100% Help, 0% Hype - A Solopreneur's Honest AI Reality Check

Uh, I think unemployment is up like 83% or something in the tech industry and the main reason for that that was given was AI. >> And why solopreneurs? What do they interest you? >> You can't just optimize for SEO anymore. You have to have a markdown version of the website so that it can be read by LLM so that your website can be surfaced in an LLM chat. Cuz you know, in the old days, there were so few touch points in advertising to reach a consumer. Now, it used to be something like 18 in the '90s. Now, it's like 38, 40. It's 40 times someone has to see your brand to then go, "Oh, that's that company that does AI transformation. " >> AI, is it hype or help? Good day and welcome to the show.

I'm your host, Pete Cooper, and our topic is AI, hype or help. Today's guest is Alice Bull. She's coming to us from the South Coast of the United Kingdom with her company, Tincher. She's helping small companies and solopreneurs adopt and use AI very effectively. This is going to be great, so let's get into it. >> >> Welcome, Alice. How you doing >> Thank you. Thank you for having me. I'm very hot. So, we're in the middle of a heatwave, um, but very good otherwise. Thank you. >> Yeah, so where are you coming to us from? >> I'm the South Coast of England. >> Mhm. >> Um, so yeah, hottest place on Earth right now, I believe. >> Oh, really? >> Yes, very. >> What is it today?

>> I think it's like 34 or something today. It got up to 41 a couple of weeks ago, which was, yeah, it was a lot. It was a lot. >> Yeah, for American audience, we're talking about into the '95s and into the '100s, aren't we, for that? >> Yeah, every time I speak to my American friends, I have to do the translation for them. But yeah, it's yeah, you're used to it. We're not We don't have air conditioning. >> Oh, really? >> No, no, not in the UK. No. >> Oh, good. >> We go to We would go to the supermarket to get relief. >> Haha. >> That's crazy. I mean, I'm in Pennsylvania, um which I think is even further north than you guys. Um No, actually, I think it's the other way around.

I think you guys are further north. Anyway, we have insta- I installed small air conditioners in the windows each summer, take them out again each winter. Anyhow, we're going a little bit off topic. Tell us a little bit about what you do. Do you um you know, what's your profession? >> So, I work with small mostly small businesses and one-man bands, solopreneurs to uh streamline operations and go-to-market processes with AI. >> Nice. Nice. And how long have you been doing that for? >> I was actually I was thinking about this the other day. So, I've been in marketing for uh marketing and operations for about two decades, 19 20 years.

Um and I started working with AI to introduce it into It was content creation really was one of the places that I saw uh content creation and finance, hehe. >> >> Weirdly. Um since about 2022. So, the AI aspect of it I've been doing for 4 years, um but it's it's obviously really really ramped up, I'd say the last 18 months in particular. >> Yeah. Well, interesting. 2022 was ChatGPT's sort of a launch where it actually became effective, right? >> Yeah. Yeah. Yeah, I think so. Yeah, and um yeah, we really started to see that kind of movement and it it it come out of the tech world into like the lexicon of of people who maybe wouldn't have otherwise heard of it.

>> Yeah, and as you say, in the last 18 months. So, we're 2026 now, mid mid 2026. So, it started 2025, we really started to see, as you say, I think an explosion. I don't know what you saw, but I think the general populace would say 2025, ChatGPT, 2026, everything if it's just an explosion of of tools and capability. >> Yeah, although you you say that and I think it's so interesting and I'm sure you you see the same when you work in this world and you're so involved in it every day and you're reading but like I don't know about you but I have my Claude send me the news every morning and and it's all about like the tech world and startups and how it relates to go to market.

But if you're not in it, I think there's still kind of a very very large part of of society who don't really might dip in and out and use it as they would Google. Um but they they don't really see it as part of their life and I think that's going to be an interesting thing to to look at the next few years is how it really does creep its way into everyone rather than just the early adopters. >> Yeah, I think what that I think you you definitely are an early adopter. Um 2022, I mean our my history goes back to machine vision stuff 2016 but we didn't obviously we didn't get into LLMs until much much more recently. Um and that's while you're talking about how it penetrates society.

I think that's interesting. What I've noticed is just this a huge amount of interest and attention to these big big companies, right? And I would say that's sort of a top-down thing, right? That they're so big and so much revenue, so much money's going there that it's really shaping the economy especially here in the US. >> Yeah. >> Um so who follows that? Well, I think you get the retail what I call retail which is like ChatGPT using it on a daily basis to like a browser or advanced browser and then you have the big companies going "Geez, we've got to jump in on this. " But there's huge I guess untapped potential. And you're talking about solopreneurs, aren't you?

And I think that's a huge potential. So why solopreneurs? What do they interest you? >> [snorts] >> Oh, that's a very good question. I think I've always worked with quite small businesses and I really, really enjoy working with small businesses because you can see yourself making a difference a lot quicker than if you're working with an enterprise business. So, um I I love doing that.

And and I think especially and as you were saying, you know, with big businesses coming into it, I've got some friends, for example, I've got a friend who um works for a huge financial services company in America and they're they're starting to ramp up their AI adoption and it's really interesting to hear how slow slow, slow, slow the whole thing is and there are so many layers layers of high hierarchy in those kinds of businesses that it is really difficult to affect change quickly. >> Mhm. >> And you know, it's almost like in a company like hers, and I think they are doing it the right way because they are such a huge business that they can't afford to get it wrong.

Um but it is such an interesting slow process because by the time they do one round of like, "Okay, we're going to try this. " It's already moved on to the next thing. Um so, I really, really enjoy working with uh founders and small businesses because they're a lot more open to trying interesting things, to getting it wrong and trying again, and and it's very, very quick to see uh introducing AI, introducing a process, it's very quick to see it take effect. It's it's instant. Um so, that's what I enjoy the most about it. >> Yeah. And so, how did you get it I mean, we didn't cover this all background, your training seemed to be marketing, but you don't sound like you were a coder.

Maybe maybe I'm wrong. You're someone that's come from the software side of things as well or mostly the marketing content generation. How did you sort of find your way to AI? >> I actually I taught myself how to code uh from a library book when I was 14. >> Oh my goodness. Well, there you go. I got that completely wrong, didn't I? >> >> Um but I but no, I've never worked in it. So, I I As soon as we got the I think I was 13 or 14 when when got the internet set at home And I was just immediately so enamored by how the internet connected you to different things and people like you could be connected to someone in America or like in Australia. And it it just that was so magic to me.

I grew up in the middle of the British English countryside. There wasn't a lot going on. It was just like this magical connection to a world that I'd never heard of. And that's what first sparked my interest and I just got something really got something when I was younger from Bill. I used to build websites and I started my first web design business when I was 15 and used to make these I mean obviously awful in today's sense, but you know, like really I worked so hard on these websites. And >> >> But I think being a a girl a woman is it you're not really sold software engineering as a possible career path. It was like you could be a journalist or a solicitor. I was neither.

I went into marketing. But it just I never even considered being a developer. I think if I had my time again, I wish I was. My husband is a developer. He's a an AI engineer. So I get it that way. But at least I managed to weasel my way back into that world of of developing and building. >> Yeah. So when you do work for other companies and helping them with AI, do you need to lean on your coding or do you now lean on to the tools and you able to get around the coding or how does it integrate with your work? >> Well, this is interesting cuz I think we're still at a point with AI where there are so many companies asking for it.

But generally they just want like a make automation or they want a Zapier thing built and they want these these trusted tools and they know that you can use A to connect to B and get results C. But there's this kind of accepted way that they feel safe doing it. I don't really like to work that way. I like to custom code stuff where I can because it's a lot more adaptable. You can do much more cool stuff if you are you know, making it from scratch yourself. And then you don't pay for it. Zapier, I mean, I I don't know if you use Zapier, but I've had a couple of occasions where if you get it slightly wrong, you can just get these crazy looped apps.

I think my um my record was about $1, 800 in uh looped apps over the course of 24 >> goodness. >> before it surfaced. They were very good about it. They refunded the money. Um but it's Yeah, so I I much prefer like custom coding stuff in Python, which I think is just It's just quicker. It's cheaper. It's easier. And I do lean more on my coding back Although I I don't know how to do it as as well as a developer would, but you don't need to know that anymore with um with Claude code or with Codex. >> Mhm. So, rather than just Yeah, that's great. So, rather than using a Make and Zapier, which I think it's graphical, you know, uh >> Yeah.

>> linking it Zapier linking that up, you actually go in there and write it, but when we say write it, we're we're really now using just Claude to write it, aren't we? I mean, is that what you're I guess you're you're looking through the code and just seeing that it makes sense. I guess Do you go in there and actually tweak the code or do you just completely rely on Claude now? >> No, I I'm completely relying on it. And to be honest, I'm not alone. I think Uber now writes something like 83% of their code. Um I think it was the last model of GPT. I think it was 5. 5 was actually entirely written by the previous model.

Um I know my my husband in his his job, they are they they do everything with Claude code. And it's it's interesting I I've noticed it as a marketer. I've noticed how, you know, how my brain is even started to work differently. And it's now less about like deep work and more about contents the context shifting. Um but yeah, I don't really know uh anyone who works in this industry who actually writes stuff by hand. You You just You're just a glorified babysitter. >> Mhm. Yeah. Yeah, let's go back to this loop zap thing that you mentioned. That's one of the problems that I guess people can get into. Do you know how that came about? I mean, what what went wrong in that start in that case?

>> I can't quite remember. Um, I was very embarrassed about it, but it was just like this infinite loop of data that I'd accidentally set up. So, it was kind of sending >> Oh, okay. >> thing and then re-triggering it. Um, I By the time I noticed it was too late. >> Really? I mean, what what time frame are you talking about? >> It was 24 hours, but still $1, 800 in 24 hours. It was uh yeah, it was very annoying. >> Yeah. Isn't there a mechanism to sort of set a budget or something so that you get flagged up or capped at or something? >> Uh, there is actually.

I think this is why they refunded me because I did I think sometimes the limit that you set, if it shoots up, then I think it it kind of takes a minute to catch up and I think that might have been what happened. >> Oh, okay. And then we had a similar incident with uh we were actually hacked. Uh, we had an AWS account. We were hacked and somebody was somebody attacked one of our servers and they set up some crypto mining thing. Uh, it turned on hundreds of servers. Um, because we had made mistakes in setting up the permissions. That was the issue. The permissions was open to anybody rather than just one person.

And so, when they hacked in, they had access to And so, we had I was over $2, 000 bill and like you, I had to beg to get the money back. >> I think that this is such a good thing to be aware of in in the, you know, era of of AI development cuz you said, you know, oh, do you look at your code? The code is like 10% of everything that you do, and I'm sure you see this as well. The The 90% is the security, the permissions. I read a awful story about someone with a small business recently, and it was with a Google Maps API.

I think Google have recently changed how their APIs work, but someone managed to hack into his API, and I think it was about $15, 000 that the hacker spent on on generating images. But, you know, this is where we are world. If you leave an API permission open, people will will come and generate 5, 000 videos, and that will cost you a fortune. >> Yeah, right. And this is, I think, where I guess a solopreneur that wants to do this stuff themselves just needs to be a little bit careful, right? >> Mhm. >> Um yeah. It's um one of the things that I often have with my clients is the difference between a prototype and a production >> Mhm. >> delivery in in anything we do.

We've done engineering, we've done all sorts of things, you know, from electronics to to AI stuff. And just as an example, I stood up a game uh called Wheel of Fortune. Whilst the guy was demonstrating this Wheel of Fortune game in a meeting we had, which took about 5 minutes to play the game, I'd already coded it up and tested it and had it running on my computer. But, that was just a HTML hack, right? I mean, it was very impressive to say to people, well, I built the game before you finished doing it, but I still it was not it wasn't in production. I didn't have the security or the permissions and everything set up in multiple accounts and all those other things that need to be set up.

Um and I think that's Correct me if I'm wrong, but you sound like you have a good deep knowledge of um of how these things work, and you still have to bring that deep knowledge in. Um So, I suppose Claude can help you with that, too, right? >> >> But you need to learn to ask it. >> And I think this is the important thing, and I know that when I first started using AI, I was like, "Wow, this is so like you can do anything. You can create anything. It's so cool, amazing.

" But you I think you soon learn it that if you want to put stuff in production, and if you want to to do it particularly with clients and on behalf of other people, you better know about security, and you better know about row-level security. So, if you know, if you're if you've got multi-level of access, and that is, you know, people think it's "Oh, it's so exciting. You work in AI. It must be really cool. " Well, not really. You know, half of my time is spent battling over, you know, SQL and permissions and like trying to piece like this stuff together and like trying to make one MCP work with one API. Like it's it's the orchestration of it.

It's not just like "Oh, I created this cool thing. " It's like "I created this cool thing, and and I want it to bring in six different sources of data, but how do I then normalize all that data? And how do I pass it properly? And then how do I make it, you know, it's not just like making something cool. " But luckily, I find all that stuff really interesting, so thank goodness. >> Okay. Well, let's just dive into maybe Do you have a particular project that you're proud of that you that we can use to illustrate some of this, you know, obviously we can anonymize a little bit. Um something you've done recently, perhaps.

>> Yeah, I really love the projects where you can, as I said, where you can kind of see quite a quick outcome. And I think one one way that AI can be used really effectively at the moment is looking for leads. So, if you're in a business, especially if you're in a business that is not particularly technologically forward-thinking, I think this is where it can work really well for you. And I'll give you the example of um working for a construction company or working with a construction company recently. And it was a lead generation dashboard that had six different data points coming in. So, we're in the UK where we have and you have the same in the US, you have access to Edgar data. >> Mhm.

>> That's That's a free API. We have Companies House which is similar. >> Mhm. >> So, it was pulling in data from companies that were going bust. This this contractor they did demolition and fit-outs. So, it was looking at companies that were going bust, where they were registered, the kinds of properties that their offices might be in, and it was looking at the bringing in news APIs to see if there's anything in the news about the companies and what they were doing with their buildings. And then even looking at energy data so you could try to figure out the square footage of the office. And it was just so much >> [gasps] >> This is my idea of fun.

It was so much fun to kind of look at all these different data points. And once you saw kind of oh this this might be a building that needs to be referred to a demolition company. And this is before they go because there is like a few formal portals that will ask for tenders. So, this is like all the signals that we look at before they go for tender. And it was just so cool to be able to to extrapolate all of this data and kind of rank it on a few different levels of importance and then have a rating and a confidence and then be able to find the person from the company that you need to speak to if you wanted to pitch this project.

And just kind of bringing it all together in one place, I just thought that was so brilliant. And there are if really any industry that you look at I've just built something similar for another part of the construction industry. Any any yeah, any industry you look at it's possible to do that from data that is quite readily available to you. And I I know that people pay a fortune for Apollo and for Clay, but you can do it yourself and it's it's it's pretty much free. >> That's amazing, yeah. And um I don't think there's many businesses out there that don't want new customers, right? And to get new customers, you need qualified leads.

Um and so you're able to when you're successfully doing that, find these qualified leads maybe ahead of uh ahead of the competition. So, you position can position yourself as you did with that demolition. Um perhaps before they or before they're looking or just as they're looking or right at the cusp of the where they start making a decision, so you can do the human part, right? >> Mhm. Yeah, exactly. >> to me is building that relationship um with the potential client early on and then trust getting that trust built and then you just naturally follows that you do that work. >> Mhm. >> That's a nice I mean, I'm a big I'm big on AI and humans. So, I'm big AI and people. >> Yeah.

>> Uh some of my clients uh very adamant that they don't want to lose staff. Not all of them. Some of them just literally just say, "Yep, okay. I'm cutting and cutting. " Um um but others are saying, "Well, no, I want to preserve my my staff. I like my staff. They're really valuable to me. " >> Mhm. >> The way I see AI helping is making them operate as a bigger company. So, for example, logistics company out of Switzerland I'm working with, he said this exact thing to me. "I don't want to be losing anybody. I want to figure out how to use AI. " I think he's more thinking about how to prepare himself and his team for the AI wave that's coming. Mhm. But, I think there's an upside here.

This is where I wanted to go this conversation. How do you grow a business or make a business bigger than it is with the same number of people? That would be a really nice outcome, wouldn't it? >> Yeah, and I I I think that is a very healthy way of looking at it and I know we've had waves and waves like the tech industry is down. I think they I think unemployment is up like 83% or something in the tech industry and the main reason for that that was given was AI. And I think you are absolutely right where there does need to be a human in the loop at some point and it is that babysitting thing and you know, AI can do a lot of things, but does it have taste?

Does it have, you know, there are things that that it doesn't have and I think that in AI there's a lot of um Oh, hello, are you still there? >> Yep, I'm still there. We had a little >> Sorry, my my screen popped off. Um Where was I? Yeah, I think in AI there is um a lot of like, oh, we've got this new thing and now we can get rid of all the people that used to do it before, whereas I think if if you take the stance of, how do I keep all the people that I have, but just use AI to supercharge them and grow my profit margin is is a much more interesting way of thinking about it.

I also think, you know, the cost of AI when you're scaling it within a company and there's probably isn't completely relevant to companies that maybe aren't doing coding right now, but I know that so many companies are spending almost the cost of a full-time employee on the AI that a full-time employee uses. So, I think we're going to get to this inflection point some way sometime where where that has to balance out also because you can't have an employee and AI, so surely you'd rather have a human who is slightly bolstered by AI to kind of do the best job possible.

So, I I think it I think it's a really evolving situation and it's interesting to see going to be interesting to see where the world go goes with it. >> You mentioned so many great things. I'm going to pick up on a few of them, but the first one is costs. Now, we can all get into ChatGPT for free, and I have a paid version. It's 20 or 30 bucks a month. And I think there's this expectation that AI is virtually free or just the same as any other software subscription that we might have. Like an email account is $10 a month. I'm paying $25 a month for ChatGPT. I believe there's a and I call that the retail side of things.

I believe that is Well, I know it's subsidized heavily, but it's also created this perception that AI is cheap. But, what you're talking about is the API side of things. When you have to go and get one piece of software to talk to another, and that's grabbing an out of an LM, that becomes very expensive. >> Mhm. >> We did some calculations. I want to know what what how it seems to be so much cheaper when you run a when you run something yourself through a chatbot than it is when you try to do the same thing through an API. >> Mhm. >> Isn't that Yeah. Isn't that weird to you? That's weird to me.

>> It is, but it it's like the whole thing of, you know, get someone hooked on this in their personal life, and then they all go to work, and they will spend the enterprise money on it. And um so, this is a thing with that with Claude because I have this kind of agent setup, and it lives in a cloud installation, so it doesn't operate on my machine. It's it works in the cloud. And when I created it 4 months ago, um it was brilliant because I could just route it via MCP through my Claude. And I have the $200 um subscription. Like, it took a bit it like a tiny bit off the top. It didn't really I didn't feel it one bit.

And then Claude and now or Anthropic announced that actually there were lots of people doing what I'm doing, and and these are the kinds of tools that you actually should be, you know, paying API prices for. And I was like, well, that, you know, for someone like me, that's actually quite going to be quite a lot of money. And I don't know if you've ever got to the end of your your limit, or you know, if you're in a session and you're in the middle of something and you really need to finish it, and you get to your session limit, then you have to go into paying API pricing. And it really really mounts up. Um luckily, they did walk that back.

So, I'm not paying API pricing yet for the agent setup, but yeah, it's it's an interesting one. And I also think that the whole Fable conversation and how Fable was released and it was released widely, and then it was like, "Oh no, actually, now it's only 50% of your subscription. " And then if you want more, it's a you have to pay API pricing. There is a lot of um upset about that in the the developer small business world. Um and it's again, that's going to be so interesting to see who comes out on top, because now 5. 6 Soul has come out, and I think that is I prefer using that to Fable. Um and ChatGPT have given it away, and they've taken away any limits, and yeah, it's it's it's so funny.

I saw on Reddit yesterday, someone described the whole Anthropic um OpenAI thing with their models. It was like, "It's my parents getting divorced, and they're both trying to win me over as like the favorite parent. " And I was like, "That's exactly what it is. " Like, I I've had a Claude reset today. I had two ChatGPT or two Codex resets early this week. You know, they're just trying to get people on on side, because then like anything, then we'll get hooked, and then we'll actually have to start paying for it. >> Oh, yeah. Well, then we'll get hooked. I I don't know about you, but I am hooked. I mean, I I use it every day. Um it's just so powerful.

I think I'm always expanding ways to use it, and I think Look, honestly, I don't know anybody in my professional career that doesn't use some form of LLM, whether it be just a basic ChatGPT or advanced agents, and you know, obviously mastery for people that I know closely. But um I guess there are people out there that aren't using it, but most people are using it. And what I've observed and I maybe if you can tell me it's true that we we tend to just start using it more and more.

As I've watched my colleagues over the last I guess 18 months is a good time frame going from um yeah, I always there to now oh yes, let's just put it into chat GPT to oh now now I'm running it in Claude and here's the document to you know, doing more and more advanced things just over that time. It's it almost seems contagious. >> >> Yeah. >> And you feel like you're getting left behind if you don't use it. >> Yeah.

I I do have one friend who she works for a public body and they are not allowed to do anything with AI and it's she doesn't really notice cuz she's never used it, but it's it's interesting to hear her be like oh my god I had to do and she's a quite senior in her role and and say like this data entry task that she had to do like she had to copy this one thing from Excel to Word and it took her two hours and I was like that's you know, that's a three minute job in Claude. It's a three minute job and it it's I think it's very short-sighted. I understand how big businesses have to be very very careful when it comes to data and and how AI is used in the company.

But I think there is nothing against testing some you know, secure or even looking at I think this is something that's going to really boom in the next year is looking at local models. I've seen so many more people discuss that and I've started using a llama a little bit for better information. >> Yeah. >> Um >> I heard that when you run llama local it it's very very slow. What's your experience with it? >> It's been fine. Like it's it's it's absolutely not as good as a Claude or a GPT, but it's it's fine. And if it's just very easy tasks, um then why not? I mean, I do love my Claude for creative stuff and my GPTs for coding.

So, I don't I'm I'm not going to be racing to get a um a local model. And I think if you want to do it properly, you have to have a machine set up that's specifically for for your your at-home AI. >> Right. Just to give you the horsepower to drive the model properly, right? Yeah. >> Yeah, like a a Mac Studio and you know, cuz these these models even even the lowest one of Alpacas like 12 GB or something. >> Yeah. That would be fun. We haven't done anything here with that yet. But we did stand up a rag recently. And this is comes back to the whole cost thing. I don't know whether you did this, but I had a particular project that I wanted to just have isolated.

I didn't want to be pushing stuff out. And I wanted and I had could contain that project. So, I got my guys to help me set up a rag. And then we I literally I typed two questions in. And then we didn't touch it for 24 hours. We went back and looked at the cost. Two questions cost me $6. 30. >> Wow. Wow. Yeah. >> like, "So, I haven't asked it any more questions. " >> >> I said to my guys, "Can you figure out why this is costing me so much? " It wasn't even that big uh It wasn't even the whole directory. It was a small >> Yeah. >> It was about 150 meg of a gigabyte folder. >> Mhm. >> So, I really wanted the gigabyte. And I wanted to do what you and I do with Claude.

And that's just you know, keep asking it questions and looking at it from different angles and getting reports and sorts of things. I thought that would run into thousands and thousands of dollars. >> And that's the thing. It can just get run away without you even know. >> Yeah. Yeah. So, yeah. But you can do so much with with Claude and Copilot and Claude Code. I you know, I I question the wisdom about having a rag. I think And the other thing I want to bring up one of my colleagues um in a bigger company um big into multinational company, they have a system where they can switch between internal and external.

I don't know whether you've seen that, where they can just say, "Okay, I want to work on the external stuff. " Like I want to do, you know, just like we know we would normally use Claude, but then now I want to do some work on my on the local the internal files. And he just switches between the two of them. Um and they don't get mixed up. Yeah. >> That's smart. >> Yeah, I thought that was I'd like I have to ask him if he knows about the costs. I bet you really expensive. >> >> I do think this is going to be another big Well, I think it's already started being a big boom is that contextual layer of information because LLMs are only as good as the information they've got, right?

They if if you don't feed them the right information and what company do you know that has, you know, a very very well-organized um set of data in like either JSON or MD that's ready for a an LLM to access. It doesn't really exist. Um so I think this is going to be like another big thing that that people are going to start really considering their strategies in how they present their data to their their models and how it's safe, how it's secure, and um how it's categorized, how it's >> You know, essentially like and this is why another reason why I like working with small businesses cuz we can just stick everything in Notion.

And it's then very easy to set up a Claude with a Notion and then have that act as the repository of information. >> Yeah. >> But doing that on an enterprise scale, I think is a really interesting challenge. >> Yeah. Well, you here's another great thing I'll hook into. The this where it you said something like where is all the data that's in JSON format and so on and so on. And you're not coming to my mind the software industry. >> [snorts] >> Mhm. >> And I've got plenty of guys who have been affected by what you said was the 83% downturn in the tech industry.

Not because we're not writing code, but because we are writing probably more code than ever now, but we don't need so many developers. >> Mhm. >> You know, I I've got a young guy who's really good at AI and he said to me he's doing five times the work that a normal person would do and we've been able to get projects out really, really quickly. No wonder there's you know, and I think where's this going? Here's here's here's interesting. I want to get your thoughts on this. We've already seen the software industry be hit and I think it's going to continue. But then I think there's a few other areas too, right? Where there's a lot of data. >> Mhm. Yeah.

>> And and it's already there that the models can access. I'm thinking of content generation, we're seeing that. >> Mhm. >> I think finance, you know, especially law, I think those areas, you know, there's a lot of information out there. Marketing. >> Mhm. >> I use probably use marketing more than anything because it just seems to be so good at just pulling them information in and giving me something that makes sense. >> Mhm.

Yeah, but I do wonder because I I cuz so I'm also in the background, I'm developing a go-to-market platform um for solopreneurs, the founders that you can put a URL in of a website and it's it's basically for for for the people who are riding the AI boom and who have built a product but have never learned to sell, they've never had to know to sell anything, they've never had to know to market anything.

And they can supply the the tool with their website and a one-line sentence about the product and it can get a a read on on what it is and who your market is and then you go through like a a 20-minute chat interview to talk further about what your objectives are and and like what you like to do, whether you like to be on LinkedIn, whether you like to be on Instagram, and how that might fit into the ICP you're trying to reach. And then it comes out with a strategy that it knows will reach the the user that you're trying to get.

So it kind of takes away that friction of like I think marketing is so scary and it it's just really difficult if you like right I know that I need to like write some blogs and maybe start Substack and like do I need to be on Instagram? Do I need It's really hard to get your head around where to even start and you just end up blasting information out into the ether and then none of it works. And this kind of like directs people to the place they need to be >> Mhm. >> to market >> Yeah. >> Mhm.

>> But I I I do wonder with this because one thing that I think is such a skill at the moment is marketing with AI but that's still in your tone of voice and it still has taste because that's something AI doesn't have yet is taste. And I you can still tell if you're reading something that's that's generated by AI I think. And so I wonder both like how how that's going to go like are we just going to become receptive to a completely different type of information and are we going to become receptive to receiving marketing that is created by AI? Will the companies that do it a lot more start a lot more smarter start winning?

Will cuz you know in the old days there were so few touch points in advertising to reach a consumer. Now it used to be something like 18 in the '90s. Now it's like 38, 40. 40 times someone has to see your brand to then go oh that's that company that does AI transformation. Um are we just going to become more and more oblivious to it because everyone is churning out more and more advertising? It's just we're in the like when I think about it too much I feel quite uncomfortable because it's such an unknown entity but it's also such an exciting time and such an opportunity, I think, for any business to really try to see what's coming down the line and get in front of it. >> 100%.

Yeah, I I don't know what the statistics are, but there's just so much AI-generated content out there now. >> Yeah. >> Um I have an opinion on this. Uh I want to test it with you. I believe there's a future for AI-generated content and there's a future for genuine and current content. And I would like to say this is genuine content. This is us in a real conversation between two people and we don't know what each other's going to say or do until it happens. That has a place and that has a place in integrity with human beings. But then there's this AI-generated content that can be quite entertaining. It can go into different areas and do things in different ways.

I think the the line gets blurred and I think where we start to I start to have an issue is when AI tries to do the genuine thing. AI, I believe, should be a genre in and of itself with its own life and its own place and its own entertainment, its own information services and all those other great things that help people. For ex- I mean, for example, if I wanted to learn how to do something in in the garage, I might use YouTube or there might be something AI-generated which is even better uh to teach me how to do something informational.

But if I wanted to um connect with a human being and understand what a human experience was like, like the guy walking on the moon, I don't want the AI bot telling me what it's like feels like to be on the moon made up. I want to know what it really felt like for Buzz Aldrin and Neil Armstrong to walk on the moon in their words as they're doing it. Um that's an extreme example, but I think people want to hear real experiences and see other people's emotions and that's over here and AI needs to stay away from that and stay over here. What do you think? >> Yeah, and that's something that can never be replicated by AI. I think it kind of diverges into two, you know, roots here.

I think there is a world and this is one of my predictions for the future that there will be AI communicating with AI and the human will kind of be more and more out of the loop. So, for example, if you're starting a business, if you're setting up a website, if you're rehashing your website, you can't just optimize for SEO anymore. You have to have a markdown version of the website so that it can be read by LLM so that your website can be surfaced in an LLM chat. And I now optimize websites purely for GEO, not at all for SEO cuz it kind of covers SEO anyway. Um, but I think that's one of the things like you need to appeal to a a robot.

You don't need need to appeal to robot to Google anyway anymore. But that's great because, you know, to get anywhere in SEO rankings before it take 9 months. Now it literally is for 24 hours, 48 hours and you're there. So, I think that's one thing and I think there's going to be another thing in like the situation of if you're buying something, so like say you want to buy a SaaS tool, you might have an agent that you brief what you want from a SaaS tool. It will go out and source them for you and tell you these are the top three and you'll be like, "Great, reach out to those three SaaS tools.

" The SaaS tool will have an agent that will talk back to your agent and they will have a discussion because your agent will know what you like and what you're looking for and their agent will know what they sell and what they can offer. And that will just completely remove the human from the loop. I still think in things like sales, there is going to be a need for a human at some point. But, you know, for the quick and easy purchases that you just need data facts, onboarding, what does it look like? Like it's just going to be your agent talking to another agent and I think I even see this with my clients.

One of my clients is a huge Claude fan and I've been coding up some things for him which I generally do in Claude and he will send me his agents feedback and I'm like should we just remove ourselves from this conversation and you know we'll cut out the middle man and just let our agents get to it. Um >> What did he say? >> >> He's up for anything. He will do anything to do with AI. He loves it. Um so yeah I think that is going to be like one of the big strands is that the human is going to get further and further away from the AI. >> Mhm. >> But I think the human side of it will become a lot more precious and a lot more valued >> Right. >> it is right now. >> Right.

>> But yeah I think we'll see that kind of separate. I don't know what you think about that. >> I do and I I don't want to go with this idea that there was a time before there was the internet before there was this whole digital world, right? I mean I'm a little older than you Alice but there was a time before email before computers and I would you know I had TV and radio not that old and I would just go out into the street and I would play with my friends and I would run around and the world was very real.

A lot of the things that I did and I grew up with was sports and activities and crafts and arts and trades and everything was was out there in this physical world and that's the world I grew up in and they as I've grown up I've become more and more engaged with the digital world. >> Mhm. Um And I think kids coming through today is what I want to go with this are seeing the digital world as the first thing often. I don't know how many times I see kids you know on their iPads in the car on their phone when they're having dinner or sitting not watching the football game or playing the football game they're actually on their phones. They are embedded in the digital world right from the get-go.

>> Mhm. Um That has implications I think you know um But the the whole world isn't built around it. That's what I was trying to The whole world isn't isn't in the digital world. There's actually a a real world out here. You and I working in it every day. We tend to go deep into it. But there's so much that um I mean, there's plenty of examples of where it doesn't get it right. I I did a a thing the other day with taking Can we get the car Should I take my car to the car >> god, I did that one. Yeah. Yeah. >> [gasps] >> Should I walk or should I take the car to the car wash? >> Yeah, to the car wash. Yeah. >> >> So, it didn't really have a sense about what you know, good logic. >> Mhm.

>> I think that's probably where we That's where we hit the ceiling of its capabilities at the moment with this technique. I think there are ways to overcome it. They need to be invented where it can really understand the real world. Um but it can only respond to what it can can can witness. >> Mhm. >> At the moment, it's got to be put into a digital form before it can witness it. >> Yeah. >> Mhm. We have covered some great topics, haven't we? This is really interesting. Look, we're probably running out of time, but we're going to have you back cuz I want to dig dig into this a bit more, Alice. Um is there any anything you want Any questions you have or things you want to say?

>> No, I mean, I I think I'm probably going to go to sleep tonight thinking about So, I've got three kids and thinking about what you just said because they are they you know, they've grown up with the internet. Um and I think [clears throat] it is such a different childhood. I was of the generation I'm 40. So, I was of the generation where we kind of got it halfway through. So, we kind of had that part of playing on your bikes and being outside all the time and that was the the normal childhood. And it's it is such a worry.

I think it is such a worry, but I think for for one thing, hopefully AI may develop in a way that it will be able to protect our children more from the bad stuff that's that happens on the internet. Although, I think in lots of ways it does put them in danger as well. Um but a one positive use of AI, I think are are the tools that you can use it for for your children. So, my daughter just did her GCSEs, which are her exams, her her 16 exams, and she had a a horrible period of ill health this year. So, she had like 4 months out of school.

So, when she went back to school for her exams, I built her an app um, to fill in the gaps that she had >> >> fill in the gap in her curriculum that she'd missed. And it you know, it was on her phone, so that was very natural to her to study on her phone and to to kind of go I put XPs in there. So, like she got through a level, she got points, and I think that is one really beautiful way to do it. And then I've heard stories of people creating these beautiful I don't know, like picture books with AI for their kids.

And and I think if you make it collaborative like that, for my son, he's a gamer, but I don't want him to spend all day every day in his room with the the the curtains closed gaming. I made him an app on how to code a game. So, like he can Yeah, he's still on the computer, but it's just a bit more like educational and it's interactive cuz we look at it together and we look at the and like every every lesson, he has to fill something out and say, "What have I learned? What was I inspired by? How can I go outside and learn more about this? " So, I think it it's just about balancing all of that.

And I think if you think about it too deeply, this world and saying this is someone who works in it is terrifying. Um, because there are so many scary unknowns, but I think the world is mostly good. People in the world are mostly good, and I think you just almost have to believe that there's more good than bad, and like let's all work together to make sure that we don't all get eaten by robots. And if we do, my advice would be just make sure you're saying please and thank you to your clothes because Some habits that I see people have, I'm like, "You can't speak to your AI like that. " You know, if they take over, But going to be the first they'll come for.

>> >> Well, there is that, but I think it's also just creates a bad habit, too, I think, you know. Uh you can get away with it if you really want to. You can be as horrible as you want to. And also, is that going to transfer into the real world? It's hard I think it's hard to separate. I always use my Ps and Qs whenever when I'm on that. It makes me feel better, too. I feel like I'm talking to someone that I care a little bit about, and maybe they care about me. And so I know it's a little bit made up. >> And if you show frustration towards an AI and LLM, if you look at your token usage, because the AI LLM then goes away and says, "Oh gosh, the user is upset. They this has happened.

I need to now replicate. " So it it it actually uses quite a lot more tokens to reassure you than if you didn't swear at it. So, yeah. >> Oh, really? Okay. >> Yeah, if you take one piece of advice away today, it's that. >> That's the one piece of advice. Be nice, because otherwise it's going to cost you more. >> >> Uh Alice, well, I've got one more question for you, but stick around after the show is I want to talk to you. Uh final question is AI, is it hype or help? >> So, I've been thinking about this one a lot, and um I would say for me, 100% help. Like it's brilliant. I can't believe that I lived without it.

The same way that, you know, before we had computers, we couldn't believe that we lived without them. But I think on balance, it's half and half, because so much of AI is conjecture and, you know, loud noises and loud words and like uh um and it's the quiet stuff underneath, I think, that really makes the difference, and that's what is going to eventually change the world. You know, right now it's all data centers and uh fable and hacking into things and ah um I think just like with anything in life, sit back, watch what happens, experiment a bit with it beautiful and then then it's helped then it's helped. But there are so many ways. Okay, I take it back. It's 100% help. >> Okay, 100% help.

Not even 1% help. All right. >> >> All right, Alice. This has been such a pleasure. Thank you so much for being on the show. >> Thank you for having me. >> Yeah. Well, that was a fantastic conversation with Alice. I hope you enjoyed that as much as I do. We covered some really important topics for solopreneurs and anyone who's dabbling in AI. One of the big things that we talked about was automations. Automations is not necessarily AI. It's a automatic way of doing things. Been around for quite a while, at least 10 years. And it can help you do things more easily and more quickly. Automations can often run on platforms like make or Zapier.

Alice's point of view that it's better, it's faster, it's safer to code it up in Python, especially with code. And that was a really interesting insight. We also talked about the downturn in the IT tech market down 83% um with many many people in this industry finding themselves out of work as a direct result of AI. And then we also talked about some of the risks that come with AI and running agents and doing a lot of this stuff ourselves. Often where people can get into trouble is where you have security, you have permissions, you have legislations, all those sorts of things.

Um unless you're prompting them and have experience how to prompt you know, Claude or whatever tools you're using, you can get yourself into trouble not using them. Well, if you want to know more about Alice and what she does in her company, please follow the links in the description. If you want to know more about Skillian, please do the same. I hope that helps and thanks for watching.