AI is in everything now! I love it as a tool and it is increasingly doing incredible things. But does it need to be in that little brochure site for a new building estate. Really, is that increasing engagement? Maybe it is and I am being cynical. The worst i’ve seen bar far, is AI in a pen.

I sort of mildly want to complain. And that is as someone who loves using AI every day in my work but is jaded by the amount of “AI can just do that for you”. “Or do you want to build the next best app and earn a million pounds? here is an AI builder”. But anyway, after the complaining I want to talk about some of the realities of running AI and where it starts to really add up.

The current landscape

So currently, as I very wittily pointed out in the intro, AI is in everything! Everyone is an AI developer and works with AI including myself. But what grates me on this a little bit is it’s not entirely true. You leverage AI, I leverage AI. I go to one of the big boys, the actual AI developers, and I say hey can you pass this to your massive language model or coding model or whatever and currently either free or for a very reduced cost. They go sure no problem. Their model does a big “think” then they pass me back the response. So far so good.

But this brings me onto the fact AI is in everything. So that AI Pen I mentioned is very likely not processing AI natively within its own electronics it is either connecting to the internet and sending the request. Or it is connecting to your phone to piggyback on its conection to send the request. So my cynical view here is no the pen doesn’t have AI, it just has the ability to send data requests.

Then onto AI companies themselves the big boys, Anthropic, OpenAI and Google (Gemini). Now effectively these guys currently have a licence to print money. And I think as of writing none have managed to hint at turning a profit (as of writing).

You Have No Idea How Screwed OpenAI Actually Is AI Companies don’t have a profitable business model does that matter

Changes on the horizon

I have a few thoughts then on the future at least of AI in its current form, pending any new major breakthrough which is always a possibility (really trying to protect myself from looking dumb here in the future). But there are a few things in the works and currently that is AI as a UI and also the introduction of adverts into AI.

Now the AI as a UI I think is interesting and lends itself to that Tony Stark style I just talk to a nondescript voice somewhere and cool stuff happens. What I do think is missing with AI as UI is the generational gap. It’s possible my daughter currently 1 might grow up in a world where you just talk to all machines and they work, or type to them like a chatbot. But currently I foresee most people finding this quite cringe outside of the home. Putting those people that talk really loudly on public transport to one side.

Conclusion

So ultimately AI is going nowhere it is here to stay. And I personally think that’s great, it streamlines loads of interesting things and gives me the ability to increase my efficiency by 10x which is not insignificant.

However I do foresee the overall “put AI in and on everything” bubble to collapse. The way the current organisations are being run and looking like they are going to change their business models suggests that beloved “free tier” or even the “cheap tier” is probably going to come to an end. I foresee lots of chatbots on websites and integrations suddenly ceasing to exist. It reminds me of when Google decided to change their Google Maps API and make it a paid-for-service. Whereby using it without a paid for key adds the development only watermark everywhere and reduces functionality. I still see this on websites to this day, which brings me back to a period of adjustment coming.

I will still continue to use and pay for AI but I am a big proponent for it being a fantastic and useful tool. It is not this magical wizard that will automatically make your application you don’t understand work or design you the next big Amazon / Google / Facebook competitor. The main point that proves this is: if it could, why hasn’t everyone done it. I know I have tried, for testing sake, sorry - it’s great and takes a lot of the heavy lifting. But also makes mistakes, misunderstands and has no context for greater functionality.

The main takeaway is currently AI is a tool and one I think will start to cost more and more as time goes on.