Keep What You Kill: Who Captures the Value of AI in Recruitment

If you actually secure an efficiency gain in your workflow, in your function, you actually get the time back and you can reinvest that as you see fit. Whether that's in improving the process, uh improving a different process, more automation elsewhere, um improving quality elsewhere, or just taking some time off, then you can do that. And I think that's a very important thing for us to recognize that the time saved doesn't automatically go back to us. We have to negotiate for that. Um, so welcome to another episode of the Harry Excellence podcast. We're in Budapest today at the Recruit Tech C event and um I'm taking some time out for a beer with my good friend Hung Lee. Hong, thanks for joining me. — Cheers to you JC. Thank you very much. Oh, so you're normally um on the other side of the mic interviewing me, everybody else from the world. Um I'm not sure everyone knows your background or knows who you are, how you became the man behind Brain Food. Would you mind maybe just start at the start? Talk to me about your career background, recruiting, how you ended up setting up Brain Food and what Brainfruit is. Yeah, fair enough, man. I mean, the origin story of Brimfield, there's actually two different origin stories. I don't even know which one's true anymore. So, um it may well be they're both true in a way. Um but I started off basically as a recruiter. So, I still think of myself as a recruiter. Uh Z 2000 to 2009. I was the agency side recruiting web one type people. So, eBay, PayPal, that early era, I was the guy recruiting web developers, what then called web producers and stuff like that. So very exciting times, similar to vibe to what it is now, you know, crazy new companies coming up. And I did that uh for agencies in London for about nine years. Uh popped out, did a bit of training type of work, early social talent type stuff I may add. Um terrible at that. So thank god someone better took that on. Um and ended up doing in-house type recruitment for these like uh 0 to 50 type growth startups. So recently funded 3 million quid, 5 million quid. I'd be the guy that converts that into — That's probably when I met you by 2010. I met you. I think 2011. — That's probably when we first connected. Yeah. Cuz I was blogging away then. I was ended up working for a few of these growth companies in London. I live in East London. So it was very natural for me to be in those places. And I did like, you know, two, three years worth of that type of recruitment. Me and a founder. founder needs to get to like core team setup, hire a replacement, I'm out and I'll do the next one. Um, so did a lot of that. Um, end up meeting a lot of engineers as a result, some of whom become friends in a weird way. Um, and uh, inevitably they come to you and say, "H, let's go build something. " Uh, and so we launched a tech platform in 2013. Um, which is a matching platform. So we were in the hired. com era, if you remember hired. Um we were a competitor of those businesses then. Um company ultimately failed. — Was Brave New Talent was it? — No, that was Workshape. — Oh, Workshape, sorry. — Um so Brave New Talent was the first in-house gig. Uh and that's where I met my founders for Workshape. — Um and Workshape basically was a talent matching platform based on uh what we thought uh the candidate sentiment was, what they wanted to do rather than what they had done. We'd visualize it and match it and then uh hopefully uh make it easier to for employers to connect with highly skilled in demand talent. Great product, terrible CEO. Um so I was always going to crash this car. But my attempts to save it started what I'm doing now which was uh you know doing content doing newsletters basically trying to stuff top of funnel um just by creating a bit more noise about the uh sort of uh about myself and breadcrumb it back to the product and in the end that became more important to the audience me and recruitment brain food was on its way. So, when was Brain Food started as a newsletter? When was the first time? What was episode one or edition one? — We can actually map that back because I haven't missed a week and we're up to issue 52, I believe. 53 coming this Sunday. So, it's more or less 10 years. Um, 9 and a half I think on October is 10 years. So, uh, so yeah, it's 201 — 16. Yeah, was when it started. So the idea was more analog than what others were doing. And arguably that more analog method ends up having more longevity than what everyone else was doing at the time. — What do you mean analog, Johnny? — You set up an email newsletter when everyone else was running away from email and saying it should be all

social, it's all about Facebook, it's all about YouTube pages and everything else. — Everyone else is gone and you're still alive and and healthier than ever, — right? Yeah. Yeah, I mean basically I thought there was space for long form reading, you know. I felt there was space we get loads of distractions. We're responding to notifications. Um but I think there's a there was a even back then I felt there was a thirst for a space that we wanted to create where we're just going to read one thing. Um and that's why I thought, okay, I'm going to just do a text based uh communication. Um I'm going to send it at a certain time of the week where I thought that space might exist. So, it's sent on Sunday morning, which is maybe an unusual time to send a newsletter. Um, but I thought, okay, when we getting the paper out and reading that over a coffee on a Sunday, that was the analog and I thought, you know, maybe we do that by email newsletter these days. — Can you remember any of the items that were the first ever episode? — Um, our first ever letter. Oh, I remember the first ever letter and it was a disaster um because I sent it I was so incompetent um at doing newsletters um that I sent it to grand total of one person and that person was me. So I had no idea how to do the mechanics of it. Sent it to myself. I thought right that was the first quit point. I got to be honest. I thought what am I doing? I spent all this time putting together this newsletter. pushed it out with a huge expectation uh and then lands on one person's inbox. It's actually my own inbox. I thought, right, I ain't doing that again. But then I remembered um a conversation I had with Matt Alder, our friend Matt Alder um and he of course longunning podcast in a recruitment game. Um and I remember him sort of talk me and him talking about sort of doing a content channel and he said, "Yeah, just do it for a year uh before you even look at ROI. " And I remembered that — and I thought, you know what, that's exactly what I'm going to do. I've made a commitment. I'm going to do this every week for one year, 52 issues, and only then am I going to look at the numbers and then work out whether it's worth it. Um, and that was probably the best and most decisive decision I made on Brain Food because if my window of analysis was 12 weeks or if it was even 3 months, I would have stopped it. explain Brain Food to a recruiter who's never come across it, doesn't know what it is, never read it. Um, tell me what it is today. — Yep. It is a curated newsletter where I'm picking out 10 things that I think are very interesting for the recruitment industry that may have come from outside of the recruitment industry. Uh, come from non- mainstream sources. There's stuff that you wouldn't find uh on your own. I've picked them out and thought and said, "You know what? This has been really interesting for me to think about. maybe this is something that can help stimulate you. Um, recruiters, I think, are very intelligent people, uh, but they're often locked in a repetitive type of function. And I'm not sure we always get the chance to exercise our mental faculties to the fullest extent. Um, and I wanted recruiting brain food to be as literally that to basically stimulate some neurological growth. You know, can we uh have some exciting things to think about? Uh, maybe even trigger you a little bit. just think about something a little bit differently uh and give you that mental stimulation for your week ahead. So that's the that's what recruiting bravery still is. — It's expanded though you've gone beyond just the I say just it's a wonderful newsletter you handcreated every week um you have podcast show um you speak a lot so tell me what it's expanded to today because people might be thinking I know this guy but I don't read the newsletter. Yeah, I mean basically I do a weekly webinar. I do a second newsletter on Monday which is more like my old sort of a free form essay original writing. I participate in a lot of events where I recruit the tech CE right now. Um I do that because I think it's all part of the job. Uh and the job ultimately is to look at the ecosystem of recruitment and all of the characters in that whether you're a recruiter, you're a hiring manager, you're a candidate, you're a tech provider, whatever you are. Um I think that ecosystem um is better when it's healthy, uh when it's high energy and where the information flows and it's high density. I think everyone benefits from that. And so whatever it is I'm doing is all geared towards that north star. I' I'd argue that over that 10 years, you've become to recruitment what maybe Lauren Michaels was to comedy in the US in terms of maybe no one's ever heard Lauren Michaels tell a joke, but every comedian you know, every movie has you you've uncovered them because of him. You're and I think you would deny this, but I'm going to say you're the most well-known and influential guy in recruitment person in recruitment in the world. No one else has that much influence. and the influence — different to most people who have massive egos, right? I have a big ego, other people with big egos. You do it

through the power of sharing what other people are saying that curation. That's an unbelievable skill. — Uh it's very kind of you to say and I would dispute some of the things you've said there. I'm sure you would. — It's very kind. Um but yeah, I mean I I learn from other people so much, man. You know, I learn from you. I learn from Clay. I learn from all these pe amazing people. Um, and I am under no illusions, you know, I don't pretend to be an expert. Uh, because the reality is I'm not. I'm a person that is just curious. Um, and the only thing that differentiates me from someone else, I think, is that I may know enough to know what's super interesting. Um, and yeah, I'm there to present it. I'm not trying to endorse a particular view. Um, say this framework is better than that framework or that product is better than this. The truth is, I don't know. Uh, but it may be, hey, this is interesting. It's worth paying attention to, folks. to check it out. Um, and that's generally uh how I kind of uh control my activities, you know. Uh, it seems to land well with the audience. It lands well with me and that it kind of enables me to continue doing it. — Tell me about how the sausage gets made. Like what goes into a week on brain food for you? Because folks might read it and go, "Oh, you know, my newsfeed gives me stories. " Yeah. So, what's all this about? Walk me through the process of getting to a Sunday morning. — Yeah. So probably the biggest revelation that most people um might not understand about the content collection part uh is that I don't search for anything. Um so I'm not online searching for anything. I'm not asking chatb what's interesting. Um the technique is to position yourself into where the information is flowing organically and then just picking stuff out. So uh again using sort of straining more metaph analogy should we say um it's like being a fisherman choosing the point in the river uh to fish. Um like the skill is choosing the point in the river. It's not necessarily casting the the line at the longest length or whatever. It's figuring out where the fish are um and then just being patient and then picking out okay that's interesting or that's a new take. Maybe that's a mistake but it's still an interesting way in which that mistake's been made that's worth talking about. Um and the education and the intelligence comes from the debate afterwards I find. So it's there to stimulate a conversation there to stimulate um more thinking and then the education comes from everyone else who's contributed to that. — So you've become like the New York Times or the the Times of London of the recruiting industry, right? I think whether you're a leader, whether you're an individual recruiter, folks know to go to you on a newsletter as a reliable and as you say, independent source, which is phenomenal. You've seen, I imagine, tons of trends over 10 years. Tell me about how the AI trend compares to some of the other trend trends you've seen. When did it kind of start for you? How has it evolved? And maybe that might bring us to your talk today that you delivered this morning. — Yeah. So I think actually um AI is a milestone on a it kills a lot of trends pre AI. So 2022 I think was the decisive era uh decisive year should we say where the past stopped meaning what it did and the new future began for us. Um it wasn't just AI that arrived in 2022. Lots of other things two other things I would say made a huge difference. um number one uh basically the end of the a zero interest rate period that we've been having for 15 years or so after the great financial crisis. So this goes back into a little bit of economic history but after the GFC basically what happened was the US Federal Reserve kept interest rates close to zero which meant there was tons of cash invested into companies typically tech companies in order to scale up. This is the tech unicorn era. Now why that's significant for recruiting is that a lot of the uh big conversations in recruiting was all about scaling tech teams. It was all about hey how do we go from zero to a thousand engineers in this period of time and maintain diversity maintain quality maintain culture. That was the dominant conversation for 15 years really in uh the tech space but because tech was so dominant it became the dominant discourse across recruitment. Wasn't there a repeat a peak moment hung when I think you put this on brain food when the amount of jobs advertised for recruiters exceeded the amount of software engineer jobs? — It was May 2021. — May 2021 was absolute peak recruiter. Um and this is just after this is the period where really post pandemic uh postlockdown um there was this feeling of okay the world's going to go even more digital than it was before. That means every process that is analog needs to be converted into digital. That means more tech hiring. That means more recruiters and we couldn't hire the tech recruiters. And suddenly uh us recruiters were getting ridiculous offers. Um you know 200k base. Uh take six weeks for the sabbatical every year whenever you want. Work remote no problem. Everything right? It was crazy

times. Lasted for about 6 months. Um so our golden era which should have been I thought a decade plus lasted 6 months. Uh why did it sort of crush down? Well, that's when the interest rates started being cranked up. I think it was bad analysis by the US Fed Reserve. And without getting too political, I think basically the Fed Reserve wanted to uh to time their interest rate rises and then do an interest rate cut in order to support President Biden's uh return uh effort against Trump. I think that's basically what happened backfired massively because it just raised costs for everyone and it pulled money out of the economy back into the banks which means suddenly we weren't on that growth spurt and the second thing that happened again apologies for bringing in geopolitics in this but of course the Russo Ukraine war kicked off in earnest in that same year and the the western response to that really was the a landmark to the end of a globalized system. Um now again you can no one supports the Russian invasion of Ukraine. Um, however, the the response of the Western world in weaponizing a lot of the the systems that were there, kicking Russia off Swift, uh, freezing assets, doing all of those things, basically caused everyone else to think, you know what, we can't trust these so-called neutral institutions because it basically uh we might uh lose our cash, we might lose our investment. We have to go a parallel route. So from that we have what I call hostile deg globalization and globalization the past 30 years have been the reason why we had this economic growth globally. So those factors basically reduced economic growth less economic growth is less hiring. Suddenly we look at recruitment differently and then out of nowhere AI lands on our plate. Suddenly CEOs are looking at that and saying hey maybe we don't need to recruit that many people. So 2022 is essentially the end of peak recruiter in terms of TA and it's opening up a totally new diff a new world. Um when you talk about how that compares the trends AI is part of that moment which ended that period from 2008 to 2022. — This morning on the stage you introduced three characters let's call them um to the audience. Um explain those characters and their respect of an AI. Yeah. Well, I read an article about sort of people who talk about AI. It was very interesting because, you know, everyone's confused and we have a situation where we have very intelligent people making very cogent arguments that seem to contradict each other. Um, and for the lay man on the street, lay person on the street like me, um, and probably a lot of people listening to this podcast, it's like, okay, how do we know which advice to take? Um, and I thought it was a useful exercise just to consider, um, to create some archetypes as to who talks about what. So AI optimist um this is typically a technologist um or typically a venture capitalist for technologists and they're talking about a future where yes artificial intelligence is going to be disruptive um but the medium to long-term outcome is a net gain on jobs uh because it's going to create massive economic growth because productivity will increase. Um so the disruptions we're feeling today painful but temporary and then we're going to have a big boost up and they're leaning on their previous reads on uh previous technology revolutions as evidence of this. Um then we have the AI pessimists which kind of don't disagree with the fundamental analysis that it's disruptive. They just don't think the disruption is worth the pain or the pain has been particularly well calculated into that assessment. In other words, what happens if millions of people are thrown out of work? Do we know what happens socially uh if that occurs? politically if that occurs? Um and how and they again point to historical examples where technology revolutions the same ones that the optimists talk about in positive terms. They say you know what industrialization produced communism it produced Nazism. It produced global conflict. Do you know how many people died as a result of all of that? Um so that's the pessimistic view. And then you have the AI skeptic view uh which is basically revolutions matter but this one it matters less than we think and it'll happen more slowly. So it's not going to be just like u a black and white shift overnight a flick of a switch but we'll just notice suddenly robots are doing everything for us or suddenly all of the information we're dealing with is actually mediated by an AI agent that we are uh control of information. Um, on that you spent the second half of the presentation not arguing for or against any one of those three modes. Again, as you say, you swim between them. You walked us through what can we do regardless of which view you

have. I thought there was some amazingly interesting observations and suggestions there. Um, can you pick out a handful that you feel are probably the most important to share? — Yeah. So I think it's an I think we've moved beyond the point where we we've got a choice about AI. Um like AI is being rolled into us whether we like it or not. Um and so that takes off the table one of the choices that was ex in existence 18 months or so ago where people could say yeah I'm not dealing with it. Can't not deal with it. You're dealing with it. Um and so um we have to um lean into it as a profession. Uh but we also need to understand whether it's good or bad for us immediately depends on whether we're running our own business or not. — Um there's no coincidence that the people who evangelize most about AI are typically people that have um ownership stake of the business, CEOs, uh the investors. If you're an entrepreneur, you love AI because suddenly your capability of operating a business is massively expanded. Um, however, your relationship with AI is different if you're an employee of that business. Uh, because what does that scope expansion mean for you? It means that you're busier than ever. It means that your expectations in terms of how your judgment performance have elevated. Uh, are you getting paid anymore as a result of that? Answer is no. Um, and so all of AI presents to you is kind of uh an attempt to to stay relevant, but it's a negative motivation. It's like, I'm using AI because if I don't, I might get fired. That is ultimately not a long-term motivation that's sustainable. We need to flip that such that you're using AI to get better at your job, to improve your conditions, but also so that you can grab hold of the value that you've created for the business. And this is what I mean by the keep what you kill policy. Uh which is not my invention by the way, not even my popularization. Ron Lev CEO Box kind of u uh articulated that. Um and he basically did that only because he recognized that AI adoption in his business was slower than he'd like and he figured out it's because the employees were frightened of AI or they were frightened of what it meant that AI was here. And so he flipped it on its head and said, "Look, if you actually secure an efficiency gain in your workflow, in your function, you actually get the time back and you can reinvest that as you see fit. Whether that's in improving the process, uh improving a different process, more automation elsewhere, um improving quality elsewhere, or just taking some time off, then you can do that. " And I think that's a very important thing for us to recognize that the time saved doesn't automatically go back to us. we have to negotiate for that and I think it's a great way where we secure something close to a win-win-win outcome um for everybody employer employee and at the end of the day the customer of the business that's being served they should get a better product better service maybe at cheaper price — you talked about how there's a risk with AI that the token costs today are being heavily subsidized by VCs um I read recently about a UK food company that a private equity firm bought into and this company was offering organic chickens I think as an example for £3 a chicken for a period of years and they were selling loads of them. The thing is it cost them six pound. Yeah. — To put that chicken on the shelf and eventually that business folded because you can't keep throwing money into it and in fact it wasn't a three pound chicken. It was a three pound chicken subsidized by a PE who spent £3 on the chicken for every time you bought it. — Correct. That's kind of what's happening with tokens today as we're learning that the tokens are being subsidized. They're not making a money not making money for the um the LLM. They're not the peas are throwing money into building data centers that aren't funding this, aren't getting a return from this, I should say. I guess the industry probably thinks it's the Amazon example. Everyone looks at Amazon going, "Don't worry, 20 years of losing money, you'll make so much money at the end, it's all worth it. " What's the risk of widescale LM adoption and how it relates to token cost that you called out this morning? Yeah, I think anybody who's been like at the advanced a edge of AI adoption probably has already personally and directly experienced this uh where you've ripped out a nonAI process. Uh you've created an AI native process that was dependent on this token budget that you had but the token budget either gets expend sort of burns through faster than you realize or it gets more expensive because you're not in control of how much it costs. Um and we've seen recently, I think the last couple of weeks, we've seen anthropic increase um sort of their uh the cost of their subscription packages at the same time as throttling usage elsewhere. Um, and I guess the main concern from a TA and HR perspective is if we're replacing payroll with all of this automated and

AI uh SAS solutions, are we actually rewiring our business uh such that we've lost control of our business because it suddenly is u a sort of anthropic deciding whether we can run that business or not by escalating their token costs uh to us. So there's a huge risk there. I think we're starting to be aware of it belatedly. Um, but it's a bigger story than I think this is going to be one of the dominant stories of the next six months at least, particularly when you've got competition with open-source uh models that are far cheaper and obviously most of these open- source models are coming out of China and that creates basically a geopolitical dimension to this which is very difficult for a lot of Western people to adopt. So you're saying it's maybe like the lock in model of the razor blade industry where you buy a razor below cost and you're locked into using their blades which where they make the margin whereas you get sucked in by replacing your process with an LLM and you're locked into the token cost that you can't unbundle and you've moved away from humans or an alternative and you can't get out of that for years or forever. — Yeah, it's no different really from I say it's not different. It is similar but there's analogies to when the internet was first rolled out. Um, I think you and I are both old enough to remember where you needed a CD to get on the internet and you used to pay like AOL2 pound9 for this CD and it was like totally worth it because you know this is uh what it is and then basically what happened with that business model was that they realized that actually it's better to just give it away free to get access and then build the value further down the chain. So all of the businesses, none of the telecoms providers reaped the benefit from the internet in the same way as the other companies that built on top of it. So the big winners of that internet period are the Google's, the Facebooks, and the Amazon of this world, all that e-commerce stack. They're the ones that harvested the value. Um, now where we're at with AI right now is that we're still kind of uh the the people who are distributing the models are basically like your old telecom providers. um that they're the ones that are trying to make the money by metering our access to intelligence. Um and that's unfortunately now developed a lot of geopolitical kind of uh narrative to say you have to do it this way. But another way of looking at it is you know what um open-source uh intelligence is going to distribute capability across the entire planet and maybe that's a similar to distributing internet across the entire planet and the value will just move later on down the stack. uh the applications that are built on top of that technology. I think that's a better world than having the intelligence metered at the gate. — I guess using to expand the analogy, right? We used to have amazingly expensive long-distance telephone charges and then someone used VOIPE and VOIPE killed all that because it was like, you know, it's good enough, pretty close, and we can make it near free over the internet and because that channel was cheaper. I've talked to recruiting leaders who give me the examples of you know these AI screening or interviewing tools that are becoming more widely available average around or started around4 uh per interview and they'll say you know what we can hire recruiters in Southeast Asia or in South America for less than that per hour — and there's no EU AI act or California act to worry about there's no worry about the AI going rogue it's an employee who's very skilled very competent and we're creating you ethically jobs for people doing this way rather than putting it into, you know, um a rich billionaire's pocket. Um that is a real thing, right? In terms of the AI may be able to do things that a human can do or even do them better, but at the end of the day, in the same way that robotics hadn't hasn't yet fully taken off in all sorts of manufacturing is that as long as labor is cheaper and robotics, labor will win. Are we saying that's probably going to be the case here? as long as the subsidy is removed. — It's an emerging outcome um that again I think a lot of people will find unwelcome um because it accelerates this offshoring of jobs that probably most countries that uh are kind of exporting those jobs don't want. Um if you for instance had a bunch of software engineers in London and you decided that you need less software engineers because AI is going to do all of this. You fire half the team then you find out that the AI costs are far more than actually the human beings that you'd fired. You're going to rehire those software engineers, but will you rehire them in the in London in the most expensive place? Are you gonna say, "Hey, I know these people in Cape Town or Budapest. They're just as good, if not better, and you pay them far less. " So, there may be a global redistribution of essentially white collar work and kind of offshoring 2. 0 um coming back as a result of AI being um mistakenly accounted for. Uh if the subsidy argument is correct, — well, I it's 3. 0 go post pandemic a lot of roles that we thought couldn't be offshored for example audit — auditors uh had to come on site and then we couldn't go on site for two years and

then the accounting firms realized we don't need to go back we did just fine and our clients now got used to the auditors being remote and analyzing our data remotely so therefore let's not put them on site they're not on site let's not hire them in Europe or the US so we did that already with a lot of jobs in 2022 when we rehired them somewhere else now we are again to your point at risk of a third wave of that where we realize the AI isn't quite as good or isn't cheap enough or the price goes up to a point where we do it. Do you think there's a risk because we've talked about a lot about me and you and others about um recruiting teams are trying to build their own tools the most which is phenomenal right — I love the innovation there but a recruiter building a tool is going to use the highest end model available and when it comes to cost an application provider is going to build it in a way that's smart using the right model at the right price for the right application in a process you know do you think that there's a moat to be created by companies who are clever about the costing of AI and the proper use of the right model for the right job that genuinely can deliver AI but at the right price point that justifies itself. — Yeah, I think so. I mean, I think the vibe coding movement I think is something I'm massively endorsed. It's very exciting. It's a democratization. I don't like to use the term, but it's a dis dissemination of capability um outside of the tech industry and into the hands of others. I think that's broadly positive. Um, however, there's complications associated with that, particularly when you're working for a company. Uh, different yourself. You know, you're putting your own business at risk. stuff at risk. But when you're sort of using vibe coded tools in a corporate context, you're actually accessing processing data that you personally don't own. Um, and it's owning the custodian for that information is actually the business. Uh, and there's a big question mark there is, okay, how do you actually police that? How do you make that compliant? Um, no different from bringing your own tech into a bank. Typically, you can't do that. You can't just, you know, bring your own computer in and start logging into banking systems even though it's more convenient and efficient for you. Why? Because there's all kind of data leakage issues, all kind of IP cyber security issues. I think vibe coding uh kind of opens up a whole new uh sort of surface area for those risks. So, I'm pro- vibe coding even though I'm ultimately uh I don't see it having a huge positive impact in the corporate setting because I think you're going to just breach too many compliance rules and the companies are going to end up being becoming quite totalitarian uh with regards to their employees use of it. you're going to start seeing contracts uh written in terms saying look if you're using AI into the business that you that you've not been given directional license to use, you're going to be personally liable for all of this stuff. So those clauses will start being inserted into employee contracts and basically I think that will knock the vibe coding in corporate out. Um I think solarreneur that's going to fly in that space uh because the risk uh and the risk appetite is different — as a solopreneur. home. Talk to me about what impact AI has had on your job, your role, and do you see opportunity to further use AI in what you do? — 100%. Um, I put my hands up uh very early to say I'm a lagged when it comes to AI adoption myself. Um, and that's not because I don't think I'm technical enough or I don't have the interest. This is purely because of a bandwidth issue. Um, and the lower lowest hanging fruit for AI is still generative AI. So, in other words, generating content, generating images or whatever it is. And that is actually the part of AI which I'm most reluctant to use for what I do. — Um, I remember having a conversation with a friend of mine in Budapest actually. Um, and he said, "H, listen. Um, give me a weekend. I'll be able to come up with something that's going to really help you out. It's going to do this, and this, this, and this. " And in the end he was absolutely right. Um but he all of his recommendations was talking about the production of the content. Um and I can't allow for that to happen because it has to be my authentic and unadulterated voice. Uh so I have to type this stuff out. Um I have to read the content in order to commentate on it. Um I have to consume in order to be able to commentate on it. So there's a chunk of work which I can't automate. Now, what I can automate is all of the back office stuff. Um, I'm terrible at the CRM aspect of dealing with subscribers. I could be much better at that. Um, and maybe I could use AI to try and like create uh different versions of the content in different format. Absolutely. Um, could I do better on the business side, the sponsor side of AI? Absolutely. So, lots of things that could improve, but right now its impact for me is not particularly high. What's next for Brain

Food as we look into the next decade hopefully or even more? Um, what can we expect from Brain Food? — I don't know, man. I mean, people rightly kind of expect a strategy from me or some sort of foresight. The truth is I don't have any foresight or strategy. Um and uh the policy of brain food generally has just been to lean in uh to where the information or the stuff seems to be happening and then just work it out from there. Um it's not like saying here here's the destination. I want to shoot for that destination. That's where it's going to be. It's more like there's a big forest in front of me. There's no path through the forest. I'm just going to try and feel my way through. Uh do you mean which is the best route? I don't know. I'm on the journey of exploration. So, the journey of exploration with brain food is going to continue. Hopefully, that's going to be interesting enough to bring everyone along with me. Um, but yeah, there's no like, oh, here's the destination, I'm done. Um, I don't see that as an outcome for us. — As our beer nearly ends, Hong, uh, final question for you because any anyone listening or watching who follows you on Instagram knows you travel the world a lot. You get to see different cultures, different parts of the world. Tell me about your favorite cities to go to, favorite cultures you've been to, maybe top food tips. — Yeah, absolutely. I'd hate to pick out some amazing things cuz every city and every culture has interesting stories to tell. Um, and even if the the experience is challenging, sometimes they're the richest experiences. You know, you can't just get the free sugar hit and think that's amazing. Sometimes you need a bit of bitterness to actually get what's going on. Um, so I'll give you a few food tips though straight away. Um, unusual ones. There's a bakery in Japan. Um, in a town called Fukuoka. Um, it's called Pain Stock or I think it's P stock. So, it's French, but they have the most incredible um, selection of super interesting flavored breads that you'd never get. One thing the Asians get right is being very creative with the — So, give an example of a creative flavorful bread. So they'll say things like salted egg yolk with caramel in this bri something like that. Um or it'll be things something called pork floss on a quason. What is that? I don't know. Um but it's this mix of salt and sweet that you typically don't get in Western Europe. Um — Japanese bakery done. What else? — Absolutely amazing. Uh, I went to um I went to Busousan in Korea um in amazing uh fish market there. I forget the name. I think it's called um is it Chamal G or something like this? It's a huge fish market all school literally have all kinds of denisens of the deep being sold in buckets and tanks everywhere. And I had some raw crab in this um in this market. sushi kind of thing. — It's just raw but marinated in a sauce. Um and yeah, you shouldn't cook crab, man. Wow. — Should eat it raw. It's the business. Sorry. It's the business. — Um and the final food thing, um I'm going to suggest um an ice cream. An ice cream in Shanghai by a brand which is unhealthily called Gelato. So, it's Chinese brand called Gelato. You try and search for gelato in on online. Obviously, it's like you're esteemed by information, but if you're ever in China, look for a brand called Gelato. And their signature thing is they're going to stack a cone like 6 in high, like massive cone, but they've got some wonderful flavors. And the flavor that I just literally fell in love with was rice flavored ice cream. — Rice cream. How stereotypical. Rice cream. — Rice, mate. I looked at that and I thought, man, are they taking the mick? Like, are they just literally playing up to a stair? I've got to buy this. But I bought it and it's like rice pudding. — Yeah. — Nice. — So, soft bits of rice grain in this uh towering uh cone. Best ice cream ever. Then the best. I will take those recommendations. Hong, it's been a pleasure. Love you to hear as a motor bike goes past. Um lovely to hear your story what you're doing. coming for the next 10 years and congrats on the first 10 years of Brain Food. — Brilliant. — Oh no. Cheers on that. — Cheers to you. — Great chatting you, man. — And thank you for listening and joining us for this special edition of Budapest. We'll be back next week with another great leader in talent. Until then, take care.