I think it is almost 7 years since I started this nonsense of trying to get along in cyber space or whatever and the whole reason I did so was two-old. First I had just (self) published a novel I had spent 8 or 9 years writing (Raiding Cooley – still available on Amazon and other sites at the bargain price of $3.95, so go on, if you haven’t already, go ahead and take a look) and I thought I needed an ‘on-line presence’ to market the book.
So I started this blog thingy, initially on Celtic Iron Age stuff, and then it morphed into new things I was learning – Curves, Travel, Music, Food and that sort of thing.
Anyway, here I am now, years later and I can honestly say that my involvement with cyber space has not taken off. I have never used – or even know what they are / do – Tik-Tok, Instagram, …. Snapchat? And, well , others like that, thank God about which I know nothing.
But all of that is now in the past because with one of these smart AI chatbots, the world is there for the picking. Just recently a bunch of pointy heads and Elon Musk wrote an open letter (to whom, I don’t know) calling for an immediate halt to AI until firmer / better regulations were in place to govern it’s burgeoning use. And recently, a major news outlet – The BBC ? – published results where an AI chatbot successfully passed the country’s Bar examinations – the entrance into the UK’s legal profession. The Economist magazine mentioned a 4% possibility of AI damaging humans by 20.. – oh, I forget the date mentioned but the future seems to be crowding in and who knows what A.I. is doing to Man’s more brutal pastimes – war, savagery, intelligent drones and so on.
Anyway, rather bravely, I though, I dipped a very tremulous toe into the murky water of AI. The website seemed keen to ‘teach me’ how to do some fairly simple procedures but after step two or so when It asked me to ‘clone a repository” I gave up and jumped to an end product which just looked – rather disappointingly, I thought – like an ordinary google query point. The website provided dozens of examples of what it could achieve if prompted and drawing upon millions of everyday interactions stored over the last several decades, it could produce a tailored, ordered examples of text with an input similar to this ‘Write a two sentence horror story with the word breakfast’ and bingo, nano seconds later it was there and … Wow, very good, I thought.
How I will eventually use something like that remains open – given my past record with FaceBook, Twitter and LinkedIn – but it might be fun to play around.
Anyway, for those who are interested
will take you to a basic starting point where there are loads of incredible examples – including the two sentence horror story mentioned above and others as well as tutorials on Building applications (!), Embeddings (?), Image generation and Text Completion modules.
However, if like me, you want to skip all the bells and whistles and just want to plug a prompt into OpenChat and get an immediate response go to
Obviously once A.I. has reached this level, there is no putting it back in its Pandoran box. Just as Atomic and Nuclear knowledge cannot be unlearned, so too A.I. has probably escaped into the wider world and will continue to learn and grow. Is the old Arnold Schwartznegger Terminator films, where an A.I. intelligence, Skynet, determines that humans are a danger to its continued existence and proceeds to exterminate them, one of those examples where fiction blends into reality? Certainly, academics and pointy heads around the world are all scratching their collective noggins as to the dangers and benefits but one thing is for sure, A.I. is not going away and is going to learn exponentially.