Introduction

Today, this is the first time I'm going to try and do this talk, because I have talked about the future of work quite a lot in the past. But I tried to rejig this a little bit.

And you've seen two pretty cool demos today. And I spend all of my day, every day, trying to figure out how do people use this technology? How do non-technical people use this technology?

in order to be more productive. And so I wanted to give you a little bit of a flavor of how at least I think about the future of work and how I think about the future of skills.

The title of the talk is Degrees Won't Save You Anymore.

And I will definitely forget a few things because I don't have my notes, but I'll try and go through that properly.

Event Overview

First thing is you probably will have all received an invite, or at least maybe it went to spam, maybe it didn't. But anyone that actually signed up to the MindStone platform, you will have received an email that asked you to opt in to potential introductions. So it was an email that said, hey, you're coming to this event. There are other people. Do you want to opt in to be matched so you meet the right people at the event? It would opt you in. There's an AI that asks you who you are, what you're looking for, and then it tries to make the right matches and introduces you in that way. And then it actually does the introduction and it puts both people in copy.

The idea being we're all here, there is 250, maybe 280 of us in this room. It's always really hard to try and figure out who are the few people that you should really meet.

Now, I want to take you on a little bit of a journey because that app was really interesting. How many of you have heard of Boardy before? There are a few in the room. This was an app that recently launched. It was an interesting one because basically it contacts you, you can have a conversation, it asks you what you want, who you are, what you can offer, who you're looking for, and it actually calls you.

The Changing Value of Skills

Now, the value of skills, going back to what this means from a skills perspective, the value of the average skill, the lifetime of value of an average skill has been going down for decades. It used to be that you learned something, and you do that in school, and you are able to apply that for quite a long time.

Recently, the average school has already dropped down to about five years, which is barely more than it takes to go to university to begin with. Very soon it will drop under that, and I would argue that maybe we're actually there at this point. There aren't any studies that put it at under three years yet, but there are studies that put it between three and five years at this point.

Now, I have a computer science degree, When I did my computer science degree, I was still taught Flash and ActionScript. For those of you that have had the bad experience of trying to go through that, it was an interesting experience, but more importantly, not only was the time it took to go through university the three years, and that's the thing you're battling with, the second bit was that I was being taught things that were irrelevant even when I was learning them.

So you're not even just battling that three-year time period. It's the fact that the actual thing you're being taught is already irrelevant by the time you're learning it. So you have to add that time lag to it.

Adaptability and AI

And what we're seeing now with MindStone, I can give you I can give you some examples around adaptability.

The possibility today, the difference between someone that knows how to actually use AI for their job and those that don't, there are two interesting stats. Right now, was it 60 odd percent or so of hiring managers would rather hire a person with less experience but with the possibility to use AI in their job than someone who's more experienced that does not. The second bit, and that is something that we see directly every day, is that there is a genuine 10 to 60% productivity increase with the basics of using this technology.

10 to 60 is a big range, I know, because it depends on the type of job that you are doing and how well can you actually use these different tools and how well they insert themselves. But the... 1This entire idea of being able to adapt, your adaptability is going to be much more important than whatever skill you will have acquired in the past.

It's not what have you done in the past, it's how do you actually react to what is right in front of you right now. Everything I used to build that app on Friday didn't exist two years ago. And so if it didn't exist two years, actually it didn't exist a year ago.

Fuck, this is crazy. That should show you or tell you just how much that adaptability is actually the key as you're trying to get things done.

Now,

I try to figure out what is a better paradigm of how you're thinking about what our relationship with AI is here when you're thinking about a skill enhancement perspective. So we've had kind of the teacher-learner model, it's one-to-many, you do the work first, you learn something, you then apply it to the later stage.

I think we're getting to a point where AI is like a personal trainer but for our mental health, for our mental capability, sorry, capacity, for our mental flexibility, for whatever task we have in front of us. Now, a personal trainer works with you while you are working.

Your personal trainer doesn't send you a message, doesn't try and teach you something, and then you go and apply it. No, it's like you actually, while you're performing the thing, you are being told, okay, do this differently, do this differently. Now, you can go and do the work yourself in between sessions for sure,

but it's actual direct feedback, personalised 100% to you, it's just in time, and in the case of a job here, and this is where you go from a physical trainer to an AI personal trainer, it's directly on the job. You are being taught whilst you are doing whatever the task is that is right in front of you.

AI in Gaming and Work Contexts

Has anyone seen this latest announcement from PUBG about their AI allies? Nobody, okay, so this is what I think we are going to end up at from an AI perspective. Just think of it, this is in a gaming environment, but this is going to be how the AI will work from a work perspective at the same time.

At Krafton, we are revolutionizing the battle royale genre with generative AI. Introducing PUBG Ally, the world's first co-playable character designed to play like a human teammate.

Hey, ally. I'm looking for a level 3 vest and 5.56 ammo. Keep your eyes peeled. Found a level 3 vest and ammo nearby. Ping it for you. Enemy spotted. I'm covering.

Hmm. Is there a vehicle nearby? Found one. I will come to you. I'm ready to get into the action now. Can you flank them? On it. Moving left.

Wait. I'm knocked out. Crawling to you now. Just hang on. I couldn't have done it without you. I know. I carry for a living.

Now, this is a gaming context, I get it. And you, actually, they're here. But I would like you to think through what was happening here, right?

There was, at some point, the user, the actual player was saying, hey, I need a vehicle. No idea where the hell that is. The AI goes out, gets the vehicle, brings the vehicle to you, kind of does a thing.

Think about a work context. When you have specific things that you need to get done, the AI ends up being that extension. So, okay, well, I know that this thing is going to come up.

I don't know. I'm doing a presentation. You know what? I really need some data on X. Get me some data on X. Boom. It goes out, knows whatever presentation you're working on, goes and crawls the data, gets you the thing back.

I really need to schedule a call with this person to go over this particular portion of whatever project I'm working on. It goes and does that. Or I need a special app that collects the data for this group of users, goes off, builds the app, gives you the thing. Is this what you wanted? No, actually I wanted a little bit different. Think about this AI in that way. from a work perspective.

Adoption of AI in the Workplace

Now, if you think that this is going to take longer, you're probably right. I am definitely, I have been proven wrong on the speed of adoption from an enterprise perspective on this. I thought it was going to go faster.

But there was a real aha moment for me in November. I did a panel at Web Summit, and we had about 700 people in the room. And I asked everyone in the room, how many of you are currently using AI in your job?

And 80% of the room had their hand up. And it really shocked me because it was absolutely opposite of what I was... We do AI training for non-technical people in corporates all the time. So many corporates don't allow their employees to use this at work at all.

So I was like, what the hell is happening? And then I asked the follow-up question, which is how many of you are using your personal accounts to go and do so? 60% of people still had their hand up.

75% of people that are currently using AI for their job are doing so on free personal accounts. And if you followed some of Jacob's talk before, the free and personal accounts, yes, they do train on all your data, and yes, that does mean that all of this data is leaking through in different aspects through all of these models.

So I look forward to looking up the latest Deloitte strategy. This is gonna be stuck up in some model somewhere that someone imported at some point. What is interesting though is that this is happening right now.

People are using it right now and it is not as far as you might think. It is also not stopping. And I fundamentally think this is, there is so much narrative for those of you that are kind of following, it's like, are we hitting a wall?

Is this just getting, are we not getting any more returns coming through? And it is just not true. Yes, there are some things, the actual general models, like the more data you throw at them, we are getting less and less

extra capability out of these models. But at the same time, there are more and more techniques to go and make these models go further, to make them smaller, to make them more accessible to everyone. So now you're looking at both the model itself, but also what's now called inference.

So you've got basically the models are thinking more before they actually give you an answer, and just the fact that they are thinking more while they're doing, you get a better result for it. You can see there's this big benchmark specifically built. I think it's trying to map or it's graduate level questions and humans tend to do about 65% or so on these different questions that come through.

This is the different models as they have come out over time all the way to January 2025. And I don't know how anyone can look at this, look at how this is mapping to different models, and then say, yeah, we're no longer getting better models out of them. And anyone that has been using them actively

You are seeing these increases every single day. It's a weird thing that last year was definitely the fastest year of evolution from a technology perspective that I have lived through in terms of capability, and this year seems to be topping that, or at least it's on path to be topping that. No one is quite really understanding what is happening there.

Degrees vs. AI Adaptability

In that world, it really doesn't matter if you have a degree. That degree was three years ago. If you don't know how to use AI, the degree is just a footnote in whatever your capability is to be able to go and use these models in the future.

I'm being a little bit direct on it. I'm not saying it's zero and there are some degrees that are about reasoning and all of that is very, very true. But to the degree that a degree is about teaching you specific skills, those really don't matter anymore. It's all about the adaptability that comes through.

Amplification, Not Replacement

Now, the good thing is I genuinely believe at this point that we're looking at, for those that actually lean in and that look at using AI in their jobs, we're not looking at replacement. We're looking at proper amplification.

Anything that I do today is double-checked, the AI plays devil's advocate or whatever I'm doing. There is no longer anything important that I do that doesn't run through AI in one way or another to try and make sure that I make less mistakes.

Now, obviously, this presentation itself was done by AI and in a way that was ten times faster than I would usually do, because normally I would have to go through everything that I think, try and structure that, come up with a presentation. My design skills are nowhere near what you saw today, and I know today wasn't dramatically amazing, but it was ten times better than what you would have had had I not had Gamma that would help me with it.

My flow today was literally throw everything that I wrote in the last six months to ChatGPT, I use 01Pro, come up with interesting narratives based on everything that I've already written, so this is stuff that I actually believe that I have put in, it's my content, but the flow itself, ChatGPT, brainstormed back and forth, threw that into Gamma, this is the presentation that came out. Obviously there's a little bit more than that. I was filtering through and figuring out which ones I preferred and how to iterate through it.

Conclusion

But even just a day like today and part of the job that I have to do doing presentations like this really make a difference. So all of that to say that I think we are looking at a dramatically different future and hopefully this was useful and it got you something.

Any questions?