A gentle ramble into the foundations of AI
I went for a walk the weekend before last, and I find it hard to resist the indescribable joy of
Shabitat, which has loads of pre-loved furniture (most is “shabby” and some is also “tatty”).
But, surprise, surprise, I don’t go for the furniture – spoiler alert – it also has a book section.
And you can have whatever book you want for a small donation. Now, I confess, “I’m a
reader!”
Not sure whether you’ve seen Bill Hick’s routine from the Relentless video called “What are
you reading for”? It’s on YouTube if you want to watch it. In typical Bill Hicks’ edgy style, and
really funny.
Anyhow, back to leafing through the shelves of books arranged higgledy piggledy, in piles
and crammed into overflowing shelves. I had a quick scan and saw a few books that I
already have, resisting the temptation to buy a second copy to give away to someone who
would enjoy it…
Then I glimpsed a really thick book sitting on top of the tallest unit of shelving, with the words
Artificial Intelligence – A Modern Approach – visible along the spine. Of course Brighton Uni is
just down the road so maybe a text book discarded by the next Silicon Valley tech start up
founder. At least that’s what I’m imagining/hallucinating.
Turns out “modern” refers to 2003, pre-historic by AI standards, you may think…
So I donated 50p and went off to a spot I often visit, in the sunshine, and sat down at a
wooden bench and table to have a good old read.
The book is by Russell and Norvig, Professors from Southampton University, if I recall
correctly. They trace the origins of AI to:
● Philosophy – explores the nature of knowledge, thought and learning
● Mathematics – underpins formal logic, algorithms, computational complexity and
probability
● Neuroscience – investigates how the human brain functions
● Psychology (sometimes framed as cognitive psychology) – examines how humans
and animals think and act
● Linguistics – analyses language in a form that allows computational processing
● Computer Engineering – concerns how to build efficient computing hardware
● Control Theory and Cybernetics – studies how autonomous systems can function
and regulate themselves
● Economics – explores decision-making by agents, often under uncertainty and with
competing interests
“These eight disciplines collectively shaped the development of AI, blending insights about
human cognition, formal reasoning, computation, language, and system design into one
unified field.” ChatGPT informs me.
Surprise, Surprise – it’s not as new as I imagined
I’ve had an interest in most of these disciplines (or fields of study) since university when I did
half a degree in Economics back in the Dark Ages of 1978-1981 although I don’t remember
any of the lecturers mentioning Game Theory By John Von Neumann and Oscar
Morgenstern, which was written in the 1944, and is one of the antecedents of AI according to
Russell and Norvig.
And Philosophy, another one of my passions since I read Zen and the Art of Motorcycle
Maintenance as a teenager, has been around for 3,000 years or more…
And Neuroscience, which when applied to leadership, is fascinating, and one of our clients is
writing a book on.
Now, I know this may sound rather strange, but this reassured me and I started to get much
more enthusiastic about AI, armed with this new insight on where it all came from. Turns out
AI did not emerge fully formed from the K-hole to confirm Mrs Musk’s prediction about her
progeny’s incipient genius.
The First Computer and The First Programmer seem to have been Charles Babbage and
Ada Lovelace in the 1800’s according to yet another book called The Creativity Code: How
Ai is Learning to Write, Paint and Think by Marcus Du Sautoy, that I managed to locate in
the Oxfam shop in Lewes on a field trip exploring rural Sussex.
Ok, enough with the history lesson, and the theory – what about the practical stuff?
What is AI good for?
My AI Bootcamp ended on Monday, and I have submitted my 5 Projects to prove that I’ve learned something!
So far I’m focusing on 3 tracks – Process Automation, Content and Creative and
Analytics/Reporting and so far all 3 seem promising. I’m creating my own custom GPTs and
I even boldly shared 1 with a client earlier this week.
The results are interesting, as long as you think carefully about the prompts, and so far it
seems to be good for Research and for automating repetitive tasks, and for coming up with
whacky pictures (you’ve probably seen more than a few of those on LinkedIn)..
I know that AI can be like drinking from a firehose, and many businesses may have one or
two false starts, so I asked ChatGPT…
Acting as an expert in AI for Small Business Owners, what are the main challenges that this
group faces in adopting AI?
And it even generated a checklist.
I’ll probably share that once I’ve checked it.
And what is AI not good for?
You can have lots of fun with automated telephone diallers, which are pretty obviously
recorded, although they can catch you out if you’re unprepared. Try asking them a question,
it’s hilarious. And I had to apologise to an actual human on one of the main AI provider’s
support chats the other day when I noticed that the answers were repetitive and incorrectly
assumed that I was typing text into a bot.
Which makes me wonder, do we still need humans or will the robots take over, as predicted
in much of the science fiction that I was reading in the noughties.
Maybe they will, maybe they won’t but, as things stand at the moment, AI still has lots to
learn.
I’m focusing on collaboration and being nice to ChatGPT, using please and thank you, just in
case. And I continue to believe in the Golden Rule, even when chatting with AI.
But, would I get into a driverless car in the foreseeable future?
Maybe I can ask ChatGPT to decide for me?
Not on your nelly!
Feel free to drop into our office at 71, St James’s Street if you want to ponder how AI might work for you or drop me an email graham@digitalfreelance.co.uk.
P.S. Can you guess whether the photo was generated by AI (or not)?
