Ep 37: How Talent Intelligence and Human-Centric AI Are Impacting Workforce Dynamics with Toby Culshaw

[00:00:00] Welcome to Elevate Your AIQ, the podcast focused on the AI-powered yet human-centric future of work.

[00:00:05] Are you and your organization prepared? If not, let's get there together. The show is open to sponsorships from forward-thinking brands who are fellow advocates for responsible AI literacy and AI skills development to help ensure no individuals or organizations are left behind. I also facilitate expert panels, interviews, and offer advisory services to help shape your responsible AI journey. Go to ElevateYourAIQ.com to find out more.

[00:00:39] Welcome back to Elevate Your AIQ. Today I have a chat with talent industry veteran Toby Culshaw, currently leading talent intelligence and strategy for Amazon stores across the globe. I've known Toby for a while now thanks to the Talent Intelligence Collective community he leads. I've always been impressed by his insights and leadership in this space. I am a member of many communities and this is one of my most highly engaged ones for sure. Anyway, I'm really glad to get him in the virtual studio. We dive into some of the big topics shaping the future of work from how AI is redeveloped.

[00:01:09] Defining talent acquisition roles to the complexities of the labor market, changing workforce dynamics. Toby also shares his perspective on balancing technology with the human element in leadership and why community has been such a powerful force in driving talent intelligence maturity.

[00:01:25] Did I mention that Toby literally wrote the book on talent intelligence? I think you'll really enjoy this conversation. Thanks for listening.

[00:01:32] Hello, everyone. Welcome to another episode of Elevate Your AIQ. This is Bob Pulver. This morning, I have the pleasure of speaking with Toby Colshaw. How are you doing, Toby?

[00:01:42] I'm very good. Thank you for having me.

[00:01:43] Absolutely. It's good to see you.

[00:01:45] Lovely to be here.

[00:01:46] We always have so much to talk about in the world of AI and talent. I thought it'd be great to finally just hit the record button and let others share in the interesting conversation and in your wisdom in the space. So thank you again for being here. Toby, just in case people don't know who you are, you want to just give a quick overview of who you are, where you are, and what you're working on?

[00:02:14] Yeah, for sure. So I've been in the talent acquisition, research, sourcing, intel space for about 20-something years. And then over about the last 10 years or so, I've been getting heavier and heavier into pure talent intelligence. And we can get into that as to what that means and the debates around it and everything else.

[00:02:33] But currently, I lead a team at Amazon. So I handle a division of Amazon called Amazon Stores, where I lead their talent intelligence function. So global team, full coverage, we'll do research on anything from logistics managers and warehousing robotics through to cybersecurity and corporate functions.

[00:02:51] So a real breadth of different roles and expertise and skills we're diving into. But I'm a bit obsessed by the whole TI space. So as we know, I do podcasts on it, I do events on it. I'm a bit obsessed by it. And I think the potential within is massive. And I think we're barely getting through the veneer of what's possible at the moment.

[00:03:10] Yeah, I would agree. And I think your obsession is shared by many in the collective, as we call it. And I think for good reason. I mean, just like there's untapped potential in us human beings, there's untapped potential in using talent intelligence to do a better job putting the right people in the right places.

[00:03:30] So thank you for all your effort there. I'm a member of a lot of communities. And this is definitely one of, if not the strongest one in terms of the cohesion, the engagement level. And then of course, the resources that you make available beyond the community. So on behalf of all the collectors, thank you.

[00:03:52] It really is a collective thing, though. It's the community that make the community. The old adjective, you can take a horse to water, but you can't make it drink. You can put some of the structure and the systems, processes, tools in, but it really is the people that make it.

[00:04:07] And I think luckily we're in a community where it's both the practitioners and the vendors are very open and very transparent and very engaging. And I often say this to people that we're all just trying to find our way on this stuff.

[00:04:20] Like there is no, no one's got it cracked. No one's got it nailed. Everyone's trying to find their feet and find their way because it is so broad. It's such a huge topic for anyone to try and address and go after.

[00:04:30] So everyone's doing kind of slightly different elements in different ways. But because of that, everyone's having to learn. Everyone's having to build things on the fly. And because of that, people are very open to talk about the good, the bad, the ugly, the problems they're facing, which is wonderful. It's wonderful. The community really do make it.

[00:04:49] The selflessness is one way to put it. People are just really generous with their time. They're generous in offering clarifications where there might be some terminology debate or where one organization, one industry or at one scale, large enterprise versus smaller business.

[00:05:10] I think it's very honest.

[00:05:13] I think it's very honest and candid feedback about what's working and what's not. Like you're going to run up against these headwinds from these people if you approach it that way or think about.

[00:05:23] I mean, it's everything from, you know, hub and spoke models, COEs, you know, things like that versus, you know, where you can put your best foot forward in terms of storytelling with data.

[00:05:36] Or it just seems like every scenario that someone brings up that needs help with, there's always a whole bunch of people that just jump in.

[00:05:44] And you don't find that a lot of communities are just, you know, you get asked that kind of question and it's just kind of crickets.

[00:05:51] And you're just hoping that one of these days, but if you've got a timely question, you may not get an answer in a lot of groups.

[00:06:00] I completely agree. I talk about this, I actually spoke with somebody about that last week.

[00:06:05] I think there's an element around the maturity of the industry and the function, which helps drive that.

[00:06:10] And, you know, 100% it's the people that are doing this and I'd never take anything away from the crowd.

[00:06:15] But because the function itself and the topic is fairly immature, we've all got questions.

[00:06:22] And so we're all open to helping each other because we know next week it could be us that's asking the question.

[00:06:28] So luckily it's quite transparent and quite fluid like that.

[00:06:32] You know, there are other elements of the TIC that are more traditional people analytics or compensation or strategic workforce planning.

[00:06:40] And you do see that the older and more mature the function, the less engagement.

[00:06:46] I don't think that's the people.

[00:06:48] I think the people are as curious and as inquisitive.

[00:06:52] But it's not quite the same kind of wild west that you see with the external labor market data and the pure TI piece.

[00:06:59] But then you do get a huge amount of conversation of how these things tie together.

[00:07:02] So how does external labor market data tie into strategic workforce planning, tie into people analytics, tie into compensation, etc.?

[00:07:09] That you're definitely seeing a lot of questions around, a lot of energy around.

[00:07:13] But I think it's great.

[00:07:14] I think it's great that the community is so open, that the vendors are so friendly with each other.

[00:07:19] You know, you don't go into it and see this kind of real animosity and a real prickliness.

[00:07:24] They're all trying to work their way through this.

[00:07:28] And I think that's, I mentioned earlier that I think we're only scratching the veneer of what's possible.

[00:07:33] I think that's because of that, the vendor landscape knows that all ships can rise with a rising tide.

[00:07:41] You know, there is more than enough for everyone to go after.

[00:07:44] Yes, it can be tight at times with, you know, the last 18 months, two years.

[00:07:48] We've seen some budgets smashed in talent acquisition and HR.

[00:07:52] So, yes, budgets have been tighter and teams have been tighter, etc.

[00:07:56] But when you think about the potential of this work more broadly than HR, there are so many opportunities to go after.

[00:08:03] And I think that's, I hope that that kind of friendly ethos and friendly vendor environment never changes.

[00:08:12] I think, I hope it really stays true.

[00:08:15] Yeah, I do too.

[00:08:16] I think about, just to your point that this has broad implications, you know, beyond just this insular kind of niche group.

[00:08:24] I spoke to John Strauss recently from Greenhouse.

[00:08:28] And this came up invariably, not just because of what they do and the role that they play in the sort of talent acquisition technology landscape as a primary sort of hiring platform.

[00:08:41] But his book called Talent Makers that he wrote with his co-founder really talks about the audience isn't, that book isn't really like talent leaders per se, like in the HR domain, although they could get value from it too.

[00:08:57] But it's like every person who's a people manager or aspires to be a people manager who will be doing hiring and unfortunately maybe some firing and promotions and putting people into leadership development or finding them a coach or a mentor.

[00:09:12] Every single person has some responsibility to think about what this all means and how the pieces need to move around.

[00:09:23] So everyone needs to learn how to play this form of chess, I guess.

[00:09:29] I think that's thinking a few moves ahead, understanding what direction to look.

[00:09:34] Am I understanding the trajectory of some of these things?

[00:09:38] Do I understand the incremental value of integrating these data sources and, of course, being responsible with all of that data and potentially AI as well?

[00:09:50] But there's a lot that people need to pay attention to.

[00:09:53] And so I think the talent space is one of those areas where, I mean, I guess you could argue technology itself is important for everyone to at least have a little bit of understanding of as well, hopefully.

[00:10:05] But yeah, it's just one of those domains that there's no one sort of spared.

[00:10:10] Everyone's gone in and out of the labor market and understand some of the dynamics there and it gets complicated.

[00:10:15] But yeah, I do think it's important to spread beyond what we're doing.

[00:10:20] But hopefully people are coming and learning from the group and then going back and dispersing back to their organizations and explaining what they've learned.

[00:10:30] I completely agree.

[00:10:32] Ultimately, everyone in an organization becomes a recruiter, particularly in an aggressive and a hungry market.

[00:10:37] So, you know, in a quieter market like we've got now, sure, I can see hiring managers not working their networks quite as strongly as they would have in other times.

[00:10:46] But broadly speaking, you know, the philosophy that everyone should be recruiting for the company kind of holds true, I think.

[00:10:52] And I think that probably will hold true across talent intelligence eventually as well.

[00:10:56] I think you should be understanding the context in which you're working.

[00:11:01] You should be understanding how your competitors do things.

[00:11:04] You should be benchmarking.

[00:11:06] You should be looking at the external market and understanding what does this mean to me as a function.

[00:11:11] And I think that's as true for corporate functions, whether it's M&A or strategy or sales, biz dev, marketing, whatever it may be, as it is for the more kind of enablement type functions and the more traditional corporate functions.

[00:11:25] Whatever we do, there is an external context.

[00:11:27] And I think for too long, as companies, we've run ourselves in a very insular way.

[00:11:34] We've looked internally.

[00:11:35] We've run things internally, whether it's our spans of control or our strategy or models, whatever it is.

[00:11:40] There's an external influence here.

[00:11:42] And we haven't really, for me at least, acknowledged that fully.

[00:11:45] But I think that really is an everybody problem.

[00:11:48] It shouldn't be restricted to the TI team at the core or a number of business analysts or people in analytics experts.

[00:11:54] We should all have a fundamental understanding of what is the environment we're working within from a competitive perspective or a market perspective.

[00:12:04] Yeah, absolutely.

[00:12:05] And I think you've got to also look at that no matter what the labor market is doing.

[00:12:12] And maybe you go through periods of high attrition.

[00:12:16] And now I think you were just writing about this recently, about now we're in this period of low attrition.

[00:12:23] People don't, they're scared to leave or whatever the circumstances are, people aren't leaving as much.

[00:12:30] So I don't know if you want to unpack some of your analysis there, because I think it's interesting.

[00:12:37] So full credit to Remy and the guys at Revello.

[00:12:41] So essentially when you're looking at the labor market, and it's like most things in TI, honestly, the data is really backing up what your gut feel would be saying.

[00:12:52] I think more often than not the labor market, people kind of feel naturally what's going on.

[00:12:57] And, you know, the reality is the market tightened.

[00:13:01] There were less jobs because of that.

[00:13:02] There were less people looking to move into those jobs and there weren't as many jobs to go into.

[00:13:08] So we had the layoffs.

[00:13:09] And then that was compounded with the fact that people didn't want to move into new jobs, be last in, first out, et cetera.

[00:13:15] So there was a nervousness.

[00:13:16] That's meant that there's a lot of people that are sitting in jobs at the moment that they wouldn't normally, in inverted commas, have been in.

[00:13:24] And, you know, attrition rates have been historically low.

[00:13:27] Yes, there are still jobs out there, but they're not really matching terribly well to the people that are looking for those jobs.

[00:13:34] So we've got this situation where everything is just a little flat.

[00:13:38] You know, the market's flat, the attrition rate's flat.

[00:13:41] And it doesn't take a huge amount to get that going again and to get that machine ticking over again.

[00:13:48] And in the research, we looked at kind of 10% to 15%.

[00:13:51] I think there's a part of me that thinks that that's a bit conservative.

[00:13:55] That was kind of the safe economist view.

[00:13:58] In my mind, I see it more as a snowball effect where I think it would probably be actually be something smaller than that that would trigger the initial point where things do start loosening.

[00:14:08] I think people still are nervous when you look at candidates' desires and behaviours, etc.

[00:14:14] Security is the number one driver at the moment, both in the short term and the medium term.

[00:14:19] People want that security of job, both in terms of getting a new job.

[00:14:23] And then once you're in a job, they want a security and a securing job company after that as well.

[00:14:28] So I get it.

[00:14:29] I think it's a market where things will change quickly.

[00:14:33] I don't think it would take too much to turn things on its head quite aggressively.

[00:14:38] I think the caveat we're all working within is what's going on with Gen AI.

[00:14:42] We're in a weird situation where I think there was a lot of promise over the last year or so.

[00:14:48] A lot of companies got very excited thinking, oh, we don't have to have any employees anymore.

[00:14:52] We just automate everything.

[00:14:53] And then each new model came through.

[00:14:56] And it was like, this is up to a high school standard.

[00:14:59] This is up to a graduate standard.

[00:15:00] This is now PhDs now, etc., etc.

[00:15:03] I think a lot of companies are still struggling to work out, how do I actually operationalise this stuff?

[00:15:07] Beyond making some fun tools and some initial time savings, how do I actually build this into my ecosystem to make this stuff work?

[00:15:14] So I think they've been holding back on hiring because of that as well.

[00:15:17] Obviously, there's a broader economy piece as well.

[00:15:19] But the broad economy, the stock market is doing well.

[00:15:21] The broad economy is actually not too bad.

[00:15:23] So I think we probably will see a phase going into next year where companies go, well, actually, if I'm not seeing that Gen AI stuff kicking through, if the vendor landscape isn't bringing this stuff out as a default for my new product stack or whatever to automate things, I probably am going to need to start hiring people until that changes again.

[00:15:42] Whether there's an interim step and you use the people power to stabilise the systems, processes, tools, and structure to then look to automate, I don't know.

[00:15:52] But I think the economy overall and the market overall is definitely, it feels like it's in the starting blocks of a race.

[00:15:58] It's sitting there, sitting on its heels a little bit.

[00:16:00] And I don't think it will take too, too much to trigger that going forward again.

[00:16:05] I guess one of the things I was curious about was just because people are staying, are we tracking engagement and things like that?

[00:16:14] Like, are they staying out of necessity?

[00:16:16] Are they staying and saying, you know what, I think I could see myself here because the company's waking up and they're starting to think about upskilling and internal mobility investments and things like that.

[00:16:30] That they weren't before or whatever.

[00:16:32] So are companies listening, you know, doing social listening and doing their engagement stuff and are they seeing some gains there as well?

[00:16:42] Or it's just like, because otherwise it's like, oh, people are scared to leave.

[00:16:46] But, you know, it's like I don't want people to stay because they're scared to leave.

[00:16:52] Yeah, it's like, you know, are you really at the end of the day, you know, do you still think of this as a win-win?

[00:16:58] Or are you just sort of riding out the storm and you're going to bolt as soon as, you know, the wave comes the other way?

[00:17:06] Yeah, I think it's a good question.

[00:17:07] So when you look at the sentiment scores online through the well-known sentiment platforms,

[00:17:14] are people leave these reviews?

[00:17:15] Most companies that were going through the squeeze in the last two years or so followed a similar path.

[00:17:21] And to what degree change?

[00:17:25] But they've followed a similar path where most of them saw their rankings and their scorings drop quite considerably.

[00:17:32] And then as things start to stabilize, it picked back up.

[00:17:35] And that can be a bit of a survivor guilt.

[00:17:39] And, you know, I'm so thankful to be here, et cetera.

[00:17:42] But equally, I do think that there is an element of, as you say, that internal talent management piece where it's,

[00:17:48] well, actually companies have looked at it and gone, we don't want to spend all our money bringing new people in that are more expensive than our current employee base.

[00:17:55] Let's invest some of that money into keeping our current employee base and upskilling and et cetera, et cetera.

[00:18:00] So I think there is an element of that going on.

[00:18:03] I think that probably still is an underlying element where people are accepting and that they probably aren't seeing other talent management levers being pulled that they'd like to be seeing pulled.

[00:18:16] So, you know, you are seeing a period where promotions, for example, are a record low.

[00:18:21] Spans of control are being changed and you're suddenly having far fewer management opportunities, et cetera.

[00:18:26] So I think you will see knock-on effects off that where people are sitting there going,

[00:18:32] I was quite happy being a manager.

[00:18:33] I didn't want to become an individual contributor.

[00:18:35] I'm going to move on as soon as I get an opportunity to.

[00:18:37] Or, you know, obviously elephant in the room, we've got RTO going on with return to office, return to teams, et cetera.

[00:18:44] That's a big, big piece where I think when you look at the data, the candidates don't, broadly speaking, want to be forced back to an office, broadly speaking.

[00:18:55] Hybrid seems to be the most common option at the moment.

[00:18:57] But the candidates, you still get a disproportionate number of applications for anything remote.

[00:19:02] The candidate base still wants this.

[00:19:04] I think as the market starts opening up again, you probably will see a pendulum swing where it kicks back and you do have that pent-up attrition.

[00:19:14] I don't think it will be like for like.

[00:19:16] So if you think most companies would be 10%, 15% attrition year on year, they might be sitting at 4% or 5%.

[00:19:21] I don't think you're going to see that whole extra 8% bundling for next year.

[00:19:27] I don't think we're going to be looking at kind of 20% attrition rates or anything.

[00:19:30] But I think that we will see a swing back in my opinion.

[00:19:35] I think as soon as that market starts to loosen, they did the appetite.

[00:19:40] Hey, this is William Tencup, Work Defined.

[00:19:43] Hey, listen, I'd like to talk to you a little bit about Inside the C-Suite, the podcast.

[00:19:47] It's a look into the journey of how one goes from high school, college, whatever, all the way to the C-Suite.

[00:19:54] All the ups and downs, failures, successes, all that stuff.

[00:19:57] Give it a listen.

[00:19:58] Subscribe wherever you get your podcasts.

[00:20:01] For return to office is just so low across that candidate base and the environment is creating.

[00:20:07] I just don't see it being a stable solution for a workforce that's got used to working from home.

[00:20:14] We had a couple of years with record innovation, with record profits, with record growth.

[00:20:19] To then say to those same people, oh, actually, it turns out you can't work remotely very well.

[00:20:24] You aren't very innovative.

[00:20:25] And they're sitting there going, well, I just was for two years.

[00:20:28] You're quite happy with me working remotely for two years during this period.

[00:20:31] I think that that's not landing as people would wish.

[00:20:35] So I think there will be a pushback.

[00:20:36] I think there will be a pendulum swing back again.

[00:20:38] I mean, I haven't been in the talent space as long as you have, but it just seems like, yeah, it's a constant, you know, back and forth.

[00:20:45] And you ride the wave.

[00:20:46] And some of it is decisions made on, you know, forward-looking indicators.

[00:20:50] And some of it is all just reactive, you know, what has transpired and, you know, forcing companies' hands to say, look, this is not working for whatever reason or for whatever indicators we're using.

[00:21:03] But, you know, all the more reason to embrace, you know, talent intelligence more broadly, right?

[00:21:09] And understand what's going on both inside and outside the organization.

[00:21:13] Absolutely.

[00:21:14] And you can get into some fairly fundamental organizational psychology when you're looking at kind of the demographic of leadership and how they like to work and how they've gone through their entire career versus the workforce coming through, for example.

[00:21:26] You know, the flip side being there is something hugely energizing about walking through an office and it's really busy and there's loads of people and there's energy there.

[00:21:36] That is a hugely empowering and energizing thing compared to when you walk through and it's a ghost town.

[00:21:41] So I can be quite a contradictory human.

[00:21:44] I get it.

[00:21:45] I can see the drive and I can see the desire behind it.

[00:21:49] I'm just not sure candidates as a whole and the workforce as a whole care about the populace.

[00:21:57] They care about their situation and their personal situation.

[00:22:00] And is that busyness in the office and that look and feel and the energy worth them sitting four hours a day on a train for?

[00:22:09] I'm just not convinced it is.

[00:22:11] I was actually in London last week and I was in a taxi and the taxi driver was bemoaning people that work remotely.

[00:22:19] And said, you know, how terrible it was that they weren't coming into London more and supporting the taxi drivers and supporting the London economy, etc.

[00:22:28] And when I mentioned, well, that their money is being spent in the local economy where they live, you know, should they spend four hours a day to come in to get a taxi with you and spend the money in London?

[00:22:39] And it's a very contentious issue.

[00:22:42] You know, it's putting whichever which way you go.

[00:22:45] It's fundamentally changing how people live, how they structure their lives.

[00:22:51] It's a very emotive topic.

[00:22:53] It's very difficult to get it right.

[00:22:55] No, I agree.

[00:22:56] You know, I think we experienced the same thing here.

[00:22:58] I mean, I'm half hour train to Grand Central Station in New York City.

[00:23:02] And yeah, there's a lot of people that just haven't or they go in, you know, a couple of days a week.

[00:23:07] But there's people that just insist like I've got to everyone's got their reasons, right?

[00:23:12] Like I got well, I got to get out of the house or too many distractions at the house or I don't.

[00:23:17] Yeah, I don't I don't want to spend, you know, hour and a half each way, you know, commuting, especially if you drive.

[00:23:23] Right.

[00:23:24] Then you can't do other things.

[00:23:25] At least if I take the train, I can listen to a podcast and read and, you know, do other things.

[00:23:30] So, yeah, I think everyone is is different.

[00:23:34] And everyone, you know, I think COVID had a lot of people just reassessing what's important to them.

[00:23:41] And you could just as easily say, you know, it's more important to me to support, you know, local, you know, businesses.

[00:23:48] This is where I live.

[00:23:49] This is where I'm raising my kids.

[00:23:52] And, you know, it just seems like you can't worry about every cab driver in the city two hours away and how this has changed their livelihood.

[00:24:02] Right.

[00:24:02] I think it's just the transparency there.

[00:24:04] And I think this is one thing that personally I always love in employees and I think is the most important is as long as you know.

[00:24:13] As long as people are transparent and you know the playing field, I think, you know, most people will look at it and go, I'm willing to make that deal.

[00:24:22] So, for example, and this is me talking theoretically, I don't want to go to London on a daily basis.

[00:24:29] So I don't go to London on a daily basis.

[00:24:30] I know that has limited my career in the past for different approaches and different employers.

[00:24:35] But I'm willing to make that sacrifice.

[00:24:37] That's a deal because it's very transparent.

[00:24:39] If I'm not in the office, I'm not going to get that job.

[00:24:41] Fine.

[00:24:41] I'm comfortable not getting that job.

[00:24:43] I think it's having that transparency of knowing what am I sacrificing for this and what am I willing to sacrifice for this?

[00:24:50] I think that's the bit where most employers aren't being truly transparent at the moment.

[00:24:55] And that's where some of the friction and difficulty lies.

[00:24:58] If that transparency was there and it said, well, actually, yeah, you aren't going to get as many promotions or you won't get seen as much and your performance review will be impacted because of that.

[00:25:07] Whatever it is, as long as we're transparent, people know what they're signing up for or signing out of.

[00:25:12] I think people are quite understanding.

[00:25:14] But having the knowledge to be able to make that informed choice.

[00:25:18] And I think that's the difficulty when you don't have that knowledge and you don't know what you're being signed up for or not.

[00:25:25] That's where it gets difficult.

[00:25:27] Yeah.

[00:25:27] I also think about you've got a plenty of people who go in and they are energized by what's there.

[00:25:32] And maybe their team is more collaborative and they use the meeting rooms and they use the whiteboards.

[00:25:39] And, you know, they go through those types of exercises.

[00:25:42] And there's other people who just come in and they're completely overwhelmed by the noise and the movement.

[00:25:48] And if you want everyone to bring all of themselves to work, you've got to appreciate that not everyone is different.

[00:25:55] We're not just all the same replaceable sort of cogs in the machine.

[00:26:00] Absolutely.

[00:26:01] Absolutely.

[00:26:01] I think that comes back to whether it's the working environment, whether it's how many days in the office per week, etc.

[00:26:09] People want to be treated like a grown up, like fundamentally.

[00:26:13] If you force anyone fully remote or force anyone fully on site, they're going to push against that.

[00:26:19] No one must be forced to do one or the other.

[00:26:21] Similarly, if you go fully plan office, people are going to push against it.

[00:26:25] If you go fully cubical, people are going to push against it.

[00:26:29] The workforce is a really complex beast with nuance throughout.

[00:26:32] We need to have that flexibility and fluidity to work in the most effective ways.

[00:26:37] And I think the overlay to that is obviously nowadays more than ever, you have fully global virtual teams that would link into this.

[00:26:44] So how do you suddenly build that team environment when if you're being pushed back to an office, for example, you could be on the next floor up.

[00:26:52] You could be the office block over.

[00:26:53] You could be the other side of the continent.

[00:26:55] You could be the other side of the world.

[00:26:56] But it's kind of irrelevant by that point.

[00:26:58] If you're not all sitting in the same specific desk allocation, as a team, you're not feeling that team unity and that team environment necessarily.

[00:27:07] So I think we've got to be growing up about all of this stuff.

[00:27:10] There's going to be trade-offs.

[00:27:12] And sadly, it's not going to work for everyone.

[00:27:14] We know, for example, data, return to office is one of the worst things you can do for diversity.

[00:27:20] Whether it's socioeconomic, whether it's disability, et cetera.

[00:27:25] It's terrible for diversity.

[00:27:28] There's got to be some kind of accommodations and trade-offs made here.

[00:27:31] Yeah, I was going to say, part of it is the accommodations you might need to make.

[00:27:35] And part of it is, depending on where you live and how expensive it is to live in different places, you may have a longer commute to get to the office.

[00:27:44] And there's all kinds of variables.

[00:27:46] I agree.

[00:27:47] One of the things I was thinking about, just sort of pivoting towards the AI topic a little bit.

[00:27:53] But we're already trying to understand from a sort of skills intelligence, I'll call it perspective.

[00:28:01] Like, what skills do we have?

[00:28:03] What do we predict are the skills we're going to need based on the advancements in technology, based on the projects that we're doing?

[00:28:11] Do we understand exactly where we might need to reskill or upskill people or hire net new people for specific skills?

[00:28:20] But then you have this AI dimension that says, well, AI is already taking on quite a few, at least, tasks.

[00:28:28] And now with AI agents, you've got some workflow capabilities beyond what a robotic process automation tool might have done eight, even 10 years ago, somewhere around there.

[00:28:40] But I was thinking about this in the context of return to office.

[00:28:45] Like, has anyone actually mapped, well, these are the tasks and these are the workflows that AI could be doing?

[00:28:54] And then also overlay that on the rationale for returning to the office.

[00:28:59] Were any of those tasks ones that previously needed to be done in an office for whatever reason?

[00:29:06] And therefore, the trajectory of AI actually reduces or lessens the rationale for returning to the office?

[00:29:16] Honestly, not that I've seen.

[00:29:17] I think it would be fascinating research to look at.

[00:29:19] I think the AI automation piece for me feels very similar to where we were when we're looking at global offshoring, maybe 15, 20 years ago,

[00:29:31] where I was looking at the role complexity and how frequently things were repetitive, et cetera, and distance to client base or stakeholder, et cetera.

[00:29:40] It doesn't feel dissimilar now where we're at that stage of what's the stuff that's fairly standardized and that we could lift and shift relatively easily.

[00:29:50] I haven't seen any study with the overlay of the return to office and whether we need to be close to the customer or not.

[00:29:56] I think where it will get really interesting, though, is with the cost of a lot of these AI systems increasing,

[00:30:02] obviously a lot of these providers are eating up the cost at the moment to try and get market share, et cetera, et cetera.

[00:30:08] But that won't be forever.

[00:30:10] As that increases and the number of actions these AI agents need to perform increases exponentially because of the complexity of the systems and the models,

[00:30:18] I could quite feasibly see a time where, yes, we could automate.

[00:30:23] Yes, we could have AI agents, but it still may be more cost effective to be running an offshore center in one of the local countries,

[00:30:31] which I think is a fascinating dynamic where suddenly we are the low-cost option versus automated tech,

[00:30:37] which feels absurd to even think, to be honest.

[00:30:39] Yes, I'm sure those costs will come down, but yes, where do you cross that threshold?

[00:30:47] It's a pretty complicated calculation to try to make with so many variables that are constantly ebbing and flowing.

[00:30:54] And I think as you get more data, more trustworthy data, maybe you've got to make some calls within a domain,

[00:31:02] or maybe it just does depend on the speciality where those functions exist.

[00:31:07] I always just think about how broadly do I need to cast a sort of warning to the people who aren't paying so close attention to all of this

[00:31:17] when we think about the impact and the potential impact when, oh, you're in this role or this industry,

[00:31:24] you should be prepared in the next two to three years.

[00:31:28] At least give them a horizon, right?

[00:31:31] Is this an H1, H2, or H3 problem for me and where I want to be?

[00:31:36] And then how do I think about that as I plan the next few steps of my career?

[00:31:40] I think it's really hard.

[00:31:41] We've seen a lot of these kind of predictions in the last couple of years, and I think a lot of it hasn't come through.

[00:31:47] I think when I look at a lot of the Gen AI stuff, I think we're probably at a bit of a crossroad period

[00:31:53] where we've had a couple of years where there's some really amazing advancement in terms of the models and the capabilities.

[00:31:59] But fundamentally, the products haven't really changed.

[00:32:03] A lot of the way most consumers will be interacting with this stuff hasn't really changed.

[00:32:07] A lot of the way that the systems integrate hasn't really, really changed greatly.

[00:32:14] A lot of it's that kind of large language model.

[00:32:16] It's the language piece.

[00:32:17] It's the prompt engineering and the language output.

[00:32:20] So I don't think we're – I think it's that piece of I'm surprised how little has changed in the, what, two years since all this started coming through.

[00:32:29] You know, you've really got very few systems, processes, tools, platforms that have really taken hold of this beyond some bits that are essentially a re-skinning of the chat GPT API or whatever it is.

[00:32:42] Having said that, I think – and this kind of ties to your point around what skills do I think are going to be needed –

[00:32:48] I think once we've got the data engineering and the infrastructure back end and we understand that better of how all these different systems hand off to each other,

[00:32:57] I think we could see things changing quite quickly.

[00:33:00] You know, there's been lots of talk already around the impact on software engineering and how much code has been generated.

[00:33:07] That feels very similar to what we saw in the very initial early days when we saw journalism being hit hard and suddenly you needed editors, you didn't need the content creators.

[00:33:17] I think we're probably going to see quite a lot of that where we say, well, actually, we need someone on making a decision point.

[00:33:22] And this – you know, we spoke in the past about human-centered AI.

[00:33:24] I think we need very clear rules and engagement and we need to be very clear around the decision points are still the human element.

[00:33:32] But a lot of that heavy lifting, you really don't need the people for.

[00:33:36] And I've written about it around talent acquisition.

[00:33:39] I think there's a huge sway to the talent acquisition ecosystem that we're nervous to let go of because we want to hold on to it because we think it's our power base.

[00:33:49] But the reality is you could probably have hiring manager self-service and some will automate a lot of that process.

[00:33:55] Does a human need to be involved in that process?

[00:33:57] Absolutely.

[00:33:59] Does that human need to be a recruiter?

[00:34:01] I'm not convinced it does.

[00:34:02] I think for the higher value, higher touch, more niche elements, 100% you want a person in that.

[00:34:08] You want a talent acquisition person tied in.

[00:34:11] But I think we're at a period where we are going to see a lot of roles getting redefined.

[00:34:15] What is the value in this process versus what is the operational output of the process?

[00:34:21] Yeah.

[00:34:22] Yeah.

[00:34:22] No, I'm glad you sort of pivoted in that direction because that has to always be top of mind when you're thinking about is it build versus buy versus bot?

[00:34:30] That kind of calculation, like which direction should I go?

[00:34:34] And you've got to play it out, right?

[00:34:35] You've got to play it out a couple extra steps to say, well, what impact would this be?

[00:34:39] If I used AI where it was kind of previously the domain of human decision makers in terms of their empathy and their creativity and other just sort of judgment, you've got to think about those things.

[00:34:55] And when you take some of those things away from people, there are going to be ramifications.

[00:35:02] There are going to be some that say, you know what?

[00:35:05] If I have to give up something out of all the things that I do, I'm comfortable with that.

[00:35:10] And there's other people who are going to be like, no, that's what I love about doing this job.

[00:35:17] That particular facet of the job is actually my favorite thing.

[00:35:21] Even if you're a people manager and you wanted to be a people manager because you're a good coach and mentor and you're a servant leader and all these great things that make you fit for that type of role.

[00:35:34] And then you start saying, oh, well, hey, we're going to deploy this AI coaching app.

[00:35:40] So I just had this conversation with one of my previous guests.

[00:35:43] I was in another group and he was like, well, I'm trying to put together a presentation to talk about management and where AI tools can complement the manager.

[00:35:54] And I actually had to pause before I said anything at all because I'm like, well, wait a minute.

[00:35:58] The AI tools that I know are actually the ones that should make a manager a valuable manager and leader.

[00:36:05] And so are you talking about just AI tools that might automate some of the performance management tasks or portfolio management over all the projects on my team or whatever?

[00:36:16] So I think to your point earlier that we're still very early days in trying to figure all this out and figure out the human impact.

[00:36:23] I want to take a break real quick just to let you know about a new show.

[00:36:27] We've just added to the network up next at work hosted by Gene and Kate Akil of the Devin Group.

[00:36:36] Fantastic show.

[00:36:37] If you're looking for something that pushes the norm, pushes the boundaries, has some really spirited conversations.

[00:36:44] Google up next at work, Gene and Kate Akil from the Devin Group.

[00:36:51] And of all of this.

[00:36:53] And I know you and I both spend a lot of time making sure that we are advancing things with a human centric and responsible lens.

[00:37:01] Yeah, I can agree more.

[00:37:03] I read a paper.

[00:37:04] It was last week.

[00:37:05] And I'm probably going to murder who did this.

[00:37:07] I think it was McKinsey.

[00:37:08] I'm probably wrong on that one.

[00:37:09] But it was around algorithmic managers and algorithmic leadership.

[00:37:13] And it was all based around the systems making decisions and managers being disempowered.

[00:37:19] And it was fascinating for me because I thought, to your point, why do people get into management and leadership?

[00:37:26] And what do they want to get out of it?

[00:37:27] And 100%, you can use systems to absolutely empower your leadership to make better decisions,

[00:37:34] to have more timely conversations, to have more timely interventions,

[00:37:38] to push on L&D pullers or levers they need to pull, etc.

[00:37:43] 100%, it can turbocharge and supercharge those leaders.

[00:37:46] We've got to be really careful that the systems don't get into a state of absolving the manager of responsibility,

[00:37:52] of management or leadership.

[00:37:54] And suddenly that manager or leader leans on the system.

[00:37:57] And the system is then managing you.

[00:37:59] The system is saying you're late for your shift.

[00:38:01] Your KPIs are down.

[00:38:03] You're going to have a performance improvement plan, whatever.

[00:38:06] We could be really careful.

[00:38:07] I think a lot of the things that we traditionally hold true of a lot of our work is operational elements.

[00:38:14] This is a, I think there are some fundamental periods of change going on where it really is calling into question what is valuable to us,

[00:38:22] what's important for us as humans in a workforce.

[00:38:26] And I think it calls a lot of it into question.

[00:38:28] And I think that's exciting and terrifying in equal measure.

[00:38:32] It's a super interesting period where roles will change, role designs will change, how we engage, how we work,

[00:38:40] whether we're in a structured hierarchical system, whether we're in an organic, fluid organization,

[00:38:45] whether we're skills-based hiring and we're fully fluid and we're only working on projects the whole time,

[00:38:51] internal gig workforce, et cetera, et cetera, right through to algorithmic managers and full-on algorithmic strategic workforce planning.

[00:38:59] I think it's a fascinating period.

[00:39:01] Yeah, I think there's so much at play.

[00:39:03] You have all these tools and you start trimming the fat in the middle, right?

[00:39:07] The middle managers, are they in jeopardy?

[00:39:09] But then you have leaders, assuming you're still in a hierarchical structure and you've just flattened it.

[00:39:16] But now that leader has many more people to manage.

[00:39:21] And so you're managing the people, you're managing some of the AI tools that are trying to help, I guess,

[00:39:28] fill in some of the gaps and even just from a bandwidth perspective.

[00:39:32] Well, if I have to reschedule your one-on-one, but you need to talk, could you talk to this AI tool or whatever?

[00:39:39] There are going to be some people that are comfortable with that if that is an option.

[00:39:43] And there's some people who are going to be like, no, this is a nuanced, relatively sensitive conversation.

[00:39:49] And you know me, Mr. and Mrs. Manager, and I know that you have the sensitivity and the empathy to deal with this situation.

[00:40:00] And I don't care how intelligent this AI agent, it's not a proper substitute,

[00:40:06] which in some cases goes back to what we talked about before, which is, you know, AI is great for certain situations,

[00:40:13] but you shouldn't just deploy it because you can.

[00:40:16] You should deploy it because you should, and it's an appropriate use case.

[00:40:20] It's not going to create backlash or jeopardize the human centricity of the effort overall.

[00:40:26] Couldn't agree more.

[00:40:27] You made me think of an article, right?

[00:40:29] I want to say earlier this year, it might have been the end of last year, I can't really remember,

[00:40:33] but it was around digital twins.

[00:40:35] And if we fully went down the path of digital twins and, you know, your internal systems at work,

[00:40:41] they could record all your phone calls, they could record all your video,

[00:40:44] they can see all your output, they can see all your one-to-ones with your employees.

[00:40:48] They could probably make a fairly, even now, they could probably make a fairly decent effect

[00:40:51] of what you are as Toby in a given work environment.

[00:40:56] And I think you probably could have that as an AI-powered sim that you could dial up or down on empathy

[00:41:02] or dial up or down, and you probably could, you know, genuinely create something

[00:41:07] that would probably do a better job managing than I can do.

[00:41:10] Whether that's appropriate or not, I don't know.

[00:41:12] But also, who owns that?

[00:41:13] That's the bit where I get fascinated by is if I decide to do that off my own back

[00:41:18] and I run my own avatar and, you know, all my one-to-ones,

[00:41:21] suddenly it's with Gen.AI-powered avatar, Toby.

[00:41:25] Should that be a company's IP?

[00:41:27] Should that be my IP?

[00:41:28] Is that one employee that gets paid if I run all my team's one-to-ones at the same time

[00:41:34] and I'm doing, you know, 10x output?

[00:41:36] Should I get 10x salary?

[00:41:37] Like, how does this work?

[00:41:38] How do we operationalize it?

[00:41:40] I don't think anyone's getting near those questions.

[00:41:44] And I think it's a really fascinating topic of productivity

[00:41:46] and performance management and output delivery.

[00:41:49] Well, actually, if I do have the power, same as a 10x engineer or 100x engineer,

[00:41:53] like, if I do have the power to turbo-guage my output

[00:41:56] through the skills I've acquired and the tools I use,

[00:42:00] should I get 10x the salary or whatever the number is?

[00:42:03] Like, how does this operationally work?

[00:42:04] And I don't think we're anywhere near addressing that stuff yet.

[00:42:07] I do think the success metrics and the way we incentivize people needs to be perhaps rethought.

[00:42:15] And that's a good example.

[00:42:17] You also just had me going off on a completely dystopian thought of forcing people to come

[00:42:21] into the office and then bringing them into these little rooms where they have their one-on-ones

[00:42:25] with a digital twin of the manager.

[00:42:28] While the manager is like, you know, the eighth hole of the golf course.

[00:42:32] It's just wrong at so many levels.

[00:42:35] The terrifying thing is you could imagine it happening.

[00:42:38] That's the terrifying thing in the current world.

[00:42:40] Easily.

[00:42:40] Totally see it happening.

[00:42:41] I don't know if that's a Black Mirror episode or, you know, the sequel to Office Space.

[00:42:45] I'm not quite sure.

[00:42:47] It's a terrifying prospect.

[00:42:48] It really is.

[00:42:49] The same as, you know, my team toyed with virtual reality a few years back now

[00:42:54] and having VR meetings.

[00:42:56] And we'd have the VR meeting room.

[00:42:58] And it was genuinely really interesting.

[00:43:01] Really interesting tech.

[00:43:02] Even then, when it was fairly new tech still, it definitely felt more like you were in the

[00:43:07] same meeting room versus a video call.

[00:43:10] Sadly, it made half the team seasick.

[00:43:12] And so we didn't last very long.

[00:43:15] But I do think that there are more interesting solutions.

[00:43:18] With the tech nowadays, I just think there are more interesting solutions than getting on

[00:43:23] a train or a bus or sitting in a car for 30 minutes or an hour, two hours to sit at a cubicle.

[00:43:31] That just doesn't...

[00:43:32] And then get on a Slack call or a Zoom call or whatever it is.

[00:43:35] It just doesn't feel like we're using the tech that's available.

[00:43:40] It's a strange paradigm where we've got this amazing tech doing some wonderful stuff

[00:43:44] and pushing so far in some directions.

[00:43:47] And yet, how we're operationalizing that and working on a day-to-day feels very juxtapositioned,

[00:43:53] some of that.

[00:43:53] I've also been thinking a lot about what tasks to give to AI.

[00:43:57] And even in...

[00:43:58] You and I are both producing content.

[00:44:02] And it's like, I think about how people have outsourced some of that work to AI.

[00:44:10] And then you just sprinkle your own human lens on top of it.

[00:44:15] And I guess I prefer the reverse.

[00:44:18] I know...

[00:44:19] Look, I get writer's block probably more frequently than most.

[00:44:23] But part of that is because I...

[00:44:25] Yes, I know AI tools are out there.

[00:44:27] And I could just put in a couple prompts and get something decent.

[00:44:31] But it just looks so formulaic and homogenized.

[00:44:34] And it's like, this is where we still need human creativity.

[00:44:38] And the writing acumen that really tells good stories.

[00:44:43] And it's compelling.

[00:44:43] And it evokes feeling and emotion and compassion.

[00:44:48] And I just...

[00:44:50] I'd rather have the AI be like the editor to say,

[00:44:54] you might want to...

[00:44:56] You could strengthen this paragraph or whatever.

[00:44:59] But for it to write and then a human come in and just tweak it,

[00:45:02] you're still going to be able to tell that AI was very involved in the process.

[00:45:08] And maybe part of this is because I have...

[00:45:10] I've been talking to my own local board of education

[00:45:14] and the school administrators about AI,

[00:45:17] even before university, right?

[00:45:19] In high school and maybe even earlier,

[00:45:21] the same way we were exposed...

[00:45:24] Well, I'm older than you.

[00:45:25] But the same way I was exposed to computers

[00:45:28] when I was in...

[00:45:29] I was probably 11, 12 years old

[00:45:32] when I got my first computer,

[00:45:33] both at school and at home.

[00:45:35] And then internet and then social media, whatever.

[00:45:37] So we've seen these waves before

[00:45:39] where people had to figure out

[00:45:41] what is the appropriate way to leverage technology,

[00:45:44] but still being with a brain that can think for himself.

[00:45:50] But it just...

[00:45:51] There's a lot to unpack there.

[00:45:53] But I just feel like,

[00:45:54] please don't give up the things that make you unique

[00:45:57] and make you uniquely human.

[00:46:00] I completely agree.

[00:46:01] And I think it comes back to two things for me.

[00:46:04] One is whether you want to be average.

[00:46:07] All these systems definitionally are average.

[00:46:09] They tend to the average.

[00:46:12] There's essentially predictive text on steroids.

[00:46:15] It's statistically what's the most likely next word for this.

[00:46:18] So they're going to tend to the average.

[00:46:19] And the problem with that is

[00:46:21] when the volume of output

[00:46:24] just increases exponentially from these systems,

[00:46:27] they're going to all tend to the average.

[00:46:29] And that's why you're going to get

[00:46:30] this kind of homogenization of tone of voice.

[00:46:32] And so suddenly you're going to have to get much better

[00:46:35] at your prompt engineering

[00:46:36] to deliberately rip away from that.

[00:46:38] But it's still going to tend to the average.

[00:46:41] It's still going to try and drag you down.

[00:46:42] So if you're average,

[00:46:44] you're probably in danger

[00:46:46] because you're going to get it.

[00:46:47] But if you're a poor performer,

[00:46:49] it's probably a good thing.

[00:46:50] If you're not very good at writing,

[00:46:51] if you're not very good at being creative,

[00:46:52] if you're not very good at

[00:46:54] looking at your ingredients in your fridge

[00:46:55] and working on what to cook,

[00:46:57] it can drag you up to average really, really well.

[00:47:00] So I think in a lot of cases,

[00:47:01] it's going to be amazing.

[00:47:02] But if you're a really creative

[00:47:04] or you're a really high performer

[00:47:05] or you're really amazing at something,

[00:47:07] all you're really going to do there

[00:47:09] is you can increase your volume of output

[00:47:10] because of the speed and efficiencies and everything.

[00:47:13] But it's going to drag that output quality down.

[00:47:15] And suddenly you're going to have to

[00:47:16] spend your time getting it back

[00:47:18] from the average back up to high again.

[00:47:20] I think where it gets more interesting for me

[00:47:22] is when you can control the data sets going in

[00:47:25] until you're not looking at large language models

[00:47:27] but small language models

[00:47:28] and saying, well, actually,

[00:47:29] I only want to train it on my work.

[00:47:32] So the average is going to be the average of me,

[00:47:34] not the average of everyone.

[00:47:35] Or maybe I'll pick someone a bit better than me

[00:47:37] to train it on.

[00:47:38] But when you can control that data input,

[00:47:40] suddenly you can make things better

[00:47:44] from that perspective.

[00:47:45] But I think we are in a dangerous situation

[00:47:48] where the bits we love doing as humans,

[00:47:49] we seem to want to automate that stuff

[00:47:52] even more aggressively than everything else.

[00:47:55] Like, if you think about all of the creative stuff,

[00:47:58] all of the imagery, all the visualization,

[00:47:59] all the writing,

[00:48:01] and it's all the stuff that should be

[00:48:04] the really rich stuff

[00:48:05] that we as humans have

[00:48:06] a really unique perspective on,

[00:48:08] a human touch.

[00:48:09] And that's the stuff

[00:48:11] that seems to be pushing ahead

[00:48:12] far faster than the boring RPA automation,

[00:48:17] AI agent,

[00:48:18] like the nuts and bolts

[00:48:19] that everyone goes,

[00:48:20] I hate that bit of my job.

[00:48:22] We seem to be getting left with that

[00:48:23] and then all the really creative nice bits,

[00:48:26] that bit's going,

[00:48:27] which is terrifying.

[00:48:28] I agree.

[00:48:29] And that's why I'm always critical

[00:48:30] when I see people using it

[00:48:32] where I think,

[00:48:34] well, if you could take anything off your desk,

[00:48:36] what made you do that?

[00:48:38] Maybe it's just what they didn't like doing

[00:48:40] and they just need to crank it out, right?

[00:48:42] If you're a new content creator

[00:48:45] and you know somebody tells you that,

[00:48:48] well, you got to be,

[00:48:48] if anything above all else,

[00:48:50] you got to be consistent, right?

[00:48:52] So you just force or train an AI

[00:48:55] to make sure and schedule things

[00:48:57] through automation or whatever

[00:48:58] and say, okay, every Monday, Wednesday, Friday,

[00:49:01] I'm pumping out a thing

[00:49:03] and hopefully I'll have a chance to review it

[00:49:05] before it goes out.

[00:49:06] But otherwise it's been trained

[00:49:07] on my tone of voice or whatever

[00:49:09] and the thing's on autopilot.

[00:49:11] And it's like, well, that's,

[00:49:11] we're talking about your personal brand.

[00:49:14] Like that's not important to you,

[00:49:17] God bless.

[00:49:18] But I mean, it should be

[00:49:20] because it seems like that's why

[00:49:22] you would get into the type of work

[00:49:24] that you're doing

[00:49:25] if that's what you're trying to do.

[00:49:27] But all the content that gets pumped out

[00:49:29] and then of course,

[00:49:30] you know, that gets ingested

[00:49:31] and used to train,

[00:49:33] you know, other systems going forward.

[00:49:35] And so it's just like

[00:49:36] this giant sort of recycling of content.

[00:49:39] But then it gets back

[00:49:40] to the prompt engineering

[00:49:42] and are people fully understanding

[00:49:44] what they're doing before they build it?

[00:49:45] So you might maybe think of it

[00:49:47] there was a Party Rock,

[00:49:49] which is one of the Amazon tools

[00:49:50] you can use to build that apps, etc.

[00:49:52] There's a Party Rock

[00:49:53] that somebody internally built

[00:49:54] and it's wonderful

[00:49:55] because the person that built it

[00:49:57] is this amazing French guy

[00:49:59] and he's got a certain degree

[00:50:01] of sass and flair

[00:50:03] and he built that into his Party Rock.

[00:50:05] So when you go and ask

[00:50:06] for information about this thing,

[00:50:09] it's as sassy as he is.

[00:50:10] It's amazing.

[00:50:12] It's amazing.

[00:50:13] So you get his character coming through

[00:50:15] and it feels very, very personal.

[00:50:16] And I think that's where I'm hoping

[00:50:20] we at least get to.

[00:50:21] You know, I don't think we're going to,

[00:50:23] it's like the rolling of the tide.

[00:50:24] I don't think we're going to stop

[00:50:25] this stuff coming.

[00:50:26] I don't think it's right

[00:50:26] to stop this stuff coming.

[00:50:28] But I think we can actively fight

[00:50:30] against pushing stuff out there

[00:50:32] that's the average.

[00:50:34] So if we are going to have,

[00:50:36] you know, these Party Rocks

[00:50:37] or whatever apps we're building,

[00:50:39] let's have them with a bit of passion

[00:50:40] and a bit of spice

[00:50:42] and a bit of unique creativity, etc.

[00:50:44] Let's not just let the system

[00:50:46] stand us and like cookie cutter us out

[00:50:49] into this weird,

[00:50:50] homogenous blob of stuff

[00:50:53] and just accept it.

[00:50:54] Let's have some creativity and push

[00:50:56] because I think there are

[00:50:57] some amazing things we can build.

[00:50:59] We just can't let it be the end editor

[00:51:03] as well as the deciding factor

[00:51:04] and everything else.

[00:51:05] So it sounds like you've been,

[00:51:06] well, logically,

[00:51:07] you've been playing with Party Rock.

[00:51:08] Any other tools or use cases

[00:51:10] that you find particularly

[00:51:11] interesting or scary?

[00:51:13] In terms of kind of productivity piece

[00:51:15] or building tooling,

[00:51:17] Party Rock genuinely has been great fun.

[00:51:19] And I was using it

[00:51:20] before Amazon had tied in.

[00:51:22] I think it's really great

[00:51:23] for someone like me that,

[00:51:25] you know,

[00:51:26] I'm not a big data engineering,

[00:51:28] BIE person.

[00:51:29] I have an idea,

[00:51:31] want to create something

[00:51:32] and it does a lot of that heavy lifting

[00:51:33] of you've got this prompt

[00:51:35] and it's pulling out

[00:51:35] all these different widgets

[00:51:36] and these outputs for you.

[00:51:37] So I genuinely quite enjoy them

[00:51:39] and I've built various

[00:51:40] different Party Rocks and stuff.

[00:51:41] I think the other thing

[00:51:42] that's really interesting,

[00:51:43] there's Note LLM by Google

[00:51:46] where you kind of put

[00:51:47] your information in

[00:51:48] and it can take all

[00:51:50] the different texts

[00:51:52] you've put in there

[00:51:52] because you can put

[00:51:53] multiple documents in

[00:51:54] and it will create notes from it

[00:51:55] and it will create

[00:51:56] different outputs, etc.

[00:51:57] And the one bit

[00:51:58] I particularly like

[00:51:59] is it does the audio output

[00:52:01] and where it kind of creates

[00:52:02] these 10 minute snippets

[00:52:04] of these two individuals

[00:52:05] having a debate

[00:52:06] around the topic

[00:52:07] you've put in there.

[00:52:08] And I genuinely think

[00:52:10] it's truly mind-blowing

[00:52:12] how fluid

[00:52:13] and how slick that feels

[00:52:15] considering the amount

[00:52:18] of information you're putting in

[00:52:19] and the buffer time

[00:52:19] and the input time.

[00:52:20] Like,

[00:52:21] the speed it churns it out

[00:52:22] is just phenomenal.

[00:52:24] Really phenomenal.

[00:52:24] I haven't done it yet myself

[00:52:26] but somebody sent me

[00:52:28] I have this group

[00:52:29] where like car guys

[00:52:31] and there was a car

[00:52:32] that I was interested in

[00:52:34] on one of these auction sites

[00:52:36] called Bring a Trailer

[00:52:36] and the guy basically

[00:52:38] downloaded the full chat

[00:52:39] and all the everything

[00:52:42] even the car,

[00:52:43] you know,

[00:52:44] history report

[00:52:44] and all that stuff

[00:52:45] and he threw it

[00:52:46] into Notebook LLM

[00:52:47] and he just basically

[00:52:48] sent me a 20 minute podcast

[00:52:49] about it

[00:52:49] and I was just like

[00:52:50] you've got to be kidding me.

[00:52:51] It's crazy.

[00:52:52] It's genuinely phenomenal.

[00:52:54] I think it's at a stage now

[00:52:55] once you've heard one

[00:52:56] you kind of see them

[00:52:57] because it's the same

[00:52:58] voices,

[00:52:59] the same characters,

[00:53:00] the same kind of cadence,

[00:53:01] et cetera.

[00:53:01] A bit like

[00:53:02] when everyone was using

[00:53:03] Chucky EPT

[00:53:04] when it first came out

[00:53:05] you could kind of

[00:53:05] spot the language

[00:53:06] and the cadence,

[00:53:07] et cetera

[00:53:08] and you can't prompt it

[00:53:10] quite yet

[00:53:11] and you can't change

[00:53:12] the tone of voice

[00:53:12] you can't change

[00:53:13] the pace,

[00:53:14] et cetera

[00:53:15] but as a raw

[00:53:17] kind of V1

[00:53:18] I guess

[00:53:19] it's just absurd.

[00:53:20] It's absurd

[00:53:21] how good it is.

[00:53:22] Yeah.

[00:53:22] Yeah.

[00:53:23] I'd say that's what

[00:53:23] I've been playing

[00:53:24] with the most recently.

[00:53:25] Yeah.

[00:53:25] Nice.

[00:53:27] So Toby,

[00:53:28] we could talk for hours

[00:53:30] as usual

[00:53:30] but I want to be

[00:53:32] respectful of your time

[00:53:33] and I know we're

[00:53:33] approaching almost

[00:53:35] the end of the hour here

[00:53:36] but I did want to ask you

[00:53:38] just as you know

[00:53:39] I've been spending

[00:53:40] a lot of time

[00:53:40] in not just

[00:53:41] a responsible AI space

[00:53:43] but just really

[00:53:44] helping people

[00:53:45] become more

[00:53:46] literate and ready

[00:53:47] for AI

[00:53:49] as it impacts

[00:53:50] their job

[00:53:51] their lives

[00:53:52] and so I'm just wondering

[00:53:54] if you have any

[00:53:55] bits of advice

[00:53:56] for anyone

[00:53:57] who's really just

[00:53:58] trying to get started.

[00:54:00] I'm certainly no expert

[00:54:01] but I'd say

[00:54:01] get out there

[00:54:02] and just play with stuff

[00:54:03] like break stuff

[00:54:04] play with it

[00:54:05] see what works

[00:54:06] see what doesn't work

[00:54:07] try good prompting

[00:54:08] and bad prompting

[00:54:09] see the difference

[00:54:10] of output

[00:54:10] you're going to get

[00:54:10] from that.

[00:54:11] I think we're in a stage

[00:54:13] at the moment

[00:54:13] and this isn't unique

[00:54:14] to Jenny and I

[00:54:15] but we're in a stage

[00:54:16] where we're going to

[00:54:17] have to continually

[00:54:18] be learning

[00:54:19] and I think that's

[00:54:20] just going to increase

[00:54:21] even more so

[00:54:23] always be pushing

[00:54:24] yourself

[00:54:25] and looking at

[00:54:25] what's new

[00:54:26] coming through

[00:54:27] what skills

[00:54:27] you need to develop

[00:54:28] or change

[00:54:28] and I think

[00:54:29] the most important

[00:54:30] thing though

[00:54:30] is really understand

[00:54:32] what's the value

[00:54:32] you create

[00:54:33] and bring to the table

[00:54:35] and what's the human

[00:54:36] element you bring

[00:54:37] to the table

[00:54:37] because if you can't

[00:54:38] articulate your value

[00:54:39] with the human element

[00:54:40] you're probably going

[00:54:41] to be in trouble

[00:54:42] it's probably going

[00:54:43] to be a difficult time

[00:54:44] because if all you're

[00:54:45] bringing to the table

[00:54:46] is something that can

[00:54:47] be automated

[00:54:48] or process driven

[00:54:48] it's probably going

[00:54:50] to be a pretty tough

[00:54:51] period

[00:54:51] so yeah

[00:54:53] I'd say

[00:54:53] go out there

[00:54:54] play with these tools

[00:54:55] have fun

[00:54:56] realise what works

[00:54:57] for you

[00:54:57] and what doesn't

[00:54:58] articulate your value

[00:54:59] and really understand

[00:55:00] your human element

[00:55:01] excellent

[00:55:02] well said

[00:55:02] Toby

[00:55:03] as always

[00:55:04] it's a pleasure

[00:55:06] and yeah

[00:55:07] thank you again

[00:55:08] for spending some time

[00:55:09] with me

[00:55:09] and I think there's

[00:55:10] a lot of

[00:55:10] interesting nuggets

[00:55:11] for my audience

[00:55:13] so really appreciate it

[00:55:14] pleasure as well

[00:55:15] and thank you so much

[00:55:16] for having me

[00:55:16] absolutely

[00:55:17] thanks everyone

[00:55:19] that's it

[00:55:19] for this episode

[00:55:20] thank you again

[00:55:22] to Toby Koshaw

[00:55:22] for joining me

[00:55:23] and we'll see you

[00:55:24] next time