Implementing AI Safely with Richard Brash

[00:00:00] Richard: What’s really interesting is when you go back and interrogate AI after they’ve given you the answer. Because if you say to them, “Is that actually true or are you guessing?” That’s the interesting answer. And we’ve used AI and got different answers to a very specific question, and then when I’ve gone back and said, “So this is what I think,” taking the o- other AI’s answer, and it’s gone, “Yes, you’re right.

I actually made up the numbers.” And I sort of said, “Well, why did you do that?” And then it answered, “It was partly face saving.”

[00:00:37] Ben: Richard Brash of Brash Solutions, welcome to the Berkhamsted & Beyond podcast. 

[00:00:41] Richard: Thank you very much. L- Lovely to be here. 

[00:00:42] Ben: You are a b- I think it’s fair to say, a Berkhamsted local. You’ve been around the town for 20 odd years? 

[00:00:50] Richard: 23. 

[00:00:51] Ben: 23. 

[00:00:52] Richard: So nearly a 

[00:00:53] Ben: local. And in… More of a local than me then, for sure. Um, and you’ve been in the same building as well, haven’t you? Business, business-wise. 

[00:01:01] Richard: Business-wise since, since moving out of London. 

[00:01:03] Ben: So- Yeah … on Northbridge Road. Mm-hmm. Um, tell us a little bit about the business, uh, and tell us how, how it’s changed since sort of 2001 to now. 

[00:01:13] Richard: Yeah, so, uh, 2001, formed my own business as a more of a consultancy, uh, for, uh, operational consultancy.

So using software systems to try and help clients automate, ’cause a lot of small businesses back in those days, uh, were very much paper-based, so paper processes, certainly through manufacturing. Uh, so I, I formed a business to really help to move from that paper base into computer systems. Uh, and that’s how it all started 25 years ago, and I was based in London.

Soon moved out of London, did the usual move, uh, and, and found Berkhamsted, uh, which, uh, through a connection with a local charity, uh, which is the Pepper Foundation- Yeah … which I’ve been involved with for many years, children’s charity. And, uh, yeah, because of s- a tenuous link to them, found a, a home with my wife in Berkhamsted, and now I live in Little Gaddesden.

But brought the business out to Northbridge Road. 

[00:02:16] Ben: And it’s a business in… I mean, automation is such a word. Are they s- It’s, it’s a buzzword at the moment, isn’t it? Yeah. Along with those two little letters. 

[00:02:24] Richard: They’ve caught up with us. 

[00:02:25] Ben: Yeah, yeah. So, a, which is what fascinates me about your business, because that automation that was originally, say, from paper to, to, um…

I was gonna say floppy disk, but paper to, to digital, uh, and is now obviously starting to revolve a- around AI. So talk to me about the different areas of the business, ’cause it’s not just that. It’s about being safe and secure as well, isn’t it? 

[00:02:48] Richard: Yeah. So, as I say, it started with a consultancy looking to help clients digitalize, as, as, as you phrase it.

But very soon, the way I worked with clients and, and where I was working with off-the-shelf software, I started customizing it to make it work for their business processes. And then I reached my limit, because my background is law and business law. So very quickly reached my limit and took on software developers.

So the software development side of the business grew, uh, as well as the consultancy. Uh, and then clients loved how we worked with them on, on that side of the business. Said, “Well, can’t you handle our IT support as well? You know all our tech. Can you not do that as well?” So- Yeah, good point. I’ll, uh, I’ll take on IT support staff and, and engineers so that then that side of the business, so now we’re, we’re roughly 50/50 between the software side and the IT support, which covers, like you say, security, um, just supporting laptops, servers, infrastructure, cloud infrastructure, and effectively now we’ve got eight staff in, in the UK.

Um, we have a development center in Budapest and Szeged, the two, uh, large cities in, in Hungary. And, uh, yeah, we can support… We almost say to client, “We’re your tech- tech department.” 

[00:04:15] Ben: Yeah. “

[00:04:15] Richard: Leave it all to us. We’ll handle it.” And because we have long-term relationships with clients- They just love that. ‘Cause they don’t want to go to someone, you know, “Well, this isn’t working.”

And we say, “Oh, no, we don’t deal with that. You need to speak to your person here.” And they say, “Well, then they need to talk to you about the hardware side.” So we just deal with it all. 

[00:04:36] Ben: I love the fact that that comes from, uh, a, a need from the client- Mm … as well, who’s basically said- 

[00:04:42] Richard: Yeah, 

[00:04:42] Ben: very- … “Well, why aren’t you doing this?” Yeah. 

[00:04:44] Richard: Very client driven. I don’t come up with very many good ideas. 

[00:04:47] Ben: I don’t believe that for a minute. I don’t believe that for a minute. Um, so m- marrying the two, the two sides, uh, uh, for a, for a typical small, uh, small to medium sized business, is that, they’re the kind of ones you deal with? Yep. It’s a whole range of.

[00:05:00] Richard: Absolutely. 

[00:05:01] Ben: Um, what are the typical problems that you’re seeing at the moment? Or, um, the, uh, pitfalls, I guess, or things, blind spots. 

[00:05:11] Richard: I think, I think your, you, you mentioned it earlier, the AI. Well, it’s a buzzword. There’s a lot of hype. And as usual, these, these things balloon into, or, a big bubble around the hype, and people jump in and expect it to work out the box.

Well, that’s not the way any new technology works. You have to work out how that’s going to fit into your business. You need to understand your business processes, and then fit in these tools. ‘Cause after all, they are just tools, and to make your business run smooth, more smoothly, or to increase the size of your business.

So it’s, it’s more about adapting the tools and helping our clients utilize the, the tools, and AI is, is a prime example. 

[00:05:56] Ben: So, go on then, let’s, but we’ll dig into AI. Yeah. ‘Cause it’s a, it’s a hot topic. What, what, what are the kind of, I was gonna say casual mistakes, ’cause I think it’s a, a good description.

I’ve got a couple of examples, actually. Uh, uh, the casual mistakes that businesses are making with where AI, A- AI is concerned. And, and let’s talk about it perhaps from a, a large lang- language model, so your Claudes and your GPTs to start with- Mm-hmm … before we delve into perhaps the agent side of things and start to probably blow my mind and some other people out there.

[00:06:31] Richard: Oh, I won’t talk technical. That’s one thing we as a, as a business, talking to our clients, plain English. That’s, that’s the main 

[00:06:37] Ben: thing. We’re, we’re the same with vi- we’re the, we are the same- Yeah … with video production. Yeah. There’s no, no, no use bamboozling with b- Yeah … with those kind of, uh, tech talk. So, so what are those casual mistakes, or, or the mistakes that, the, the easy mistakes to make?

[00:06:50] Richard: Okay. We’re finding a lot at the moment clients are letting loose their staff with AI. “Go and try it out. See what you can do.” Which as a business, especially when you’ve got governance, security, data security, obviously GDPR- Yeah … you know, letting your staff loose on these tools and, and having no policy, AI policy within the business, where’s your data going?

How’s it being processed? Is it being, is it being held in Europe, UK, America? So in ChatGPT, for example, where are their servers? Or Claude, you know, don’t have their data processing in Europe or the UK. Yeah. It’s all in America. So that’s fine if you’re happy with that and where your, where your data’s going to be processed.

Um, how your data is processed, is it going to be used in the large language model, or are you keeping it private? So if you’re uploading a contract, you need to know is that, is that gonna be used in the large language model and for anyone to effectively get information from that data. So you need to understand your settings, how, how it’s all set up.

So that’s one, one of the key things, the sort of governance around it. But then it’s your staff as well. If they’re going off doing all sorts, and all you’re doing is creating chaos, you may be creating some chaos and doing more, but i- doing more chaotically, that doesn’t really help things. And s- as usual, people…

You know, if you’re running a disorganized business, bringing in AI on top of that is not gonna help organize it. It’s actually probably gonna exacerbate the problem, ’cause it’s gonna be more chaotic and disorganized. ‘

[00:08:34] Ben: Cause I think there’s a, there is a general feeling, right, that people just think AI’s now gonna solve…

Th- there’s kind of two camps, I think. Well, there’s probably more than two camps, but there’s, there’s two ideas that I, I think of is there’s business owners that think, “Oh, AI’s here and it’s gonna solve my problems,” or, “AI’s here and it’s gonna take all my business away,” ’cause… A- and it’s, it’s how you approach it and adapt it, right?

[00:08:55] Richard: Yeah. I mean, we’re, we’re seeing our clients that run their businesses really well, know their business processes, their workflow, they, they’ve got their business organized properly. They are go- they are the ones making most use of it, or it’s, you know, it’s working for them. The chaotic businesses, they’re, they’re struggling.

So it’s, it’s again, you know, it’s the word crap in, crap out. 

[00:09:22] Ben: Yeah, yeah. 

[00:09:23] Richard: It’s gonna be- Yeah … the same with, with AI. 

[00:09:26] Ben: Um, so are there any c- catastrophic or just like, ugh, face plant kind of stories from clients that you’ve, you know, obviously without protecting the innocent, as it were? 

[00:09:40] Richard: I th- I think really not knowing the, the, you know, people running away with things and going off on their own and saying, you know, “I’ve created this agent” or…

But then trying to apply that back into the business and making it useful for everyone is where people are struggling. So often now we’re getting, you know, calls from clients saying, “Well, we’ve got this far with AI, now we’re stuck.” Okay, so you expect us to sort of just pick up the pieces. But if you’ve got some plan of how you’re going to implement it, we can come in and help with that.

We can help devise a plan and make it, make it work in a structured way. Um, so not sure if that answers your question. Catastrophic? 

[00:10:28] Ben: Yeah. I do. I was trying to be dramatic, you know, for a- Yeah, yeah … for a podcast promo short. Yes. In… We, yeah, which will go viral. 

[00:10:36] Richard: Yeah. 

[00:10:36] Ben: Um, but, no, it, it does. I’m just wondering, um, okay, so there’s gonna be people that are a bit, that are watching and listening the pod- listening to the podcast who are, “I know my large lang- I know what a large lang- language model is” 

[00:10:49] Richard: Yeah.

[00:10:49] Ben: I might even be able to say it. LLM. 

[00:10:50] Richard: Yeah. 

[00:10:51] Ben: Yeah. An LLM. That’s, yes, that’s why it’s got an acronym. Um, and there are those that are like, “Oh, I’ve heard of these agents, but is it not, is it kind of not the same thing?” I wonder if you can explain that a little bit. 

[00:11:03] Richard: I’ll, I’ll try my best. So the large language model is, is, is what you would in the old, old days called the database where all the data is held, and you’re interrogating it, and it is using your prompts.

It’s, it’s effectively looking at the words you’re putting in a prompt and trying to predict what you require as a, what your question is and, and giving you the answers. Now- What it uses a lang- large language model for is to get reference points. So it’s, so it’s saying, “This type of question, this is, this is what the answer,” or, “This is the type of answer we think you’ll, you’ll need.”

What’s really interesting is when you go back and interrogate AI after they’ve given you the answer. Because if you say to them, “Is that actually true or are you guessing?” That’s the interesting answer. 

[00:11:58] Ben: Because essentially, large language models are … They’re just predicting the next word, aren’t they? 

[00:12:05] Richard: Yes.

[00:12:05] Ben: It, 

[00:12:06] Richard: it, in, in the simplest terms And it’s an aggregation of data as well. So it’s … I mean, I’ve done it several times. Had a really interesting sort of argument, discussion with a friend over a few beers, and we’ve used AI and got different answers to a very specific question. 

[00:12:23] Ben: Mm. 

[00:12:23] Richard: And then when I’ve gone back and said, “So this is what I think,” taking the o- other AI’s answer, and it’s gone, “Yes, you’re right.

I, I, I actually made up the numbers.” And I sort of said, “Well, why did you do that?” And then it answered, “It was partly face-saving.” 

[00:12:43] Ben: It’s … There’s a couple of things, things there, because that … I mean, f- first off, that’s a prime hotspot for people mucking up in business, is just throwing, you know- Yeah … if the AI gets things wrong you must kind of check them.

Um, but when you talk about the prompting, going back and re-interrogating it, a lot of the time I, I was read- I’ve been reading some stuff recently about how you probably don’t even notice, but you’re influencing the, the model just by the way you ask the question. So from a sense of, “Oh, I’ve heard that- Yes

this is, this is true- It’s 

[00:13:21] Richard: confirmation 

[00:13:22] Ben: bias … can you find out?” Yeah, confirmation bias. 

[00:13:24] Richard: It is. And, and one of the things we’ve done is we use Claude and Copilot. In the settings you can … I, I have instructed Copilot across a whole organization in saying, “You are not here to agree with us. You are not being used to affirm what we’re saying.

We want a critical response of everything we’re putting in, and we want you to categorize your answer.” So now at the start of every single answer, each paragraph, I’ve asked it to say, “Is this fact? Is this your, what you think is likely, or is it just a guess?” So every paragraph now comes back in square brackets at the start so you know, “Okay, it’s unsure on this.”

And that’s a really, really useful tool to … And you can change that in the settings. 

[00:14:14] Ben: I’m appropriating that. Yeah. That’s, that’s, that’s a great one. I 

[00:14:17] Richard: can send you some stuff on 

[00:14:18] Ben: that. Thank you. Yeah, yeah, yeah. That’d be really cool. So, so on the … On that then, what are, what are some dead basics if you’re in Claude, GPT, Gemini- the Microsoft one that I never remember the name of ’cause I’m not in a Microsoft world.

Yeah. Yeah, Copilot. What are some dead basics that y- you should follow as far as setting them up or- 

[00:14:41] Richard: Yeah, as an organization, you said it there, the settings. We’re so used to, “Oh, oh, yeah, just agree.” 

[00:14:48] Ben: People- ‘Cause we’re just th- yeah, it’s like- It’s, it’s- … it’s the, it’s the Apple- Yeah … contract, isn’t it? Yeah, yeah, of course.

Don’t read. You can have my children for 

[00:14:54] Richard: the rest, yeah. But you have to go into those settings. It’s really, really important security, governance-wise, the setting up and the way it answers, as I’ve said, getting the instruction right across your organization. You can set your branding right from the outset, your tone of voice.

I mean, we’re non-jargon, so that’s essential in an- anything that is generated by AI, otherwise it will sound like it doesn’t come from us. We still check it all. We still edit it, but getting those settings right, that’s just crucial, and it’s, it’s not a huge amount. You can f- you can ask other AI on the AI you’re…

about the AI you’re using and what settings it would recommend- Yeah … if you’re unsure yourself. So yeah, there are tools to use, but cross-referencing is really key. 

[00:15:44] Ben: I found myself doing that a lot now, cross-referencing between the, between the different- Yeah … models. So we’re in a Google world, so we’ve got, we’ve got Gemini.

Gemini, yeah. It… Kind of Claude at the minute is my go-to, as it seems to be for, for a lot of people. But I am sat here thinking, “Hmm, maybe I should just go and revisit tho- those settings.” 

[00:16:02] Richard: Yeah. And, and you’re finding now that they’re, they’re, they’re overlapping. So Copilot is now using Claude or Anthropic in their, in their modeling, in their co-work, which is the agent side.

So- You know, you, you should be able to recognize what each one is good for, but you might be able to use it through one platform 

[00:16:25] Ben: Um, let’s move on to the agents, um, ’cause agents is something that I, I kind of understand but I, but I don’t. So let, let’s just say I don’t for the benefit of the pod- benefit of the podcast and the, and the listeners and the viewers.

When people talk about deploying agents, uh, kind of what are they and, and what s- scenarios could small businesses be using them in? 

[00:16:47] Richard: In simplest terms, it is automating repetitive tasks, and so you can set it as a schedule. You can create something to actually… W- when you’re not… You know, obviously everyone knows about the chat side of, um, AI, where you’ve got to prompt it.

What you’re doing is generating an agent that will run on its own at certain… So if I get an email from this or every, every morning. So I have an agent in, uh, Anthropic, and I, I have a prompt so that I can put what next, and it will look through Teams, it will look through calendar, tasks, and all my emails, and it will work out as a traffic light system what I really should be focusing on.

Absolutely critical, then this is urgent, and then don’t worry about it. So at any point during the day, if I think, “Okay, well, I’m not sure the priorities”, I should know probably, but sometimes it’s good to get someone else to tell you the, what, what you should be focusing on. So that’s, that’s a good agent that, that will do that, and then on the back of that you can automate the next step, the process.

[00:18:04] Ben: So to put that in tangible kind of terms, what ha- what does it look like on a, on the u- on the, the user interface of a, of a laptop and… Uh, do you see those sorts of things? Yeah, so, so- People can get 

[00:18:15] Richard: their head around it … okay, so you can have agent that w- will actually work in the background. So the only out- the only thing you’ll see is the output.

So, um, you know, an agent to, to, uh, if you get an email from this person, maybe it needs to be filed automatically in this place and a note go to Teams or, or something like that. Yeah. Or you’ve got a, a new starter in the company- The, within Teams, let’s… I know Microsoft, but, um, within Teams your employee can ask the agent in Teams anything about their new role or the policies.

Yeah. “Please remind me about my holiday or sickness. What’s the- Yeah … what’s the policy on sick?” And it will give that answer automatically. So it will inter- You, you’ll have data that it’s 

[00:19:07] Ben: trained. Please remind me about the company’s AI policy. Yeah. 

[00:19:09] Richard: So, so that’s a more- … an interaction one, but you can al- it’s the automated tasks.

[00:19:15] Ben: Um, so, uh, what, what are the sort of things you ro- are you rolling out for clients? What are you seeing them, them looking for and, and asking for as, as far as that, that kind of support? 

[00:19:25] Richard: Well, as, as far as that kind of support, we’ve got, um, an actual, uh, an AI officer, so people can’t really… You know, a small business, you can’t afford a, an executive role or senior role of someone just handling your AI.

So we, we provide different levels of a, a- an AI officer. So it could be as min- minimum as four hours a month just helping you keep on the right track, just understanding your business and, you know, finding out where, where you’ve got to, and just trying to keep, keep it all working correctly, maybe dealing with any problems, up to, you know, two or three days a month, um, actually fully embedded for those days and making sure that we’re driving it forward and you’re, you’re reaping the rewards of the investment in your AI.

[00:20:14] Ben: And what are you fi- what are you finding, uh, common across, across your clients’ areas where, where you can really help, areas where- There’s a second part to this question. I, un- answer that first, please. 

[00:20:27] Richard: So the first … Yeah, the m- the main areas initially are those repetitive tasks. Yeah. The, the more admin based, the lower level.

So that’s where, yeah, it’s the low-hanging fruit, isn’t it? Find the things that can easily be, um, adapted or, or created for an agent or, or automated, and then you’re looking at the more complex stuff that is actually, uh, a lot more, uh, looking at data, looking at databases, and actually drawing out, maybe automating reports, et cetera, and then prompting someone on the basis of KPIs to, to do the next task.

[00:21:08] Ben: Um, I knew I’d do this. I forget the second part of the question. Mm. Uh, but I, I think you, you were kind of then alluding to, so the more, what are the more complex things that I think maybe that, that business owners wouldn’t think about? So m- it might, it might not necessarily be complex, but the st- quick wins, I guess.

Yeah. Or like you say, the low-hanging fruit. 

[00:21:29] Richard: Yeah, exactly. It’s, it’s just the … Probably the thing to focus on is the things that you struggle with as a business. Okay. Because, you know, it’s just examining those and saying, “Well, why can’t we … ” You know, if it’s policies or, you know, onboarding of a, a new member of staff, well, why

You know, that’s all … It’s a process driven- um, operation. Yeah. So why, why is it that that can’t be more automated? So, you know, setting up of a, of a user, so, you know, giving them the orientation around the systems, w- is there any reason why that can’t be… You can’t use AI to do that? So it’s really the questioning, I think, that, that’s, uh, the more that y- you’ve ticked off the low-level stuff- Mm

and now you really interrogate, “Why, why are we doing it like this?” Or, “Why do we struggle with that area of the business?” 

[00:22:22] Ben: It’s come round, AI, the… I mean, the, the growth and the implementation of it is so… I think n- no one’s gonna disagree that it’s come so quickly. Mm. Particularly in the last 12- Yeah … 12, 18 months.

And obviously GPT was, uh, the, the biggest growth thing ever. Yeah. That’s a really rubbish description, but- … you, you know what I mean. It, it works. It, it was more… Yeah, it was the biggest growth thing ever. That’s all we need to know. It was, it… There were more users of it- Yeah … it was like faster than Facebook.

Yeah, the sign-up rate. Yeah. Fast- The sign… That was it, the sign-up rate. Within, within months- Right … it, it, it had done what Facebook and- Yeah … and Twitter and other stuff had done in, in years. So where do you see it g- I… So that is… In a way, this is a daft question, but in a way it’s where it’s not, ’cause where do you see it as an, as an expert, d- is that rate gonna slow down?

What’s gonna happen over the next few 

[00:23:16] Richard: years? So- Yeah, I think there’s a, there’s a initial everyone jumping into it and doing bits and pieces, and then it’s like any shiny new object, it fades a bit and people find, “Oh, well, I didn’t get it to work properly”- 

[00:23:31] Ben: Yeah … 

[00:23:32] Richard: go back to their old ways. That, that’s generally what happens.

And then those, as I say, who are more organized, have more of a plan, uh, are able to take it to the next level, which is actually making it use- useful in your business, measuring your return on investment, you know, measuring what is, what it’s doing to your business, how much is it improving. I think that’s where we’re probably at now.

That initial wave of just everyone jumping on it and creating useless stuff- 

[00:24:05] Ben: Yeah … 

[00:24:06] Richard: which, which has happened, but I think we’re over that now, which is a good thing, ’cause 

[00:24:11] Ben: it’s- ‘Cause- … 

[00:24:12] Richard: just let’s, let’s concentrate- Oh, we’re all- … on the useful stuff … 

[00:24:14] Ben: and we’re all fed up with the slop. 

[00:24:16] Richard: Yeah. Exactly. Especially in the, in the creative and media stuff.

[00:24:20] Ben: Yeah. 

[00:24:20] Richard: Some of it waste, waste of time. 

[00:24:23] Ben: It’s just a w- And a waste- It’s a waste of- Yeah … resources. Yeah. Exactly. It’s a waste of the, you know, environmental resources. Yeah. All that, all that kind of stuff. I… You are a very keen environmentalist. Yes. Yeah. So how do you, h- how do you weigh up, um, AI and, and the envi- and the cost to the environment?

[00:24:43] Richard: It’s a really good question. I think, um, AI is hugely competitive- Hugely, you know, commercial 

[00:24:52] Ben: Yeah 

[00:24:54] Richard: AI can’t afford to be inefficient. Any AI platform has to be the most efficient and inno- innovative. Part of that is energy efficiency and sustainability. Now, they’ve got an awful lot of money to spend and build in these data center, centers, and there’s huge, um, numbers out there about how much energy they’re all gonna need.

But the pressure is massive to, to get them efficient. You probably hear stories about these new chips coming that are just, you know, using, using photons rather than, you know- Yeah … electrons and, and I’d probably get that wrong, but- That’s… But, well- You, you, 

[00:25:38] Ben: it 

[00:25:38] Richard: sounds 

[00:25:38] Ben: good … I know what you mean. 

[00:25:39] Richard: Um, but- 

[00:25:39] Ben: Well, a, a really good example of that is in, in my world, in my industry, the cameras that we’re recording on here, the audio that’s being recorded, is all recorded onto the memory cards or- Yeah

solid state, state drives. Can’t get ahold of them for either love nor money or, or it, it… ‘Cause, because AI’s ta- it’s been- 

[00:25:58] Richard: Yeah … 

[00:25:58] Ben: the, the resource. Not because they’ve stopped making them or, or there, there, there’s- It’s demand … the man- they can’t find- 

[00:26:03] Richard: Yeah … 

[00:26:04] Ben: the, they can’t, they can’t find the raw materials.

[00:26:07] Richard: Yeah. And I think this is one of the big thing about the sustainability, but also on the, on the commercial side, you know, everyone has jumped on, and this is backtracking to your previous question, but people have jumped onto the bandwagon. 

[00:26:19] Ben: Mm. 

[00:26:20] Richard: Set certain things up without necessarily having a plan.

Everything is gonna be much more expensive to use because they’re going to limit it, and they’re getting people hooked, and then there’s going to be, you know, per processor tokens charging, uh, on a, on a lower level. 

[00:26:36] Ben: Yeah. 

[00:26:37] Richard: So a lot of companies are gonna be caught out by that, all of a sudden finding that their costs of using AI are increasing massively.

And again, it’s a commercial thing, and it, it will be a drive between competitive platforms to make it as, as cheap as possible, but they’ve got to recoup some of these hundreds of billions that they’re spending. 

[00:26:58] Ben: What are your thoughts on lower earth orbits? This is ’cause, yeah, this might sound weird to anybody listening who’s not heard about it, but, uh, lower earth orbit data centers.

[00:27:09] Richard: I’m… So there’s difference. Uh, on a, on a business commercial sense and a, and a techno- technical sense, it, it, it will work. It will. It will, it will make sense because of certainly the cooling and the- 

[00:27:23] Ben: Yeah, the idea behind it- Yeah … is the, the data centers need cooling. Yeah. Well, it’s cool up there. 

[00:27:27] Richard: Yeah. 

[00:27:27] Ben: Uh, and they need power, and, uh, well, there’s solar panels- Yeah

attached to these things. 

[00:27:32] Richard: Yes. So from that it will work. I’m not sure it’s… I, I… From an environmental point of view, I would like that not to happen. 

[00:27:42] Ben: Mm. 

[00:27:43] Richard: Certainly as an advocate of clear skies as well. 

[00:27:45] Ben: Yeah. 

[00:27:46] Richard: Um, I’m very keen on, uh, not destroying any more of our- Of our, our world, if our atmosphere as well.

Um, so I would prefer that not to happen, and I don’t think it needs to happen because I think there is innovative, renewable sources of energy with that sort of investment and finding, you know, battery storage as well. They, they need that. Yeah. But they’re finding that these data centers, they, they, they could actually help the grid.

They don’t need to be at full capacity all the time, so they can be used in the load balancing of a renewable energy grid. 

[00:28:23] Ben: Yeah. 

[00:28:23] Richard: So in actual fact, they may be a, a blessing. Might not look like it now, but I wouldn’t be surprised in five years or so that they’re actually looked upon in, in quite a positive way for helping load balance a renewable grid.

[00:28:38] Ben: ‘Cause as often happens with these things, it’s you, you … What, what is a problem then also- Yeah … becomes a solution to something else. Yeah. 

[00:28:45] Richard: And they’re so motivated to create a solution. 

[00:28:47] Ben: Yeah. 

[00:28:48] Richard: You know, they c- they can’t afford not to be. Yeah. So it’s not like a sort of governmental thing. These are private companies worth now trillions.

Um, they’re gonna roughly get it right. They will make mistakes along the way, but the commercial pressures will, will, will resolve it. That’s putting a lot of hope in them, and I understand that. As an environmentalist, that’s a, a dangerous thing to do. Um, and- Mmm, I, I’m confident that it, that it will be the case 

[00:29:19] Ben: Moving away from AI- 

[00:29:20] Richard: Mm

[00:29:21] Ben: and from the business to the other side of- … Richard Brash 

[00:29:26] Richard: The dark side. 

[00:29:26] Ben: The… well- 

[00:29:28] Richard: No … 

[00:29:28] Ben: I mean, Captain Crazy, actually. Um- 

[00:29:31] Richard: It’s bonkers … 

[00:29:32] Ben: the, yeah, Bonkers Brash. Uh, is it… have you got that website? If not, you should have it. Yeah. Um- I’ve got 

[00:29:38] Richard: the book. 

[00:29:39] Ben: Yeah, so talk to me about the book, and talk to me about cycling, running, and swimming around the, uh, circumference of the British Isles.

[00:29:48] Richard: Yeah, so the island of Britain, so that was 2015. I generally do, I do lots of silly events and, and, and adventure races, et cetera. But once every 10 years, just to mark a, a bigger birthday, I generally get out and do something that’s, uh, pushing the boundaries. So 20 years ago I did a, a solo, uh, Lands-, uh, John O’Groats to Land’s End, um, self-supported in, in a quick time.

Uh, 10 years ago, in 20- 2015, 11 years ago, I kay- I actually kayaked and- Ah, kayak … kayak, swam, ran, and cycled around the island of Britain, um, taking as, as much, you know, following the coast as, as much as humanly possible. Um, and completed that in 28 days, um, which is 2,800 miles- 

[00:30:46] Ben: That 

[00:30:46] Richard: is am- that- … around the beautiful island we live in.

That 

[00:30:49] Ben: boggles my mind that you did it that quick and got time to see it all. 

[00:30:55] Richard: Yeah. I just want to take it all in and see the whole- I think you’re, you’re, you’re immersed in it completely. But, yeah, I mean, there’s some a- amazing, amazing parts of the coastline that I was very privileged to see, that not many people get to see because it’s, it’s still remote.

There are, there are certainly parts- Yeah … of western Scotland where, you know, you won’t see someone for, for quite some time, especially if you’re clambering over moorland that no one has probably trodden on for hundreds of years. 

[00:31:24] Ben: Um, so many questions. So many questions on this one. It’s great. Uh, why is gonna be the first one.

[00:31:30] Richard: Um, I like to- I like to find my limit. I like… Or try and find my limit. I like to… I- I enjoy seeing what is possible. I, I do think we’re, as a h- as human beings, we, we generally live in our comfort zone. And I think that’s- there’s a good reason for that. There’s a really good reason, because it’s survival.

And, and from our sort of primitive days, y- your, your brain is wired to, to stay comfortable- 

[00:32:03] Ben: Yeah … 

[00:32:03] Richard: and, and to stay safe. Um, but I think we’ve got all these luxuries of modern civilized life, uh, that it’s actually does us good to, to have, um, some risk in our, in our lives, and to, to, to push ourselves outside our comfort zone.

It was very much a theme in my book when I returned and wrote it, was how the… My theory is that we’re, we’re born with a, a balloon of life. And when we’re young, we, we s- we experience new, new things. And every time we do, we fill that balloon. And s- so the, every puff of air is a new experience. And this balloon, by the time you’re sort of 18, getting on to sort of early 20s, your balloon is fully, uh, fully blown up.

And, and that’s fantastic. You’ve got all these experiences. But then you get these little pinpricks, holes, that are, that are doubts. You know, mortgage. I’ve gotta pay… You know, I’ve gotta look af- I’ve got children, a family. And they just start little tiny pinpricks of holes in those balloon- the balloon.

And, and it, and it deflates. And if you’re not careful, it be- can become one of those sad party balloons outside the end of s- someone’s driveway when two days after the party it looks all limp and shriveled. I was 

[00:33:24] Ben: really liking this analogy- Yeah … and now I’m feeling- Yeah … I’m feeling sad about it. 

[00:33:27] Richard: Yeah. So yeah, it’s great.

So, so, so in order, what you have to do, I feel, through, sort of later in life or, or whenever, is keep filling that balloon with new experiences to, to compensate for the little pinprick holes that are letting air out. So keep filling it with new, exciting stuff, and then that, that I think is a… You know, our, our lives are a privilege, what, what we have, and I think it’s a real shame to, to waste it.

So that’s my theory of my big balloon theory. 

[00:33:59] Ben: I like it. I love the theory. I lo- and it’s very well explained. Uh, I, I can’t help noting th- noticing though, ’cause we’re gonna talk about another one of your, uh, endurance adventures- Mm. Um, that you’re

doing these on your own and filling this, filling this balloon on your own. But I, I, is there a- Obvi- a part of that is challenging your- yourself so that- 

[00:34:20] Richard: I, I, I get the feeling you’re volunteering to join me on the next one- 

[00:34:23] Ben: No, def- … because- No, that is d- 

[00:34:25] Richard: The reason I do it on my own- ‘

[00:34:27] Ben: Cause no one else will go with you

is 

[00:34:28] Richard: when you say, “Would you like to come around the island of Britain? It will probably take a month, and- Oh … you’ll need to be self-propelled, cycle, swim across rivers and estuaries. Um, there might be a bit of a few ferries that you’ve gotta dodge as you’re swimming across, a few jellyfish, the seals.

There’s a lot of climbing. It’s probably about five times the height of Everest.” 

[00:34:50] Ben: Can I nominate someone? Yes. Well, we know, you know I’m gonna nominate- Yes, yes … Ben Morton. I, just this- Yeah … I know, I, this, the swimming thing might not be great for him. He’s been 

[00:34:57] Richard: working on it. Yeah. That’s right. But no, when I went round Britain, some friends, uh, Richard Hillier came along and joined me, uh, who you know well.

Yes. Uh, he came and joined me for a week, and he was sort of in the, in the van ’cause I had a, a support vehicle to carry the kayak ’cause I couldn’t carry it on my back. Yeah. Um, and he helped with some of the logistics- Brilliant … and he loved it. I mean, a- and a few other people joined me for a day here and a day there around the whole island, so it was, it was amazing.

But my latest one was on my own. 

[00:35:28] Ben: So tell us about the, the adventure. Yeah. 

[00:35:31] Richard: So, uh, last year I had another big birthday, so, so to mark, mark that, um, in my 50th year, I decided to take on the, the, the widely regarded as the toughest mountain bike race in the world, which, uh, starts in Banff, on, in Canada on the second, second Friday in June.

The clock starts ticking as soon as you cross the start line on your mountain bike- Yeah … and you have to carry everything, so you’re unsupported. You carry 

[00:36:05] Ben: your tent. It’s totally self-sufficient. 

[00:36:07] Richard: Yes. Uh, well, you do need to resupply for food. Yeah. You can stop at anything that’s commercially available- Yeah

so available to every rider. 239 starters when I- Okay … started. Um, that’s allowed, but if someone hands you something on route that is not handed to everyone, that’s not allowed. So you have to carry everything but resupply where you can. 

[00:36:33] Ben: Uh, this is called the Tour Divide. 

[00:36:35] Richard: The Tour Divide, yeah. 

[00:36:36] Ben: Uh, and it is, it’s, it’s well known within mountain biking circles- It’s, it’s a-

and with, within those ad- adventure… It’s the pinnacle, right? 

[00:36:43] Richard: It’s, it’s the ultimate endurance, mountain bike endurance or ultra-endurance race, yeah. 

[00:36:50] Ben: I had a little look at the website before we’re having this chat- Yeah … just to go through. 

[00:36:55] Richard: It’s about 20 years old. 

[00:36:57] Ben: Yeah. I was like- It’s- … “Oh, maybe, maybe you could- Yeah

help them out with that.” Well- But I was like, the amount of information on there to, to go through is like an adventure itself- Yeah, yeah … to start with. Um-

I, I, yeah, I… Just, just tell us about the- Yeah … experience of doing that. So 

[00:37:13] Richard: I haven’t said- ‘Cause- So we’d start in Banff. 

[00:37:15] Ben: Yeah. 

[00:37:15] Richard: Um, and it’s 96% of it is off-road, so it’s gravel tracks, single track, like mountain passes and, and you’re following the spine of the Rocky Mountains. So unfortunately there’s a lot of climbing and descending.

Um, and you finish at the US-Mexico border in what’s called Antelope Wells, which is just a border station, uh, by Trump’s wall now. 

[00:37:45] Ben: Nice. 

[00:37:46] Richard: So I’ve seen that. Um, and, and that’s in New Mexico. So, so you, you cycle through, um, Alberta, uh, tiny bit of Alberta, um, right down through, uh, Wyoming, Montana, um, Colorado, um, and then New Mexico.

So it’s 2,750 miles, and again, it’s 55,000 meters of climbing. 

[00:38:13] Ben: So again, like loads of questions. Uh, like the, the why we’ve, we’ve covered ’cause- Yeah … you’re looking for s- to find, find that, that limit. Yeah. So how, how close to your limit does a, a journey like that bring 

[00:38:26] Richard: you to? Yeah, so- What’s the- Um, so your, your…

Uh, I averaged s- actually in the saddle 13 and a half hours a day. 

[00:38:33] Ben: Whoa. 

[00:38:34] Richard: And I completed it in 23 days, 8 hours, and 3 minutes. Amazing. And I’ve got the T-shirt. Um, and you just find it, you, so you, you, it’s up to you. The, the, you have to follow the route. 

[00:38:47] Ben: Yeah. 

[00:38:47] Richard: If you deviate from the route, you’d be disq- well, you can deviate, but to go and get some resupply, and then you have to rejoin where you 

[00:38:55] Ben: left.

Join at the same point. 

[00:38:55] Richard: So, um, the clock just keeps ticking. So you can stop and go and find a motel and stay in a motel if you want, or you can pull your, your bivy bag out and sleep by the side of the, the trail in amongst the bears and the, and the rattlesnakes and, and the mountain lions and- … spiders or you know.

So it’s completely your choice. And, um- I’m guessing you chose bivy most of the time- 

[00:39:19] Ben: Most of the time … 

[00:39:20] Richard: most of the time. Yeah. I stopped in some nice, uh, well, relatively, uh, nice… Well, everything’s nice compared to- Compared to sleeping in a- But, uh- … bivy in the side 

[00:39:28] Ben: of 

[00:39:28] Richard: the road … I, I stopped five, five times in a, in, in undercover.

[00:39:34] Ben: Yeah. 

[00:39:34] Richard: But w- the weird thing is towards the end, I was just- Hang on a 

[00:39:37] Ben: sec. Five times in 20- 

[00:39:39] Richard: Three … 

[00:39:40] Ben: three. 

[00:39:40] Richard: Yeah. 

[00:39:41] Ben: Wow. 

[00:39:41] Richard: Yeah. But towards the end, I, I just wanted to be in the, in the te- in the natural environment. I took a very s- ultralight tent, and, uh, much… I just got the, I just enjoyed… ‘Cause you can just stop wherever and, uh, the freedom, it’s, it’s just incredible so long as you’ve got a bit of food and water.

But water is one of the big problems. There is… Uh, because it’s in June, I mean, you’re, you’re, you’re going up to- 12,000 feet 

[00:40:09] Ben: Yeah … 

[00:40:10] Richard: maybe 13,000 feet is the highest point. Um, and there’s snow up there. And then you’re down to about, um, three- 3,000 feet the lowest point and, and it’s 39, 40 degrees certainly in New Mexico.

So you’ve got this huge range. Uh, one day, one day it was 28 degrees as I was climbing the mountain, and then went in my tent, and the next morning open my tent it’s snowing. And I go to get… You know, and it’s about minus one. I’m knocking the ice off the, off the tent. So it’s, you know, it challenges you in many different ways.

[00:40:45] Ben: So what, what is the most challenging thing about that? Is it… Uh, ’cause it- Mm … talking to you about it, it doesn’t sound as if I was gonna say, “Oh, what, what was the lowest point, Richard?” And it, it doesn’t feel… I don’t feel when I talk to you about it that- 

[00:40:59] Richard: And the brain is very, very smart. Not mine particularly, but, you know, we’re wired to forget the pain.

Yeah. We can’t actually remember pain 

[00:41:08] Ben: Well, that comes down to childbirth- I don’t know … 

[00:41:10] Richard: and raising 

[00:41:10] Ben: kids, right? Yeah, exactly. If you, if you remember the pain- 

[00:41:12] Richard: Thank God we can’t … 

[00:41:13] Ben: pain 

[00:41:13] Richard: of- But even us blokes who are supposed to be not so good with pain- 

[00:41:17] Ben: Yeah … 

[00:41:17] Richard: you can’t actually replicate pain in your brain. So if you try and think, you can remember a painful moment, but you can’t replicate the pain if you break your arm.

You can’t rep- it’s impossible. So it’s very easy for me to now look back and say, “Well, it was, it was tough. I remember, you know, I couldn’t sit down on the saddle for three days, so I had to spend 13 hours a day standing- 

[00:41:40] Ben: Whoa … 

[00:41:40] Richard: cycling and trying to repair the damage, you know, at every, every point and, you know, off to a, to a pharmacy or whatever to try and get, get things- What you 

[00:41:50] Ben: needed

[00:41:51] Richard: things to repair, as you can imagine. But that’s part of- 

[00:41:54] Ben: Trying not to 

[00:41:58] Richard: reflect actually … yeah. I mean, but, but the attrition rates, only less than 40% finish it. I mean, one guy- Incredible … you, there are, there are downhill sections- For sure … when you’re on a, you’re carrying all your stuff, and people… I, I had four crashes.

Not… One was relatively big, but I mean, one guy got airlifted. He’d broken his back. I mean, it does, it’s a, it’s a proper race. 

[00:42:19] Ben: Um, and what a f- I mean, congratulations. What a, what a feat. Thank 

[00:42:23] Richard: you. 

[00:42:23] Ben: Um, I, I guess, unfortunately we’re, we’re, we’re nearly out of time, um, ’cause we could go- That’s a shame … on and on with this.

Yeah. We’ll get, we’ll get you back. I’m talking about the bears. We’ll get you back to talk about this, the bears as well. Yeah, the bear stories. Go on, let’s have a quick bear story before we finish. Oh, 

[00:42:36] Richard: it’s just cycling up a, up a, up a hill, and, uh, s- occasionally you’re, you, you meet another competitor. So I was generally sleeping more and cycling faster.

So most days I would catch competitors up, spend a couple of hours with them, cycling for a, for a bit, chatting, getting to know them. Brilliant. And, uh, and then maybe carry on and y- you’d maybe see them at some point later in the race. So there was this chap, Carl, who, uh, became a very good friend from Caribbean, um, complete nutter as well.

I think we all are. And, uh, we were just cycling together, chatting away and, and I’m, you know, fiddling with my handlebars or whatever, trying to, trying to get things to, to repair work. And he, he goes, “Hey, Richard, I think we got a problem.” And, and I, and I’m looking down at his bike. “What, what’s, Carl, what’s, what are you complaining…

What do you want me to fix now?” You know? And he, and he, “No, we got a problem.” And he, and I’m s- his gaze is up the, up the track, and, and I look up and there’s this massive black bear- 20 meters in front of us just looking at us as we’re cycling and, you know, of all the things to say, “I think we’ve got a problem,” rather than, “Bear!”

Yeah. 

[00:43:52] Ben: Or, “Ah!” 

[00:43:53] Richard: Yeah. So as we, uh, we cycled towards it, it, it decided to turn and, and run with its huge- That was good, wasn’t it? … bumous backside sort of… Aha. And then went off on the trail, but what an experience to see something that’s the size of a, of a Fiat 500 car just blocking your way. 

[00:44:10] Ben: Wow. 

[00:44:11] Richard: But black bears are, are a bit more shy, so they generally run off.

The grizzlies, one I cycled past, didn’t see it ’cause I was messing with my music, and the guy behind saw it and got his bear spray out. Oh, you have to carry bear spray, by the way. That’s one of the things that you must carry on the… Anything else is optional, but bear spray is the only compulsory, apart from a helmet.

[00:44:35] Ben: Amazing. 

[00:44:36] Richard: So that’s my bear sp- bear story. 

[00:44:38] Ben: Thanks for the bear story. Yeah. Last question for the podcast: How did the family feel about this, Rashid? How did- 

[00:44:44] Richard: All right. Long-suffering wife. She’s, uh, she’s amazing really. I think we’ve always… I, I’m, I like doing these things, and I think it’s, it’s always been the case.

On, uh, my wedding speech, I announced one particular, uh, adventure race that I was planning, so I think she’s used to it. But yeah, she’s, she’s, she’s brilliant. She doesn’t share the passion for, for the, the adrenaline stuff that I do. 

[00:45:10] Ben: So she’d rather you off doing it on your own anyway. Yeah. 

[00:45:12] Richard: And, and I think one of the things is, is inspiring the children to do the same, you know, go, go and, go and seek adventure.

I think it’s a great thing. We’re, we’re very lucky that we have that opportunity, so I don’t wanna miss it. 

[00:45:28] Ben: I think that’s a perfect piece of advice to end- … the podcast on. Uh, honestly, we could’ve gone on for ages and, and, uh, we’ll get you back. 

[00:45:37] Richard: Okay. 

[00:45:37] Ben: Thanks for joining us today. Brilliant. 

[00:45:38] Richard: That would be very nice.

Thank you. 

[00:45:40] Ben: Cheers. 

[00:45:40] Richard: Thanks a lot. 

[00:45:42] Ben: Thanks for watching this week’s episode of Burco and Beyond. If you’d like to be a guest on the show, then drop us an email, podcast@burcoandbeyond.com. And don’t forget to like and subscribe wherever you watch or listen to your podcasts. We’ll see you next week.