher education sector. I work with a lot of universities. So Oxford University, Imperial College, a lot of the Oxford colleges delivering incredible world-class buildings. you know, my original plan was to spend a couple of years in industry and then go back into academia to talk about technology in construction and in project delivery. And I totally failed at that because 20 years later I was still delivering projects.

but I never really fully scratched the itch, if you know what I mean. And I was given the opportunity to start moving into a more technical role within my current company, the company I've been with for quite some time, to help sort of deploy technology throughout that company, which was around the same time as AI and ChatGPT came to the fore, even though...

We were experimenting with AI all the way back in sort of 2016. So I got, there's a story about that, which I can come to later. So it's been a funny kind of route. I've kind of arrived at where I wanted to really be, but 20 years later, but it's actually worked out really well because now I've got a very, and I think this is something that a lot of people miss out a very, good understanding of the domain I'm trying to work in and the problems they're trying to solve. Cause every industry will have different issues to deal with.

in my particular industry, which is mainly construction, infrastructure, energy, there are major blockers. We're not the only industry to have these blockers, but coming into the tech side and deploying the tech, deploying AI solutions, deploying agents with that knowledge of knowing the frustrations and the blockers and the issues you've got to resolve has actually helped me in quite good stead. So even though it's been a non-traditional, quite strange path to where I am, that's kind of how I got there.

THE HUMAN PROTOCOL (03:37.399)
That's a very interesting background, James. Thanks for sharing with it with us. And I'm sure, you know, as with anything in life, where you're coming from and the different backgrounds, it's really you bring a different perspective of how you are absorbing AI and technology. So very much looking forward to this conversation with you today. And if you guys stick with us, we're gonna be covering some of the critical shifts that we believe we are seeing, which is the capability gap.

James Garner (03:40.844)
Yeah.

James Garner (03:57.72)
Thank you.

THE HUMAN PROTOCOL (04:07.639)
the move from chatbots to physical AI and what that means, and the emotional pushback that comes with it, right? So stay with us and I'm sure we're gonna be going over with that. And to to start this conversation, James, what do you think is this capability gap for our people to understand and our audience to understand it? What do you think is this capability gap that we're seeing today with AI?

James Garner (04:36.599)
So what we're experiencing at the moment is unlike anything we've experienced before. So not the change, not the fact that we've got a new technology that we're all grappling with. That's always happened and is nothing new in some sense. The difference is the speed at which it's moving is just unparalleled. We've never had anything like this before. If you go back to other technologies, say the printing press or even the internet, you you're looking at a

quite a long period of time for a technology from the invention of a technology to that technology being embedded into society. know, sometimes take the internet. I think it's about 17 years between when it was invented in the early eighties to when we could say it was mass adopted using explanation marks. Those people not on video. So the difference with the AI, well, this, this flavor of AI, I like to call it the generative AI revolution is that's probably happened in a number of months. So

What that means in terms of the capability gap is that whilst there is untold amount of money being invested in this technology from everyone and everyone's on the bandwagon, everyone's deploying AI in some sense, you know, whether that's properly or not is another question, but people are saying, yes, I'm adopting AI, I've got an AI strategy. So they're adopting AI.

But the capability of the people within the organization to use that AI is not going at nearly the same pace. So we're having this gap and it's getting bigger and bigger all the time. If you imagine a scale, you know, a graph, you've got this kind of hockey stick thing going on with the adoption of AI to the point that nearly everyone's adopting AI. And then at the same time, the capability gaps kind of crawling and that gap's getting bigger and bigger. Now this is important because when

THE HUMAN PROTOCOL (06:09.815)
Mm.

James Garner (06:28.883)
the AI chatbots first came out, they were really easy to use. There wasn't a huge amount of training to learn how to use them properly. It was pretty much common sense. Yes, you needed some help with prompts and things like that, but we're talking about natural language. So the difference between how to use the tool and the capability was quite small. We're now moving into AI agents, AI orchestration, all these other kind of...

paradigms which involves AI. And there's a lot of people who haven't even made it past that first rung. So now it's becoming harder and harder for those people to get on to the ladder. And I fear my fear is that's going to get worse and worse. And then we're going to have this separation in society between the haves and the have nots. And the frustrating thing is it doesn't need to be that way. Really, everyone can get onto this. There's no reason for

doesn't matter what position you are, what company you are to be able to upskill yourself. So it's a frustration for me really.

THE HUMAN PROTOCOL (07:33.471)
It and I share with you that frustration and I also I'm I'm also seeing this same pattern and this same behavior and this gap kind of, you know, increasing in a way because the technology is moving so much faster than really what our potential on the human side is kind of picking up to, right? We're not picking up the pieces of how AI on the technology on the technology side is accelerating because we are not keeping up with our pace of learning and adopting.

James Garner (07:48.695)
in

THE HUMAN PROTOCOL (08:02.993)
this new technology. And this is what we're talking about today. And it's not good. Every time that you have a gap so wide, and the wider it gets, the worse. Because how good is the technology if you cannot have mass adoption on it? And this is what we're seeing. This is what I'm frustrating as well. Because even though there is a lot of conversations around AI, I'm still seeing a huge, huge gap between the what the technology can do

And the adoption rates that we see on the human side of things.

James Garner (08:34.731)
Yeah. And you know, the weird thing is that it's kind of uniform. You know, people have done these studies of, where are these gaps occurring? Is it, it in older people? Is it in younger people? Is it in corporate companies? Is it in small companies? And I'll be honest with you. I mean, I've read lots of studies. I'll read all the studies for the newsletter that I do. But in my actual experience, there is no, there is, doesn't seem to be any.

major pattern that it seems to be spread out. You know, you'll get some older people who you would say should be like technophobes doing amazing things with AI. And then on the other side, you've got people, know, at university or your college who are massively anti AI. You might have heard recently the recent speeches by some of the tech founders at the universities where they were booed on the mention of AI. You know, there's this weird

THE HUMAN PROTOCOL (09:24.671)
Mm. I I saw

James Garner (09:26.219)
Yeah, I mean, it's what that happening is so crazy because you wouldn't expect that you'd expect the younger people to be embracing it. And actually they they're seeing a threat to their future and they're pushing back, you know. So that's the weird thing for me is that there is no pattern in society that you could you used to be able to say, you know, this is a more young person's game or offers game or you've got to be part of a big company to adopt it or you've got to be.

a sole trader to be able to adopt it. It seems to be relatively uniform across the industries and across the age groups in terms of the old Reed Hoffman, Duma, Zuma, Bluma, Gloomah categorization. You know, I haven't seen a huge difference.

THE HUMAN PROTOCOL (09:57.343)
Mm.

THE HUMAN PROTOCOL (10:09.953)
Yeah, it's it's the same. I'm I'm I'm with you on that. And I think to add to it, the the more widespread things are and the gap is, for example, it's the more difficult it becomes to tackle something. And it should not be this way. You were mentioning the news that that you probably a lot of people have seen this in in the news, you know, the college graduates actually booing the speakers in every mention of AI.

James Garner (10:36.855)
Yeah.

THE HUMAN PROTOCOL (10:36.863)
And you know, it's that's not what you expect from that generation, because usually they are more subject to, you know, adopting news changes and adopted to technology. But I think to add to it, James, it's just the end of the world, the fear-mongering news, abundance of news that are out there. And that's what I want our people and our audience to focus on really what is it that we need to do? And what are the steps that we can take today to start kind of bridging that gap?

James Garner (10:52.525)
is love.

THE HUMAN PROTOCOL (11:05.271)
'Cause that's really what we need to focus on. AI is not going anywhere, but we ne we have a lot to catch up on.

James Garner (11:12.801)
We do. if you take that example, you know, I actually blame the institutions because they wouldn't be getting the booze if they weren't getting very mixed messages from the universities or the institutions. They're being fed from these institutions that AI is a bad thing, right? And I've spoken to, we've had professors and people on our...

THE HUMAN PROTOCOL (11:28.905)
Mm-hmm.

James Garner (11:33.42)
our podcast and don't get me wrong, there's some of them who are really good, but a lot of them are hamstrung because they want to adopt AI, but they're dealing with institutions that were built for a world that no longer exists, right? A lot of these universities have been around for some of them hundreds of years, hundreds of years. You go to Oxford thousands of years, you know, so they find it hard to adjust and they're kind of given very mixed messages to their students around the use of it. I know this for a fact because I've got a who's at university in the UK.

And you know, the messages they're getting is, you know, almost, it's almost like a dirty word, the use of AI. Everyone knows they're using AI. And the students probably know that the professors and the teachers are using AI to mock the homework, but no one kind of says it. And then when you get to this, this weird thing happens, right, it's so wild because it's like a sin. It's like you don't speak about it. Everyone just knows. But

THE HUMAN PROTOCOL (12:19.063)
Yeah.

THE HUMAN PROTOCOL (12:22.891)
It's like a sin, right? It's a sin to

James Garner (12:28.749)
I did speak to a couple of professors recently, we had them on the podcast from, think it was, can't remember which university, might have been Westminster University in the UK, who said that, you know, they get really stuck because they then say, well, the only way we're going to really be able to test whether a student understands what they've been taught is to bring back good old fashioned exams within vigilators, I think you call them proctors in the US, you know, actually proper

people testing what they're doing, but then they get that, that gets pushed back because they say, well, you know, the students are now the clients and that they talk about things like anxiety from having to have proper exams. So they call, implement these exams, which would probably be the answer to the question, because to me, there's no problem in using AI to learn things as long as you are learning it. And you're not just using it as a way to basically to pass off,

assignments and learning as your own. So it's the way that you use it. But they going back to the point around these students booing. I blame the fact that they're probably getting really mixed messages. So the institutions need to do something about this. They need to be very clear about the strategy around AI and to not kind of turn it into a dirty word. I think they need to embrace it and say, is okay. And give the students good guidance.

Because even my son, says, I'm not quite sure what I should do. How much do I use it? And he says to me, dad, how much do I use? I say, yeah, absolutely, you should use it. But use it, for instance, in a way that makes you understand it. So I was given in some hints recently on his assignment, which is you can use it to help you with some of your research. But then when you've done your assignment, put it back into one of the AI tools, something like Notebook LM from Google, and get it to quiz you. Get it to ask you questions to make sure you really understand it.

THE HUMAN PROTOCOL (14:18.433)
To push you, to challenge you, exactly.

James Garner (14:20.447)
Yeah, to challenge you exactly. And then you're using AI to really kind of solidify the learning that you've been through rather than just saying, you know, damn, I'm late for an assignment. I'm just going to pass that off. So it's not as simple as like AI is bad. It's how do you use it? How do you deploy it?

THE HUMAN PROTOCOL (14:37.13)
It's hot.

It is key. And we it's not only with with AI, right, James? It's like really with any technology. At the end of the day, it's the use that you make out of it. And people are used to think that AI, you know, when I hear when I hear people talking about, hey, people or students are getting dumber and dumber because they're using AI and they're afraid that they're not exercising, you know, the the brain as it used to be when they used to

James Garner (14:46.221)
course,

THE HUMAN PROTOCOL (15:08.747)
you know, when we need to memorize everything, I keep thinking myself is that that's not the only use case for AI. AIs can challenge you. You can use AI to assist you on, you know, challenge you on the things that you know and the things that you don't know. It's all about how you use it. And this is what it's key.

James Garner (15:26.699)
Yeah. And it's about responsibility. Yeah. It's how you, so you can take two approaches to it. You could say, you know, AI is bad because I'm not learning properly from it. And I've, that's your, your choice to use it in that way. Or you could be the person that says, you know, I am going to look at this like I've got a 24 seven best professor that you could ever imagine on tap.

that I can ask any question to, you know? So it's all about attitude. It's like, what attitude are you going to have with AI? Are you going to have the attitude that's a negative attitude that looks at the things that it can be used for as a harm to you? Or are you going to look at it as a way to supercharge you and turn you into a super superhero effectively?

THE HUMAN PROTOCOL (16:12.683)
Yeah. And and and you know what's funny, James, since we are talking about this gap and you're saying we're not seeing a pattern, it's everywhere, it's widespread. You know, when we bring this challenge to the business world, right, and organizations adopting or trying to implement AI, we also see a gap. We also see a gap sometimes from leaderships not understanding what this technology is about, what this technology is capable.

James Garner (16:20.3)
and

James Garner (16:36.435)
Yeah, so much.

THE HUMAN PROTOCOL (16:39.343)
And and what it takes to really implement AI. So I'm curious to hear your thoughts about what do you think leaders are underestimating right now?

James Garner (16:49.772)
Yeah. mean, everything we just talked about with that individual kind of student, you know, that gets timed by 10 or 20 when you bring it into the boardroom, because now you're dealing with probably 10 or 20 different personalities, all of which are going to have a different view, all of which are going to have a different opinion on AI. And that can often lead to just kind of stagnation because you can't find a clear way through. So it goes back to what I was talking about, the Reid Hoffman framework, which I really love. I always use it for those who don't know.

THE HUMAN PROTOCOL (17:04.534)
Yeah.

James Garner (17:19.404)
what I'm talking about. Reid Hoffman, who was the founder of LinkedIn in his latest book, I think it was called Super Agency, really good book. I'd thoroughly recommend it. He talked about these four personas that exist within, within a attitude towards AI. So you've got the Doomers, they're the people who think that AI is an existential threat to humanity. It's going to wipe out the human race, et cetera, et cetera. Right. Then you've got the, the gloomers. Those are the people who think AI is coming from my job and

THE HUMAN PROTOCOL (17:34.465)
Mm.

James Garner (17:46.349)
You know, I'm going to be, I'm going to be without a kind of purpose in life. Then you've got the bloomers. They're the kind of balanced optimism. They are the people that, you know, they want to embrace AI. They see it as an opportunity, but they're a bit worried about how to deploy it properly. And then you've got the zoomers. Those are the people who are like full steam ahead, incredibly annoying, always going on about AI, probably like us. You know, always going on about AI. So the reason I raise those personas is because if you take a boardroom.

Not any, it doesn't matter where, all those personas are going to exist. again, this comes back to decisions. How are you going to address those decisions? That can either turn into a source of conflicts where you can't really navigate an AI strategy that is workable because there's so many conflicting opinions. Or you can take the approach that everyone's opinion is valid because

We're very, we're even though it feels like a lot has changed, we're still very early on in this journey. And the truth is nobody really knows how it's all going to pan out. There's lots of opinions and actually listen to, you're a doomer, you need to listen to a zoomer. And if you are zooming, need to listen to a bloomer. Listen to other people's opinions because they always come from somewhere. So if you're a, if you've got someone on the board who's a, let's say a gloomer.

THE HUMAN PROTOCOL (18:58.327)
Hmm. To balance things out always, right? Yeah.

James Garner (19:11.494)
And that fear, because the gloom is all around fear, right? They're fear of jobs too. That comes from somewhere. So you need to address it. You need to find out, what is the issue that they're really worried about? And then make sure when you're dealing with an AI strategy, and I hate the word AI strategy, by the way, it should be a business strategy that is powered by AI rather than just an AI strategy, right? First of all, but when you're dealing with that strategy that has AI in it,

You're addressing those concerns. Otherwise you're going to end up with a situation where people feel unheard and you won't have a unified approach. But if you take the approach where everyone's opinion is valid and you can, I do believe that you can make everyone feel relaxed about this because there's a lot of panic going on at the moment because things are moving so fast. And sometimes what you've got to do is you've just got to stop. You've got to take stock. Don't rush into bad decisions because

THE HUMAN PROTOCOL (20:00.191)
Mm-hmm.

James Garner (20:09.132)
just because it's going crazy out there, sometimes the best thing you can do is just be calm and see what's happened, take stock of what's happening. That doesn't mean you shouldn't move fast and react to what's happening, but I'm also seeing a lot of panic decisions going on, both within the tech world, then the business world, and in the personal world. The worst thing you can do is react too quickly to any of this. So I think in the business, going back to your question about the business world, I think it's...

It's just a microcosm of what's happening in wider society, actually. This is happening across the board in education, in business, in even religion. We had the Pope talk on the weekend, right? So every part of life is being reflected in this conversation.

THE HUMAN PROTOCOL (20:47.179)
Yep. Tristan Harold, yeah. Yeah.

THE HUMAN PROTOCOL (20:56.667)
absolutely. And this is where we this is where the future is heading. And you know, the psychological part of it, it's so important. And thanks for for making a reference to that book, because this is really we make decisions based on our beliefs and who we are as people. This is still human beings implementing a technology. So based on where you're coming from and and what you believe is the technology and your point of view of it, that's gonna probably determine the strategy, good or bad.

James Garner (21:11.775)
Yeah.

THE HUMAN PROTOCOL (21:25.619)
on how you're gonna implement it. And and what we want people to think about is we don't know all the answers. And it's always a good idea. It's probably the best idea to, you know, to talk to people that may differ from your point of view. Because that's how we learn and there's always good and bad about talking to people de from with different points of views. Right? But but this is where we're heading and and what I see especially also in the business leaders

James Garner (21:47.852)
7.

THE HUMAN PROTOCOL (21:53.553)
It's really the gap is really on understanding what this is. and it can it can be sometimes technical limitations, right? And and organizational limitations that that p that these companies have. But at the end of the day, these decisions are being made. and we hope that the business leader making those decisions, they have technical people to really explain to them what it takes to implement AI.

James Garner (22:00.588)
Yeah.

James Garner (22:23.946)
Well, you're raising an important point there because first of all, you talk about it being a technological kind of implementation, which it is, right? There's no getting away from that. We're dealing with a very, very powerful technology. But I think where I see it failing and the MIT study that came out the back in the last year kind of alluded to this was that it fails where you try and treat this as a technological rollout rather than as a change management piece. Because yes,

Obviously, it's a powerful technology, but what you're really asking people to do is change the way that they work. Particularly when we're getting into agents, which I know we're going to talk about later, but that is the fundamental. And this is where people's resistance comes up because you're asking people sometimes who might have been in a job or career for 20, 30 years to suddenly, very quickly change the way that they've done things. And we all know we hate change, humans hate change.

THE HUMAN PROTOCOL (22:58.689)
Yep.

THE HUMAN PROTOCOL (23:16.159)
Mm-hmm.

James Garner (23:22.068)
Not surprised. I hate change as well. It's really hard to address that. So that's where those personas come in. And usually drives one of those kind of outlooks on AI. But I think it's really important to recognize, especially if you're leading a business, that you get various opinions, because, you know, you'll get one opinion from the IT team who will probably be very focused on security and safety. Rightly so, rightly so.

THE HUMAN PROTOCOL (23:51.753)
Right Lit So. I agree.

James Garner (23:52.385)
But, but that is not the book because they'll probably say slow down, stop, you know, do everything. But then you also got to look at the people, the people on the front line, the people delivering projects, what, what's their, their requirements, you know, then you're going to have, various people throughout the company, maybe in recruitment or maybe people with the ERP, you know, how does it impact them? And you've got to take all those opinions on board and not just take the IT teams position on board because.

THE HUMAN PROTOCOL (24:06.803)
Mm.

THE HUMAN PROTOCOL (24:16.097)
Mm-hmm.

James Garner (24:22.154)
whilst they do a great job, the IT teams, they will be coming at it from their paradigm, which is, I am going to make this as safe as humanly possible, even at the detriment of deployment, like deploying it properly. So you've got to take a whole suite of different opinions when it comes to this. there's so much good stuff out there, right? Obviously podcasts like yours, but what I always say to leaders in this space is that

THE HUMAN PROTOCOL (24:28.023)
Mm-hmm.

James Garner (24:50.41)
There's no excuse. It's not like the old days. There's so much good information out there. All the big frontier houses, Google, open AI, they all release fantastic information for free. Then there's a multitude of different podcasts out there, multitude of different YouTube channels. So you can find all the information that you need really easily. It's all out there.

THE HUMAN PROTOCOL (25:10.891)
Yeah, it's it's it's definitely a lot of information out there, a lot of noise, but a lot of good stuff. If you take the time to try to dive in and and try to find good sources of information and for free, you're probably gonna find a lot. that that's definitely not something that we are lacking on, thankfully, right now. Yeah.

James Garner (25:26.796)
Exactly.

James Garner (25:30.7)
I know. It's almost the other way around, isn't it? It's like information overload. Where do I start? That's the problem.

THE HUMAN PROTOCOL (25:35.895)
Exact. Trust me, we we go through it every day, James. It's it's the nature of things, right? But I wanted to pick your mind in in in kind of the next topic that I want to talk about, which is this change or this transition that we're going from being very shotbot driven to physical AI, right? And you know, it's in the past it used to be

James Garner (26:00.013)
Okay.

THE HUMAN PROTOCOL (26:05.405)
AI and automation, it was putting a chatbot on a website and you you you you were done, right? But what does physical AI really mean in practice?

James Garner (26:16.556)
So you're talking about humanoid robots and robotics, or you're talking about agents?

THE HUMAN PROTOCOL (26:20.383)
Right. That talking more about the physical AI on how we are moving from shot interfaces to machines and agents that actually act.

James Garner (26:24.374)
Yeah.

James Garner (26:30.154)
Yeah. Okay. Yeah. So I would separate the, in your definition of physical AI into the two. There's the, there's the, the agentic AI and then the physical AI as in robotics. So let's take each one in turn. So where we're moving from the, so from when chat GPT came out to, you know, around the beginning of this year, it was all very much quite passive experience in that, you know, you put a prompt in.

And then you got an answer back and you could have a conversation, but it was very much, okay, let's take a really simple example. I want you to help draft an email for me and it would draft the email and then you could take that email and you can cut and paste it into your email browser and then you can send that email. Right. So that's the way people have been interacting with AI up until now. Now what we're moving into is an agentic AI. So what does agentic mean? Well, it simply means that you're giving the chat bot

whether that's chat GPT or Claude or copilot, you're giving it agency. That's where agentic comes from. You give an agency. What does that mean? You're allowing it to do things. So let's take that really, really simple example, email. So an agentic version of that would be not only does it write the email for you, it then goes and sends it on your behalf. That's the agent. That's the agent. Immediately you can see two things there, hopefully. Number one, that's incredibly cool and amazingly powerful. On the other hand,

THE HUMAN PROTOCOL (27:47.047)
Mm-hmm. Yeah.

James Garner (27:57.388)
That's bloody dangerous. Right. So, yeah, it's because you're letting it do things.

THE HUMAN PROTOCOL (28:00.309)
Damn, yeah. Especially if you don't insert yourself in the loop in the middle of it somewhere, right?

James Garner (28:08.157)
Yeah. So, so this is the thing, right? So that's a really, really simple example. Now you expand that out to multi-layered workflows. Take anything that you're doing where you've got a number of different steps. You start actually putting an agentic workflow together where you've got various agents working together to complete a full workflow. Now, now that gets even more cool and even more incredible. And at the same time, even more dangerous and even more

THE HUMAN PROTOCOL (28:21.793)
Mm.

James Garner (28:37.547)
more of a security threat because you know, any point in that chain, things could go wrong. Now you raise the point human in the loop, which is what everyone talks about as the way to ensure that we're still in control and we're signing things off as, as things get done. Now I actually hate the word human in the loop. I tell you why, because that implies that we are somehow subservient to the AI.

I prefer the word bookending, which is we're taking responsibility for what we're asking it to do and the final output. But semantics really is the same issue. I just don't like that terminology because it does imply that the AI is kind of on top of us. you know what I mean? Someone raised that to me, a very clever guy called Sir Robert Buckland who

THE HUMAN PROTOCOL (29:26.273)
Yeah. Putting it that way sounds like now you made me think.

James Garner (29:34.316)
and my podcast raised it to me and I've never been able to forget since he said that. And I thought you're absolutely right. Absolutely right. Yeah. Because it does, if it gives that, but anyway, the principle is right. You know, there's a human somewhere in there controlling the steps. So we've already seen this start to happen. If anyone's got copilot, they can create a very simple agent themselves. And then we're also moving to full agent workflows. And then...

THE HUMAN PROTOCOL (29:40.811)
Now you're gonna make me think every day about this.

James Garner (30:03.397)
even within things like Claw, if anyone uses Clawed, there's things like Clawed's cowork or Manus is another one, or even within ChatGPT now with Codex and things like that. These are like generalized agents. can do an incredible amount with them. You could ask it to go and build a website for you and it will go and build a website for you. And it will come back, back and forth. It will ask you questions and it will, you know.

It will really sit. It'll be like an assistant. will be like the clever assistant you've ever imagined constantly doing stuff for you, constantly badgering you saying, is this okay? Is this okay? Is this okay? Unless you go to ultra dangerous mode, which does stuff by, by itself. wouldn't recommend that, but at the moment it's very much saying, are you happy to go down, down the street? You're happy to go to the next step, which then that's the safeguarding. So that that's where we're going with agents and that is going to become the norm very, very quickly. And.

there's a massive gap emerging between the, like I said at the beginning, the availability of this technology, which is freely available now, pretty much, and the people who even know about it. Most people don't even know it exists, right? So that's the gap. So that's the agent side of AI. And then you talked about the physical AI. So this is the next step, which is a little bit further along, but it's also moving very, very fast. When I...

THE HUMAN PROTOCOL (31:17.047)
Mm.

James Garner (31:29.035)
When I talk about physical AI, I'm really talking about AI in robotics. So we've seen a huge amount of press around humanoid robots and all they are, are physical AI. They're robots that perform functions and it's got AI built into them. So if you take the Tesla one, which is called Optimus, it's built on top of the GROK AI, is Elon Musk's AI. Yeah.

THE HUMAN PROTOCOL (31:56.971)
They're mm-hmm.

James Garner (31:58.846)
So it's basically a physical embodiment of that AI and they are becoming incredibly powerful. Like there's still a few years away. I actually met my first one the other day in Amsterdam, came up to me and greeted itself and shook my hand. was incredible. Yeah. But when you take industries like the industry that I work in construction or manufacturing, I think the BMW factories now got humanoid robots working in it. These robots,

THE HUMAN PROTOCOL (32:14.375)
boy.

James Garner (32:27.775)
they can go pretty much 24 seven. Yes, they have to charge, but they take themselves off to the charger, plug themselves in and then carry on. that's going to be, when people talk about AI taking white collar work, it is going to have a major impact on white collar work, but blue collar works not too far behind either. So I think both of those things in terms of agents and physical AI are happening very, very fast.

I'd say agents will be commonplace. Most people will be using them by the end of this year. And I'd say humanoid robots will be like, I think I said at the beginning of this year on my podcast, when I had like a year review, I said to my guest at that time, my prediction this year is that everyone will at some point during this year, meet their first humanoid robot in like a normal scenario. like just at the shop or in the shopping center or,

you know, doing going around their daily business. And it only took until April for me for that to happen. So, you know, that that's speeding up. So it's incredibly exciting time, but I can also understand why it's quite overwhelming to.

THE HUMAN PROTOCOL (33:38.867)
It it it is. Imagine if the use of the technology, let's say just understanding what AI means, it's already overwhelming for people. The vision of having AI agents doing things for us, and not only one AI agent, but an orchestration of AI agents working together for a lot of people maybe too much of a science fiction, but this is real life. This is happening today. And what are g yeah. No.

James Garner (33:47.936)
Mm.

James Garner (34:05.309)
It's real life and we're not ready for it. That's the other thing.

THE HUMAN PROTOCOL (34:09.639)
And and and I think by the end of this year, it's you know the the biggest part of the business world we have one way or the other, from sixty to eighty percent of their businesses already with AI agents performing them. It's what I'm looking is and and I'm seeing an acceleration of businesses really implementing and developing AI agents to be able to run their business.

at least the their main processes and more businesses are looking at how I can implement AI agent to start not only automating but also being able to make decisions. If they're doing it right or wrong, you know, making sure that there's safeguards in place or not, that's a different conversation. But what business leaders have in mind is really it's AI, AI agents, how we can implement it. And this is what

This is heading down.

James Garner (35:09.513)
Yeah, it's here and we're seeing the results that are ready. The number of graduate level entry jobs are going down because that's going to be the first thing that these agents do, kind of manual, menial work that we get the grads in to do. And that's already happening. So all of the big four consultancy firms have reduced their...

grad intake, a lot of the big law firms reduce their grad intake. That's what it's coming for first, but it won't be the end at all. And I said before, we're not prepared for it. So what do I mean by that? Because whilst we've got the technology to be able to use these agents, most people are lousy managers, right? And somebody still needs to orchestrate these agents. if, believe it or not, if you're lousy at managing humans, you're probably going to be lousy at managing agents as well.

Right. And so we've never really got over that. And that's why he says we're missing this major piece, which is if you want to really thrive in this new world, actually what you need to learn is management skills. And yeah, okay. You're not going to be managing humans, but you're going to be learning to manage a bunch of agents. And believe it, they are even more difficult to deal with than humans because humans, you know, can interpret what you mean. You may say something and they go, I know what he means.

The agents will be very, very specific about what you ask it to do. So when you get a bad result from agents, and I've definitely been there, it will always come down to the fact that you haven't briefed it properly. You haven't prompted it properly. You haven't given it enough context. And we are lousy at it. And it's a skill that all humans are going to have to learn if you want to get on top in this new world.

THE HUMAN PROTOCOL (36:40.609)
Yep. Yeah.

THE HUMAN PROTOCOL (36:50.667)
Yeah. All that that you said is extremely important and something that I've seen and I think we're gonna hear and see a lot more in the future, James. And and actually recently I I was reading something I I it just popped up on my my phone that Microsoft it's actually just announcing that it has plans to restrict the use of AI by their engineers. Why? Because of rising costs.

James Garner (37:15.913)
Yeah, yeah, token costs.

THE HUMAN PROTOCOL (37:18.947)
And and and the cost part of this technology, it's something that at least I haven't been able to find someone with a great explanation on how are we calculating the cost which is running an AI solution. You have people saying, you know, it's the cost is based on the time that you're using, how many workflows are being running in the past, but these LLMs.

James Garner (37:35.147)
Yeah.

THE HUMAN PROTOCOL (37:47.677)
And these companies are charging differently. And if we are gonna be moving from a budget that is managing people to a budget that is managing people plus AI cost. And the understanding for people's leaders and managers and leaders to understand the cost associated with AI when it's not clear, that adds a different set of challenges that way. And I think we're gonna see a lot more.

James Garner (38:13.855)
Yeah. Yeah.

THE HUMAN PROTOCOL (38:16.171)
When costs associated with AI are gonna start coming back up.

James Garner (38:21.097)
Yeah. This is a difficult one for me to have a view on because I think we are being led into a false sense of security at the moment because it's a bit like when Netflix and Spotify came out, they gave it away for pretty much free to start with, got you hooked and then slowly start raising the prices. Right. now we're all stuffed and we're stuck with it. Right. So it's really easy to say, I wish I could turn it back, but you can't once it's gone, it's gone. know, blockbuster video is gone, isn't it? So it's, that's the way of the world.

THE HUMAN PROTOCOL (38:36.829)
Mm-hmm. Yeah.

THE HUMAN PROTOCOL (38:45.355)
Yeah.

James Garner (38:51.051)
And I do at the moment, we know all these front houses are making an enormous loss, right? They're all running at a loss and they're all competing against each other to try and get the market dominance. At some point, things will settle down and the costs will start going up. And, you know, that's when we'll be faced with the situation of the reality. What does it really cost? Not is what is the cost that they're currently passing on to us.

but what is the real cost at the moment? They're measuring cost in tokens, right? If it's sort of an enterprise solution, which is basically, you know, number of words and we're being really inefficient with the way that we're using AI at the moment actually, because no one's really worrying too much about it. Because if you've got like, let's say you've got loads of oil, you've got to

you're not worried about the use of it. You'll just keep using it. If you get to the point where there's a scarcity of oil, you start being really careful about how you use that oil, right? And it's the same thing with AI in that at the moment we feel that it's freely available. There's nothing that we need to worry about. And so we're just using it without worrying. But at some point that tap will switch off and

THE HUMAN PROTOCOL (39:58.305)
Yeah.

THE HUMAN PROTOCOL (40:11.712)
Yeah.

James Garner (40:17.06)
everyone will be going, God, we've been using it really, really inefficiently.

THE HUMAN PROTOCOL (40:19.263)
It's sure.

It should. And and and you know, to your point, James, it's just to bring it down to very human pay, you know, humans using, for example, one of these LLM shot GPT or Cloud, whatever. This is why prompt engineering it should become more and more important. The more you prompt, the more you use it, right? In the back end, the more computing being used. So instead of you taking the time to understand what you want to achieve and try to come with the best prompt that you can.

in one or two maybe people are just prompting prompt they don't care or it's b X it or it's free or it's twenty dollars a month at least here in the US unlimited prompting you know what I mean I think we're probably gonna see some restrictions coming in the future because we cannot keep the whole world prompting prompting and prompting where are we gonna end up

James Garner (40:55.243)
Yeah, because there's no jeopardy.

James Garner (41:12.363)
Unless, unless of course what we didn't talk about is the local model, you know, the local model. So we're talking about the big frontier models. Of course you've got all these Chinese open weight models that are in effect free or very, very cheap. And then you've got the likes of Google releasing things like Gemma four, which was their, you know, their local model. I think that's going to play a big part when people realize that, you know, calling these APIs are going to become very, expensive. They'll look at ways of

becoming more efficient with their AI usage. And you can imagine a world where, you know, maybe 80 % of AI queries will go through their local model. And maybe they'll only call the more expensive models for very, complex tasks. I think that's gonna be a part of the future. This is unsustainable. It is unsustainable alone.

THE HUMAN PROTOCOL (41:52.097)
Mm.

THE HUMAN PROTOCOL (41:58.517)
Mm.

THE HUMAN PROTOCOL (42:02.119)
It is. It is. There, you know, energy consumption, water consumption, there is a lot going on. That's why there is a race for everybody now. Countries, right? Governments, they don't want to be left behind. Companies, they don't want to be left behind. And everybody needs more computing power. And that's where these big humongous data centers are coming into play. The future of it depends a lot on the computing power that you have to back you up your operations.

James Garner (42:30.345)
Yeah, and data centers in space.

THE HUMAN PROTOCOL (42:34.109)
Or underwater or a little bit above water, floating data centers. I I was reading the other day floating data centers. You you know what's funny, James? I was I was reading the other day about these floating ideas, floating data centers. And the main concern about the reporter was, but the the main concern is because since you are floating, then you will need to use a satellite connection to send information back and forth.

James Garner (42:38.218)
Well, no.

No. What else am I gonna go? Mad, isn't it?

THE HUMAN PROTOCOL (43:04.009)
And that will take longer than your instant response. Okay.

James Garner (43:07.583)
I know we've got very greasy haven't We've got so used to this technology in such a small space of time that we moan when it takes anything more than a few seconds to get on tap intelligence.

THE HUMAN PROTOCOL (43:22.749)
Are you kidding me that that's the speed of the response? I had to laugh.

James Garner (43:27.115)
It's mad. It doesn't surprise me though, because we're weird people, human beings, you know, we get very used to... Have you heard of this thing called the Overton window, which is an amazing theory, which is basically that what sounds like crazy today becomes tomorrow's reality very, quickly. And then we get used to that new reality really, really quickly. So, know, wind back 20 years, right? And just the thought of the things that we live with now...

THE HUMAN PROTOCOL (43:33.396)
Yeah.

THE HUMAN PROTOCOL (43:38.199)
No, what what it is.

James Garner (43:56.607)
you know, mobile phones, are basically the most powerful computer in our hands. Yeah. Like ridiculously fast internet, artificial intelligence, AR, VR, you know, talking about quantum computing breakthroughs, humanoid robots. People have thought you were mental talking about that. And we've just got used to it really, really quickly. And then we, and then we'll, as soon as we get used to that, we'll start like laughing about the next big breakthrough.

THE HUMAN PROTOCOL (44:00.31)
Mm.

James Garner (44:25.522)
as if that's ridiculous. Well, there's no reason at all why that should be ridiculous. I was talking to somebody the other day about, you know, nanobot technology and the fact that at some point in the not too distant future, these nanobots, the chips in them are going to be able to be small enough to put into like almost something the size of a cell and then we're to be able to put it into our bodies. And this is exactly Ray Kurzweil's hypothesis. If you've read any of his books, the Singularity books. But, you know, these people are just...

good friends, but they're just absolutely laughing at me saying it would be ridiculous. I'm thinking, yeah, go back 10 years, you would have been laughing about what we're doing now, right? So we're strange creatures.

THE HUMAN PROTOCOL (45:03.265)
I I think we've gotten two e two way easy. We've gotten two way easy, especially with AI, to the point that that that we worry about that. But you know, at the end of the day, it's it's humans, right? This is this is who we are. This is the species we belong. And that's kind of a good segue for for my next egg for a next topic and question that I want to ask you, which is the the human response to this.

James Garner (45:08.99)
Yeah, absolutely.

James Garner (45:15.934)
Yeah, we are with.

THE HUMAN PROTOCOL (45:32.449)
Curious to see what have you seen, what type of emotional response have you seen when when people start using AI and their capabilities grow? Any any anything that you can share with us that you have seen in people?

James Garner (45:47.999)
Yeah, I mean, I've seen all kinds of responses. I've had somebody in tears. So I helped somebody with an AI, a very simple AI agent. And to, I won't go into the specifics, but it was to solve something for that particular person. And when it got, when it was working and it ended up being deployed, that person burst into tears because...

behind that was a major, major issue that they were unable to get past. They didn't know how they were going to get through this problem. And this AI agent basically solved it for them. And it was such a relief that they expressed themselves with tears. So I've had that. And then I've had fear. Definitely people visibly look shocked when you show them what it can do. you sit, kind of, don't say anything, but you can see this kind of emotion of

THE HUMAN PROTOCOL (46:26.017)
Wow.

James Garner (46:44.714)
Oh shit, my jobs. That's my job. We've just automated. So I've seen that. I've seen the ridicule, ridicule from like I just expressed where you talk to people about where this is going and they literally just laugh and think you've gone off your rocker. I've had that quite a lot. And then I've also had a little bit of anger where, you know, people genuinely react in a way that can only be described as

THE HUMAN PROTOCOL (46:46.807)
Yeah.

THE HUMAN PROTOCOL (47:01.351)
Mm.

James Garner (47:14.27)
with anger because they feel that this technology is not being governed properly. You know, they're not angry with me, but they're almost angry with me as a proxy because I'm a proponent of it as if to say, this is all wrong. We're not doing it properly. And I do have sympathy for that argument as well. I do understand where they're coming from because I do feel that we're doing things that are maybe sometimes a bit too fast without thinking about the consequences.

THE HUMAN PROTOCOL (47:20.139)
Yeah. Right.

THE HUMAN PROTOCOL (47:26.487)
Hmm. Mm-hmm.

THE HUMAN PROTOCOL (47:43.275)
Yeah. And and and I think, you know, at the end of the day, human resists change, but and and and about technology in particular, human receives what it takes from them, right? or they are gonna fear losing their jobs and with that losing who they are, right? Because we know that if jobs changes and roles changes and you're not gonna be doing the same thing you have been doing for the past five, ten years, of course, the nature of things and your job.

James Garner (47:43.828)
Sure.

James Garner (47:54.078)
Yeah, always.

THE HUMAN PROTOCOL (48:12.225)
may be a totally a different one. There's challenges inevitable, right?

James Garner (48:16.91)
inevitable. Inevitable and that always has happened, right? So when people say that, I just push back and say, look at the jobs that used to exist that don't exist anymore. Right? Did you know, I don't know if you had this in the States, but in the UK, there used to be a job back in the late 1800s, what's called a knocker upper. Have you heard of this? A knocker upper. Yeah, this is a true job. Their job was they used to go around the houses with a stick in the morning and knock on the

THE HUMAN PROTOCOL (48:24.822)
Yeah.

THE HUMAN PROTOCOL (48:34.047)
No what is that?

THE HUMAN PROTOCOL (48:43.646)
Yeah.

James Garner (48:44.254)
windows to wake people up because this is before alarm clocks. Right. And there was another one called that lamp lighter. They used to go around the Victorian streets and light the lamps. You know, obviously those jobs are gone now. And I'm sure those people at that time were worried about this new fangled technology called electricity that came along that took away their job. So it will always change. Things always change.

THE HUMAN PROTOCOL (48:48.518)
THE HUMAN PROTOCOL (48:59.851)
Yeah.

THE HUMAN PROTOCOL (49:06.675)
It will always. It's what the email and the internet did with, you know, the postal offices, you know, the changes are coming and this is not the only time that will come. After AI will probably come something else. And you know, it's it's a matter of time, basically. It's

James Garner (49:25.098)
Well, it will just speed up. It will just speed up, I think, because we're on that hockey, that exponential hockey stick that Ray Kurzweil talks about that, you know, we're going to get this record like self-improvement happening soon where the AI starts making the AI get better. So he's going to change that. It is going to speed up 100%.

THE HUMAN PROTOCOL (49:42.771)
And and this is the same that I tell my friends, my family, you know, people that I work with, we have a chance. Or we look at the positive or we look at the negative. It's all how we look at things. And based on how you look at it, there are actions that you need to take. you're not gonna be able to be in the middle because this is a transformational technology. This is transforming our life in many different ways, right? Because this is being adopted.

in all industries, you know, personal and your professional lives as well. So it's just a matter of how you are gonna approach to the changes, how you're gonna adapt to it and and your attitude to it. Yeah, absolutely. Awesome James. what do you think AI companies should focus on? What are the things at the top of your mind?

James Garner (50:23.186)
Attitude, attitude is the... yeah.

THE HUMAN PROTOCOL (50:40.107)
That if you have in front of you OpenAI or Anthropic or Gemini leaders, you're gonna tell I would like you to focus this technology and pay attention to this. What would that be?

James Garner (50:54.196)
What would I tell those AI companies to focus on? That's a really good question actually, not one I really thought about too much. mean, I would say to them that they need to focus on that they all seem to be in a war against each other. And it's getting to the point that they're all trying to outdo each other with new releases all the time. And I would just say to them, you focus, focus on rather than quantity.

THE HUMAN PROTOCOL (50:57.001)
Yeah. Mm-hmm.

James Garner (51:21.758)
focus on quality, like just get some of your features done really, really well, then just chuck, particularly Anthropic at the moment seems to be chucking new stuff out every day and people can't keep up. I would say focus on real use cases and give people some space. I know you wanna get stuff out there. I know you're in an arms race, but remember that people are humans and they can't take all this stuff in. So I would just tell, say slow down, focus, work on the use cases.

THE HUMAN PROTOCOL (51:52.597)
All right. And what would you tell an eighteen years old James seeing him today, especially going into this new feature with AI?

James Garner (52:02.154)
Well, I've got a 19 year old mini me, so who's my son, who I do talk to. Yeah. So I do say to him and, um, you know, embrace it, embrace, embrace it. Don't, don't resist it. Work at like we talked about before in terms of education work. What is your attitude? There's a great book called the gap of me. I think it's called the gap in the game, which is around your perspective. Whether you look at

THE HUMAN PROTOCOL (52:05.335)
Or night or night.

James Garner (52:28.84)
you know, what you've achieved versus what you've got yet to achieve really determines how you function in society. And it's the same thing with this. It's like, choose your attitude towards this. Yes, there's dangers you've got to be aware of, but look at the positives. Look at the things that it can really, how it can improve your life and use it to improve your life.

THE HUMAN PROTOCOL (52:50.323)
Awesome. Amen to that. And and James, my friend, thanks for being with us today. This was very this was very insightful conversation.

James Garner (52:58.218)
Oh well no, thank you so much for having gone. It's been fun. It's been a lot of fun talking to you.

THE HUMAN PROTOCOL (53:02.983)
Awesome. Appreciate you. And if this was a powerful reminder that the future of AI is not just about what machines can do, but what humans are willing to face. Share this with your friends, with your family, with your loved ones. And I talk to you guys next week. Thank you all.