Category: Daily Report

Currently there are three AI projects going on. Without giving too much away right now, I’ll call these:

• KTX
• DXP
• SQA

KTX is my platform built in Laravel and I see it as a strong reliable backbone for everything I do going forward. I see it linking in with other AI services. It will serve the main web interface that eventually paying customers will be using.

Today’s main focus was on getting monthly subscriptions configured and setup along with Stripe integration. I’ve got this done but there’s plenty of QA, configuration and understanding I still want to do on it. But the basics are in place and eventually I will be able to take monthly payments from people. I kind of have an idea about what I can do with this in the short term.

DXP is a prebuilt platform that I am very grateful to have the opportunity to work with and help drive forward. Now’s not the time to delve too much into that, but past few days we’ve got it to a point where we can start taking it for a spin. A lot of it is around content generation, which is obviously highly competitive, but there’s quite a lot of good points to the platform and it would be naive to not investigate its potential.

Today I had a conversation about high level strategy around the DXP.

SQA is a potential third platform that a friend is prototyping using bolt.new and supabase. It’s amazing to see what no-code tools can output now. Time will tell how resilient the code is when it comes to actually being used by people but I can’t deny how cool it is to see something like bolt help someone with no programming experience helps get their ideas down into prototype form – not just UI, but actually functionality including Supabase.

Today I had a brief look at the progress on his prototype, and was amazed at how it was starting to come together.

Yes these tools help us make UIs faster, but are they just digging themselves into a hole faster? Time will tell. Programmers without AI are now at a huge disadvantage, and I know of one good friend who pretty much just directs an AI in everything now. But he has the experience and knowledge of how to structure an app, and WHY it is structured in such a way.

Anyway this is a conversation that’s going to drag on for a while. Many programmers will refuse to think AI is going to take their job, some will think it supplements their job, and others think it has fundamentally changed their job.

We can build UI and functionality far faster now … but it’s always been about making software that’s wonderful for humans to use. It’s about the design, not just aesthetically, but how the software itself is engineered to user flow. AI is the best pattern recognition tool out there, and it will recognise best practices; the good news is that humans will have intuition to point it in the right direction.

Hopefully anyway.

That’s it for today.

I’ve made the decision to attempt to build an AI Company. The other options were to:

• Find an normal job at a decent startup as a web developer in a leadership role. I have done plenty of project management, product ownership and development over the years so whilst there are many better programmers out there than me, I do have a lot of skill and experience to offer any startup. And I love it, still, 30 years on from building my first website on GeoCities. Problem with employment is I do start to feel trapped in a box, and having to be somewhere every Monday morning.
• Carry on as freelance but improve scheduling workload so I wouldn’t constantly be juggling multiple clients and projects. Freelance is probably the worst way to go self-employed because over the years it’s a war of attrition on your energy levels. If you don’t get it right, it can really grind you down.
• Find a decent contract. Contracting was super attractive because you can make thousands a week if you find the right one, and you get to focus on one project at a time – which is much easier on the mind, and you are able to give your fullest to it. And once the contract is over you can take a well earned break.
• Build a web agency up from scratch. If you are going to manage multiple projects, you are better doing a web agency of sorts. But then you have to deal with people and building a team, and that comes with all sorts of problems. A lot of agencies now are struggling because, in reality, most of the digital transformation has now been complete, and the availability of tools for websites and software, together with the increase of supply of capable people mean the profit margins have been hit. Of course, the best agencies with the great service and product, will do well… but building one from scratch now, was something my heart wasn’t in.
• Do something completely different. Learn a trade such as plumbing or electrician. I may still get training for these qualifications since I’d like to know how to wire up a house or fix plumbing. I think AI will probably shrink the information / service workforce and going back to trades is a good bet for many.

I’d been considering all of these things at once since the start of the new year, perhaps longer. In the end my heart wasn’t in any of them and that’s why I was dragging my feet.

I’d always wanted to build a product, and whilst ten years ago would have been an ideal time to get started, and whilst right now is a really challenging time in my life to do so… it’s also the best time. So I’ve got a hard deadline, a bit of runway, going to do my best.

Some Further Rambles

I haven’t yet got around to doing videos on the tools available to us right now, but some of them are awesome and transformational/paradigm shifting. I will do videos eventually, but the blog is good enough for the moment.

The challenge of building an AI company has a few components:

• Yourself and your energy
• People who want to work on the foundations with you
• Gaining exposure
• Understanding the technology
• Keeping up to date with the rapid progress
• Being willing to keep on working despite realisations of intense competition
• Using AI within the company to build the company
• Working out how to give value to people already running companies in a way that’s different from the plethora of options already out there

It seems to me that people who work hard to keep up to date with the AI world are going to become the experts at making recommendations as consultants, so there’s clearly an opportunity for web design and software agencies to ‘upgrade’ themselves to ‘AI transformation experts’.

Side-note: Brave Browser Leo

Let’s you summarise webpages.

Side-note: Really Simple Example of AI Monetised

I was shown a really simple WordPress plugin today that charged users for filling in alt tags on images. You send the image off to the server and it sends back a really good description. I don’t know how much impact alt tags have on SEO anymore, but it saved significant chunks of time. This is a really simple example of using AI to create an income stream.

Building an AI Startup could involve:

• Optimising the algorithms with which the language models are made, to reduce the cost of creating them in the first place.
• Creating more efficient hardware with which to process the language models.
• Training the language models
• Building software that uses the language models to create new markets, or to optimise current business delivery

It’s unlikely that I can do any of the first three, so the last one is the most obvious. For the moment I am just talking about Language Models here. Deep learning / data analysis is part of the AI game but for the moment I’m just focusing on Language Models, since they are all the rage.

We also want to look at image, video and audio creation; since these will significantly disrupt industries that use them day to day. My last job at homemove.com was an amazing example.

Beyond their excellent use of AI agents to enhance sales, one thing really stood out and that was an AI generated video that was created fairly quickly and inexpensively. It was TV commercial quality (ish). That was enough for me to know how disruptive this technology is going to be.

Long Term Aims

• We are each going to have a choice of Virtual Assistants. These will help us navigate through life, and will abstract away a lot of mundane tasks; and will act as a firewall (protection) from the exponentially increasing amount of information out there. It will use processing power to help us make sense of the enormous amount of information out there.
• Large corporations will also have virtual assistants for each employee. And each of those virtual assistants will likely be linked to each other, and especially those of their managers! As you go up the tree, there will be some sort of overriding ‘AI God’ that is monitoring it all, reporting on it, and most probably taking action.
• Governments and militaries, if not already, will have extensive AI monitoring systems. The capacity for summarising overall sentiment using just their ‘backdoors’ to Facebook, is quite frightening.
• I think buildings and cities themselves will have overriding AI brains. Call them AI agents if you like. They will be real-time systems and I suppose not much different from current software excepting they will have surveillance analysis, real-time number crunching and some use of LLMs.

I don’t think AI needs to always be done in a creepy or sinister way. AI surveillance if done correctly can actually anonymise privacy if we chose to want to do that. Currently the UK government is intent on facial recognition down to each person, much like the Skynet program in China.

Anyway, these are also just projections from within our own paradigm of life currently. We’re already seeing fully automated sea ports and internal logistic delivery centres, so I expect that fully automated companies driven solely by AI (with an overseer) to be running. The nature of a corporation as we know it may well change. But for the moment we just have to ride the wave and see what happens.

My long term aim is to find a niche within the Virtual Assistant / AI Agent market, both for individuals and for corporations. There are plenty of people jumping on this bandwagon already, but the market is going to be huge for it.

Anyway it’s been a long day, so will leave with just a short work diary below.

It’s a Saturday but I still managed to get some time in.

Invested today attempting to get some more Docker configuration fixed. Difficult when you are on a brand new tech stack, and stabbing in the dark with console logs. The main problem is the local stack doesnt seem to like my host environment and it’s become a slow process figuring out how to sort that.

Most of my time was spent on configuring the DXP software, which is an existing AI driven platform that I can help take over and run with. Ultimately it listens to RSS feeds and generates websites from it, but the technology there is adaptive to many other things and is built to scale.

That said, haven’t been able to get it configured so spun up an Ubuntu box on Digital Ocean with Gnome Desktop. Just about got it and running but its now maxed out at 100% CPU and I can’t login to it currently. Going to let it rest for a while, then come back to it.

I have some ideas now how to fix it.

Also Digital Ocean has its own AI Agent Capabilities. Something to look at in the future

https://cloud.digitalocean.com/gen-ai/welcome?i=6b4c9b

I’ve had quite a few ideas for a while about the way forward. But I’ve been hesitant to commit to any of them until last night when I realised I just had to burn bridges behind me so I had to move forward. This is a small reference to a chapter in Think & Grow Rich, which I read a long time ago. Shortly put, you have to burn your bridges behind you so there is no source of retreat. You either win or perish.

In my case, I’m not going to fully perish (hopefully), it’s just that I have given myself a hard deadline of October 2025 since that’s when my available money is going to run out. At that point, if nothing has come of it, and it really seems like I am up against a brick wall… I will concede my entrepreneurial dreams and go work on a funny farm. Or more probably, I’ll go back to employment.

Client Work Gets In The Way

So in the next eight months, I’ll avoid as much client work as humanly possible. I have mostly reduced my commitments, but there are two projects that I’ve made a commitment to support and don’t want to give them up. I potentially will do consultancy in the interim to get some cash in, but ultimately the main challenge over the last few years is the balance between client work and having an entrepreneurial seizure. Client work is always the easier route to putting bread on the table but the time and mental energy it takes to do that, is taken away from the singular focus you need to build something yourself. So I’ve made the mental decision to allow myself to focus just on building, and use up my financial reserves to keep on living. I will be pretty much broke by October 2025, but that’s half the fun!

Public Commitment

When I realised this idea last night, the main theme of it was making a public commitment to going down this route. So I’ll be making a story out of it on LinkedIn, in combination with this website, and across social media. It’s somewhat daunting, and I’m sure plenty of people are going to have negative things to say; and many will expect failure; but it feels right to do this. I expect to make a YouTube video each day on my findings, and then put that out on LinkedIn daily. It’s a bit up in the air at the moment, but I’ll figure it out as I move on. Will also need to get on Insta and Twitter but will deal with these later on.

Team

I do have some really good people around me, and have been chatting with them recently on starting something up. Various ideas swirling around. But I realised that I needed to take the lead and march/bumble forward, and if they want to join me on the journey then that’s great. Some of you reading this will recognise you are in that group of people. Up to you to see if you want to join. For the moment it’s just me.

AI will take my job if I am not aware

I’ve used AI helpers in programming for a few years now, and quite frankly it’s been an eye opening process. I’ve programmed and worked on the web in general, starting as a hobbyist, then professionally, for 29 years now. To see most of my knowledge advantage now annihilated by AI, and to realise that my programming career is all but over in its current form … this took a while to get over the sadness of it. And I’m certain almost everyone is going to be affected by the rapid onset on this technology. If you aren’t aware and looking into it, and you work in the knowledge or information service industries, you will be washed away.

Once you get beyond that, it’s pretty amazing the superhuman skills it can give you. But I’ll leave that for some of the forthcoming videos.

I’m aware that I am completely up against a wall

Companies have insane amounts of funding pouring talent into this sector. Many people are way ahead of me technically. Younger minds have more neuro-plasticity to adapt and less junk stored from the last 30 years of the internet. At 41 I am aware it’s a good age to start again as a new CEO, I’ve got experience and wisdom that I never had twenty years ago when I started a small web agency. But it’s still tough to constantly keep up to date, constantly reinvent yourself, and still find time to actually live life as its meant to be (i.e. not behind a computer screen).

I have no product market fit currently. I have some ideas but nothing approaching something that fixes a real world problem. I have only enough funding to afford to cover me up until October. And I have zero market advantage. I am basically starting from zero knowledgewise, I have very few major connections, no social media profile and not much codebase to go on.

And I’m up against the worlds best with all my weaknesses. So, it’s a big ask but screw it you only live once. I have given myself a hard deadline to build something within 7 months.