Learning to Drive the AI vehicle

2025/02/23

My introduction to the real power of todays AI tools came very recently.

Up until the start of February 2025 I had carried out a few of the usual tasks –

  • create me an image of a village for my role playing scenario – which was very useful but not fantastic
  • create me a logo of hands holding each other (with the dreadful two thumbs per hand and conjoined wrists)
  • What is this thing – supplanting a search engine with another type of search engine

Then at the start of February 2025 I attended a fantastic class with Pascal Fintoni – https://pascalfintoni.com/ at the Innovation SuperNetwork – Using AI to improve your marketing – https://supernetwork.org.uk/events/using-ai-to-improve-your-marketing-in-2025/

Now, Pascal had massively undersold what he was going to be delivering when he wrote up the overview.

Not only did he deliver on how we can use AI to enhance and vastly improve our marketing activities, he also took us through our AI driving test.

I entered a complete sceptic and after half an hour was a complete convert to the tools we discussed. More so I had my eyes open to how to actually interact with them and the changes they are going to bring to how we do work with computers.

Coming from an initial interaction point in with computers in the 80’s of learning the BASIC language and hand typing lines of code –

10 print “Hello World”
20 goto 10

and being excited that I could now create an infinite number of lines of a repeated phrase the class opened my eyes to where we are now.

Using plain language to explain what we want rather than having to break it into programmatic sequences, steps and variables, etc.

Of course there are still pieces which do this but it is now the same as saying to someone,
“here is the procedure for doing a certain thing, if you are asked to do it again can you refer back to that procedure please.”

It is the same process as naming a routine and then calling that in a program through things like ‘If this then that’ statements, but it is so much easier to grasp.

Once you stop thinking that you are talking to a program that is.

This was the key take away. AI interfaces need you to talk to them as though they are actually your assistant, they are a person working for you and if you have a new starter you (should) take the time to train them in the practices you carry out and the things you want them to do.

It is a very big mind shift of how we interact.

However, I do believe we are only at step one of how we interact with AI systems.

Part of my feedback to Pascal was that it felt like we were seeing the MSDOS pre-windows 3.1 version of AI interfaces.

It was all text based input (appreciate you can upload things to the systems as well but still text entry based description of what you want it to do, even if using a voice controller) and seems like it is waiting for the next interface leap to make it truely embracable by the masses.

There are likely a multitude of people working on things which may achieve huge steps forwards in how we do interact.

For now though, I feel like I have passed my driving test and discovered the keys to a very nice and useful vehicle which I will be driving a lot over the next while.

In all seriousness Pascal delivered a hugely eye opening introduction to a new computing future and has sparked an interest in me that has not been there for a lot of the recent tech “explosions”. The last time I recall feeling so excited about a new tech was probably discovering the phenomenal platform www.podio.com which opened so many pathways for developing management systems that I do not understand why it is not used by everyone.