So I founded the company about a year and a half ago and I started it because I’ve worked in tech for the longest time and I understand SEO, I’ve done just about everything you can imagine and all of the things that really would be required to know how to run a blog and so I started my own blog and I was like yeah this time I’m gonna do it great I’m gonna really focus I’m not gonna get distracted I’m gonna really produce something that’s unbelievable so I made my blog I started writing optimized it I made sure everything loaded fast and I was really enjoying it it went well I was happy.
Then about two weeks later I was like, oh my god, this is a lot of hard work, I can’t believe that I’ve got to do this, there must be an easier way! Being in the development space I obviously knew that you could use AI to produce some great things. So at the time Jasper / Jarvis was around, they were using OpenAI text generation models like Ada, Babbage, Curie and DaVinci models. They were doing a pretty good job. I was relatively impressed by it but I knew they could do much much much much better and I was surprised that they weren’t, especially given the amount of funding that they must have had behind them. So I tried to have my own go see what I could do to see if I could make something better I turned out I could, and I was using that on my blog it was doing really well.
I told some friends about it, they were really interested and they asked if they could use it too. So I put a front end on it and from there very quickly discovered that what I was writing about didn’t translate very well to what they were writing about. I then had to adjust it to make sure that the niche could be more flexible and that the content produced could be more flexible to match those different niches. That worked really well and then those friends told their friends and then I had more people asking to use it and from there it just grew really. Obviously as I added more features I needed to add more UI to it and then at some point I was like hang on a minute if all these people using it I’m sure I could charge people to use it. People were already offering to pay me at this point so it just seemed like the next sensible step.
Ultimately I just wanted to have a blog that would produce content for me automatically that would be high quality and would have all the SEO done for me. Every time you produce a bit of content there’s a big checklist of things that you need to do I was making sure that I’ve done those things every time.
As a developer I know anytime we’ve got to do something repeatedly, you can automate it. So I did that.
Relating to the current state of the AI industry in terms of growth and customer acquisition, I think it’s very much a growth sector. Obviously there’s a lot of hype around AI right now. The interesting things to me is that a lot of it is is cheaply veiled glorified ChatGPT prompts so there’ll be entire websites that are like “we’ll write blog posts” and ultimately all it’s saying is “Chat GPT write me a blog post about cats” and that’s it they’ve done nothing else extra to it other than add a pretty UI to it. I think that hurts the industry by them doing that, however in terms of the actual players that are doing really great things; they are doing incredible things. It’s really interesting to see what they’re doing, it’s all ramping up, however I’ve noticed that no one seems to have done anything to lock their market share in. At the moment, as cool as ChatGPT is, it’s not that hard to switch out to a new player.
Let’s say that Bard takes off and becomes unbelievable, not seeing that happen just yet, but there’s nothing to stop people just swapping out. It’ll be really interesting to see if they can come up with some kind of lock-in. At least with the ChatGPT UI there’s very much: you can use the web pilot, you can use plugins – they’re great examples but that doesn’t exist in the API right now. Lang chain provides abstraction layers to allow people to do all those sorts of things, which therefore means, you don’t need to use OpenAI. Maybe Claude takes over, we’ll see what happens.
Re: the challenges that startups need to overcome in the future to remain competitive – that’s just having something like I said something that locks in to your product something that you do that’s better than everyone else. I hate to see people compete on price because it shows that you’ve got nothing to offer, if you’re just the cheapest then brilliant that means your products are not really standing out above anyone else’s. We don’t aim to do that, we aim to offer the best product. I fully think we are the best product on the market and it’s only growing from here.
What accessibility to various types of AI technologies enable entrepreneurs to compete with companies that have larger funding and resources – so this is a great question. I actually think AI has been the great equalizer at work; I know someone that doesn’t know how to code but they’ve used AI and they produced some high quality code. It’s scary!
As a developer that’s had many many years in development and coding, I know I can do a good job of producing really complicated systems; but there’s people that don’t know how to code and they can produce something pretty good. At the moment they’re limited by context length, so they can only have so many words that can be output, but that will change – we’re only at the start of the AI revolution. Give it not too long and they’ll be producing some unbelievable results that can span multi files.
I’m using AI right now to write my unit tests for me, I couldn’t do that before – it’s fantastic and this is a very very early stage of the change. It’s only gonna get better from here and it’s a great equalizer.
I’m someone that knows what I’m doing, but there’s plenty of people that don’t and they will get up there. Our company provides a tool that allows people to write high quality blog posts that do SEO for you. A person doesn’t need to learn SEO anymore, it’s just handled. Think how good that is! People spend years and years and years trying to learn that skill and it’s just done for you. They don’t think about it. They don’t need to know that it’s going on. They just know that it has happened. It’s fantastic!
In terms of disruptions in the current AI industry – I think obviously products like mine are gonna completely change how the internet works when search generative experience (SGE) from Google comes out. That’s going to be really interesting for what it means for websites, because who knows where it goes from here. If you don’t need to go to a website, what does that mean for websites? And if we don’t make websites, then where does the generative search experience get its data from? So it’s a bit of a catch-22 that I’m not sure how it’s going to play out in the long term. But hopefully people are still going to want and trust personalities and content producers to give them that that real experience, and that will give a reason for people to still produce the things that they produce.
I also think the way that hardware works currently is going to change significantly – so right now we’re all bound by graphics cards whereas things like ASIC (Application-specific integrated circuit) will come out and they will be designed specifically to attack machine learning problems. Right now graphics cards do a good job but they’re not specific for the actual task at hand, so an ASIC, just as they came out for Bitcoin, will come out for machine learning. They’ll be fantastically powerful, they’ll be cheaper to run, they’ll be more efficient, they’ll run faster. It hasn’t happened yet but I’m sure there are people working on it, and it will come out.
Another thing I think that will change is that right now computer hardware is very very very much focused on being fast, but memory is the demand that is causing all the machine learning models to struggle. Right now you’ll get 40GB large cards in the data center, you get 24 gig at home if you’re lucky and buy the very expensive ones, but the model sizes if they get bigger that’s going to become a problem. Realistically what’s likely to happen is we start having expandable GPU memory. Why struggle and have to have 12 cards just to have the video RAM you need, when you could just have one card of expandable memory. Maybe the memory is slightly slower, because it’s expandable and it’s not on like next to the chip, but it will be good enough. It will get you a lot further than not having that expandable RAM will. So I really think that will change just the design of systems going forward, it will be interesting to see what happens.
So I guess in terms of specific skills that I think individuals should focus on development: I still think software dev. As much as I’ve said it’s the great equalizer, I still think it’s very good for logical thinking and logical thinking is problem-solving. How machines run and work better, as much as they’re creative themselves, it’s very good and as you may have read whenever a large language model is trained on code: it actually becomes better as a model. Not just on code but also on text; because of the logical thinking that’s required. I also think that psychology is worth learning unusually, because actually how you talk to the machine is very important so if you’ve ever said to someone they “word vomit” and they should “think before they speak” – this is a great example. When you take large language models and you say “do this for me”, and you expect it to do it straight off, that will produce a bunch of nonsense (or more likely nonsense). Whereas, if you say “I want you to think about your answer, I want you to take this action, then this action, then this action” – generally it will produce a better result. Same as if you “think before you speak” – you get a better result. So we just want to make the most of that really. Finally I guess creatives are probably another good place to be – as much as you could argue that these models have come out and they’re really really creative to begin with. I think it’s nice to have a reign on it, and I think a lot of the models at the moment sit on the fence. So they never have an opinion. So one of the things that’s really good is to just say oh I I really like red – so let’s roll with red. Whereas a lot of the models now are like “I like red but there are other colors”. That’s not that’s not good for product development, you need to take a stance, and roll with it. Which is cool but that said (and I’ve recorded this recording twice now), there’s another thing that we need to think about. So originally I talked about reigning in creators and stopping them like reigning in an AI; and stopping it doing something that’s completely crazy.
Actually there’s a concept of a GAN, so a generative adversarial network, and that’s where you have two AIs competing against each other. One’s trying to do something, and the other one’s trying to say “hey you can’t do that”. So in this instance we can have one that comes up with the wackiest craziest product idea ever and another one that says whoa hang on “cars can’t fly”, or whatever it might be. Actually there’s already some really good examples of stuff like “GPT Engineer”, “Meta GPT” and “AutoGPT”. All these things that create fake companies, and they have different roles. So they have product managers, they have developers, they have code reviewers, they have testers, and they will all work in a big little little cluster of people and they will produce code for you. They’ve run similar experiments with entire cities built out of these larger language model AIs, and they all interact together and they produce something unbelievable. It’s very interesting to see where the industry is going, I don’t think anyone fully knows yet, everyone just knows it’s crazy and it’s going to change everything. It’s very exciting!
In terms of funding, I was lucky enough that I was able to support the funding of the company, thankfully. This is nice, I’m quite happy with that situation, maybe as we grow and we need larger marketing spend, maybe that’s something I’d look for outside funding for, but I’m not really seeking VC funding. I’ve been offered, by quite a few people, funding but actually I think we’ll be alright for now but I’m sure there’s plenty of companies making a lot a lot a lot of money from VC funding. Even in this recession, AI is the buzzword right? People want to be in on it because it is the next amazing thing.
Some examples of users of Article Fiesta are people that want to rank well in the search engines but don’t know how to. One of the things that everyone knows is you should be producing blog content for your website, everyone hates it, no one does it but the people that do do it, get a whole bunch of free traffic from Google. If you don’t, you have to pay for adverts. One of the things we do is we allow you to just find your keywords, set up a schedule then put your feet up on your desk whilst your website does all of the hard work for you. You just get free traffic, it literally is a win-win. So that’s one of our best use cases. We also have people just looking for a side hustle, like people looking for affiliate income on the side; they create some website, for example reviewing cat cups or something and then ultimately people will go to an Amazon link, buy a cat cup and you will make some money, everyone’s a winner.
A particular customer segment that we’re targeting? Actually quite broad, which I know is not always the best idea for a business, but we appeal to any company with a website really. Or anyone that wants to earn money online – which again is really really broad but really fantastic. Just the same with sites like Wix or any website builder – there’s so many people that could use their service.
There are so many things in the product roadmap. So one of the tools that I really like, and have taken a lot of inspiration from is Surfer SEO. So Surfer SEO analyzes the top results in Google and it will show you what keywords have been used throughout the text (or key phrases or n-grams) that ultimately are being used to help them rank in number one. We do a lot of that behind the scenes, but I would love to expose a similar sort of UI to how Surfer works just because I think that it just looks really nice. I just like it, and it shows what goes on in the background – because at the moment a lot happens in the background but where the users not exposed to it it just feels like it’s writing an article. I want to highlight all the cool things that go on in the background. If only to make it more obvious.
There’s also a bunch of stuff that is on our radar, for example seeing it written in real time (actually already coded just not rolled out), being allowed to select your own specific Amazon product and see a review of just that one, there’s so many things! I don’t want to go into all of them because I don’t want to give all my competitors ideas!
What sets our product apart from the competition? That’s easy. We do so many extra things, and we do actually rank. Like I can say there’s a lot of blog creators out there, and cool they produce a bunch of text, but it doesn’t rank. Look at any youtuber that tries to review their product over a long period it doesn’t work. We not only write the content, we optimize the content for SEO by looking at all the other top results. We make sure we write the meta description, we make sure we create or use existing categories. If we create a category, we make sure we do the meta descriptions for that. We do the tags. We include images we include videos, we create dynamic graphs within the text, the images that are produced are ridiculously high quality. They’re custom. It will have text overlays on them, they will have dynamic little quotes from the article. We use tables, we use bolded text. We do so much to make sure that we match the EEAT requirements of Google! Basically if you want something that you know is going to do a good job, we’re the people to come to. If you want just a bunch of text, you can use one of those cheaper writers that are out there but just don’t expect to rank.
In terms of partnerships, I’ve not actually reached out for any but I am debating reaching out to maybe like the SEObility’s of the world. Something that would be a nice complimentary service, so something that does a site audit or something maybe SEMrush or Ahrefs something like that I think we could work really well together, but in honesty I haven’t had time to reach out for partnerships. I’ve been so focused on making sure the product is as good as it possibly can be, but it’s something that we can definitely explore in the future!
As for frequently asked questions we have? I think most of it is handled by the website. If a question is “how can I write high quality content”, then it’s just enter the keyword and click go. If it’s in terms of finding the keywords, there’s a few videos on our YouTube channel that can help resolve how to do this. We give some great suggestions in terms of like internal linking etc it’s all done for you literally this product is so hands-off and it just handles everything you possibly can think of.
But anyway, that’s hopefully everything that everyone wants to know but I am of course available for more questions and I’ll catch up with you soon!
Ash