My A.I Mistake

// The Thought

Everyone uses A.I for shortcuts right?

This week, we (Yammayap) kicked off a month of 'events' that we're either attending or arranging to help other businesses understand a little more about what we can do and how we can help and it all started with an event organised by the South West Manufacturing Advisory Service (SWMAS) hosted by Bournemouth University and the Arts University Bournemouth - the event was to promote a 'fund' for manufacturers to help move towards a 'Smarter Future'.

Manufacturers are notorious of leaving things as they are with their tech as they'll do a job for so long, the sales will grow and suddenly, things have become legacy and then, everything doesn't work - it's what we're fighting against but there is so much work in getting people to understand that they don't need A.I - they don't need to put all their money in an 'off the shelf' solution that is designed for everyone..

But anyway, I'm going into pitch mode.

The irony of this piece of writing isn't lost on me - I was attending an event to move people away from the conversation around A.I and here I am, out here using A.I to 'help me'.

// The Reliance

I'm not addicted to using ChatGPT but it is becoming my new 'Google' - I use ChatGPT daily and I use it to streamline/speed up the work I carry out - it's become a crucial tool in helping me deliver more with the same amount of time. I never use ChatGPT to write this newsletter, I made that conscious decision a while back as I never wanted it to replace my creative process and I wanted to keep my 'eye in' when it came to writing this newsletter.

Are there times when I don't have to use ChatGPT but I do out of 'habit'? Yes.

Are there times when I should just use my own judgement and write something without the aid of A.I? Definitely.

But it's so easy - I've learned though, what I can use it for and what I can't - it's a great tool for summarising large sets of data and using this to help with my work.

I can see how it's going to level the playing field in some regards - the 'creative' output of ChatGPT is awful - I've played around with so many functions that I can see how much I could use it for but never would.

For example, when designing a logo - I've used it to kick out some real basic concepts to riff on but I'd never use it to create a final version - it's just not capable of the nuances of moving small areas of a logo or creating something with thought - it's bland.

If I need to conduct bland work - thats what I use it for.

// The Mistake

The time available every single day reduces with every task added and there is only a finite amount available - there are things that have to be shortcutted and sometimes, I rely on my 'talents' to wing it - what I mean by that is, if I have been able to do something without prep before, I'll often be less likely to prepare as well as I should because I can get away with it.

Presenting or being stood up in front of many people and talking is one of those 'talents' - I don't get stage-fright, I like getting up and presenting and I will happily stand up and do these things without much thought.

There are very few occasions where I have fluffed my lines, not come away with a sense of doing a 'good job' because when you're stood up at the front, no-one knows what you're going to say so when you lose your way a bit, people tend not to notice because they didn't know that it wasn't a part of the prep in the first place.

Except people must have known with my 'presentation' last week.

I fluffed it a little.

Relying on my 'talents' didn't work - I didn't prep and it showed.

// What I Did (or didn't) Do

I've delivered this presentation before - it was a presentation that I'd given a few months before and it went ok before - I'd managed to rescue the fact that the presentation was a little light by talking around the main points on the screen (which had been written by others but with input from me) - the points being displayed were sensible when I did the presentation before.

This time around, I wanted to refine - get rid of the parts that didn't work so well before and add in some words around what I needed to talk about - so I uploaded the whole thing to ChatGPT.

I told it what I wanted it to do (without much input on what I wanted to be honest) and I asked it to refine it.

It did what it was told.

Now, I did look at what it delivered because I typed the words into my Powerpoint and I created (not ChatGPT) handwritten notes around the points I wanted to discuss, which I then typed up.

But I did this on the back of someone 'elses' thinking - which is where I think the mistake comes in. The thought didn't originate from me, it wasn't what I had thought about and then created so I couldn't, when I stood up in front of people, confidently deliver a presentation around it.

I got up there and my mind went blank.

The structure - the underlying narrative on what I needed to talk about, just wasn't there.

It struck me the moment I stood up - I didn't know what I'd written and because the forethought (or at least a little thought) hadn't gone into those words, I was suddenly up a creek without a paddle and I needed to get to safer land.

The best thing to do in these situations is to talk about the things you do know and that's what I did but I didn't deliver what I wanted to deliver.

// My Point

The thinking that you put into something is 80% of what you produce.

You can't expect A.I to create the idea and you polish the turd - if you do, you're negating the 80% of the work that you'll need to put in.

I made this mistake last week - I was expecting a robot to come up with the idea, structure it in a way that made sense and then implant that into my brain so that it made sense to me - it didn't.

I tried to shortcut the process and I failed.

Now, I'm not sure that other people in the room knew this - they might have seen a confident presenter up there because I still projected, I still made eye contact with the people in the room and I still moved around like I knew what I was talking about but for my own confidence, I took a hit.

I didn't do my best and that's what stuck with me after I sat down - I immediately began scribbling things down to improve my next presentation and get my thoughts out to make it better.. and thats the moment I realised my mistake.

When I was scribbling - its at that point that you're slamming the clay onto the table and you start shaping it - I didn't do that before. I just handed the clay to someone else, it handed it back to me in a shape and then I had to come up with all the thought behind the shape and tell people about it.

It's very hard to do that.

// The Worry

I'm worried about how the next generation will become reliant on ChatGPT or other LLM's - the lack of thought or mental hardship that they'll have to go through will be the decider - if they don't create something themselves then they won't understand it.

I mean, I've been through failed presentations before and I've delivered match winning performances but without the baseline of what's good or bad, I couldn't have recognised that the one I gave last week was poor.

If I only ever deliver ChatGPT created content - how can I tell what is good enough? The answer - you can't.

You can't tell whether your work is good if you don't create it - you can't get the feedback from the room if you didn't put yourself there.

Tools like ChatGPT aren’t the enemy. They’re brilliant in the right hands and for the right jobs. But they’re not a substitute for thought.

That’s the crux of it - the hard graft of thinking is where the value lives. It’s the difference between knowing about something and actually understanding it. And when you shortcut that part, it shows - maybe not to everyone in the room, but definitely to yourself.

That’s why last week stung. It wasn’t that I bombed, it was that I knew I hadn’t put in the work where it mattered most.

So, when it comes to the future - I now understand that A.I can clear the rubble, but I still need to build the house. It'll help me with those large datasets and the manual sorting that happens but the clay has to be slammed down and shaped by me before I ask a tool to help smooth the edges - I need to have the thought first and use the shortcuts second.

Because if you skip the part where you wrestle with the problem, you’ll never own the answer. And if you don’t own the answer, you can’t stand up in front of a room and deliver it with conviction.

That’s the balance I’m trying to strike. Use the tools, yes. But don’t hand them the steering wheel.