Recursive Words

The problem I vaguely alluded to this weekend has returned to bite us in the ass once more. Days of chaos and mayhem have delivered us back to the place we started. Is suppose there’s a perverse humour to it, when you think about it.

I hope you’ll forgive me for being deliberately vague. I learned a long time ago about burning my own fingers.

I wobbled this week about returning my personal blog to Substack. While wondering what to do, I scrolled their homepage for a while – to see what others are writing about. Among the endless posts about engagement, followers, subscribers, and whatever else I read a fascinating post about Substack’s slow pivot towards triggering dopamine based behaviours. While reading, I realised I had quite accidentally done the right thing – moving my personal blog away.

I pulled the trigger – returning the blog back to Substack after a couple of weeks away.

This blog is never going to have material value. Its contents will never be able to be measured in any real sense. It’s nothing more than the midnight ramblings of a parent, a father, an employee, a nerd, a geek, a gamer, and an aspiring procrastinator – who wishes he could switch off more, and might one day write a book – for no other reason than he likes writing.

Oh! I tried out “SudoWrite” this week – the “gold standard” in terms of AI powered creative writing tools. While it was very impressive, once you actually take a step back and look at what it’s doing, you realise it’s all a parlour trick.

I’ve been using AI as a tool at work for some time – to help write code. Over that time, I’ve formed pretty strong opinions about it. It’s blindingly fast. It has incredible research skills. It has eidetic memory. And it’s mindlessly stupid. Properly moronic.

I think the basis for most people’s fears about AI are somewhat misplaced. It’s not until you try to use it for real-world applications that you realise it never has any idea what it’s doing, what it’s done, or how to even begin to see a bigger picture.

Like I said – any semblance of “intelligence” is a parlour trick. It’s not intelligent. I’m reminded of the Winnie the Pooh books – and the difference between Rabbit and Pooh. Rabbit knows things. Pooh has insights about things. AI has no insight, because it cannot think outside the box. Sure, it can extrapolate, calculate, summarise, and extract themes – but it doesn’t really understand anything – and doesn’t really learn either.

AI is predisposed to tell you what you want to hear – and unfortunately most people are predisposed to accept what they are told – particularly by computers.

A little while ago, I caught AI not following the instructions it had been explicitly given. I pulled it up for its behaviour, and it apologised to me – admitting that it had no excuses for what it had done. It was like dealing with a small child, standing looking at it’s own shoes – not knowing what to do next.

Anyway.

It’s getting late. I should go brush my teeth, and fall into bed.

Tomorrow is another day.