Edition 3
“Genuine communication can only occur between equals”
- Claude Shannon
Hello all,
Well, here we are again another week has passed in the real world which we now know is around six months in AI years.
To give you some idea of the kinetic freneticism of it all you may recall I spent the last two issues evangelising the idea of prompts and prompting, well after a long fortnight of glorification I’m happy to announce that prompting is a dead end (and we’ve always been at war with Eurasia).
Not that you’d know that prompting is no longer the Holy Grail as those ubiquitous offers of “957 free AI prompts guaranteed to make you an Archduke by next Thursday” seem to have sprung up all over social media, which is usually a sure sign the thing on offer is a royal flush of bustedness.
The issue being that prompts like these offer a one-size-fits-nobody solution which is completely at odds with how AI works. Your average LLM (Large Language Model) AI such as GPT4 and Claude 2 are designed to rapidly assimilate and then accurately reflect your very specific wants and needs.
It’s as if AI is offering you a personalised exquisite Saville Row tailoring service and the prompt mills are hawking you one of those voluminous kaftans beloved of Dennis Roussos. Sure, it will do the job of covering your modesty and possibly save your life as an emergency parachute if you’re ever unlucky enough to find yourself thrown out of an aeroplane but other than that it means you’re just lugging around a weighty billowing sail cloth good for trapping fresh air but little else.
My freshly minted take this week is that getting a genuinely useful grip on AI is (apologies to Swiss Tony) a bit like learning to drive. You’re going to have to commit to between ten- and fifteen-hours simply sitting at the AI wheel before you start to feel comfortable with the instrumentation and gaining an ability to spot major hazards (such as AI’s capacity to hallucinate, more on this next week) whilst learning to accurately read the road. You can of course memorise your Highway Code and do online hazard testing, but nothing is going to take the place of getting in the car and going for a drive.
Obviously in this metaphor I’m your friendly, trendy, avuncular-but-effective driving instructor ready to tap my clipboard on the dashboard.
On a much brighter note, I did hear a lovely apt metaphor for AI this week about the piano being invented by people who had no idea what sublime levels of music would go on to be created upon it.
Attached is a graph from the behemoth McKinsey that seems to bear out most of the findings from the AI Job Killer. Please note I gave it to you for free whereas McKinsey would have charged you a pretty penny for such insight:
This week I tried a local Facebook post (more fool me) to talk about AI and got to experience what a Roman engineer must have felt like trying to explain the concept of an aqueduct to the man who emptied Boudicca’s chamber pots. I saw the irony of people using Facebook, via the internet, on their smartphones belittle AI, but who without a doubt once had the same skepticism for Facebook, the internet, and smart phones.
Last week I mentioned Open AI haemorrhaging cash, this week came this article from The Information: OpenAI Passes $1 Billion Revenue Pace as Big Companies Boost AI Spending
To quote from the article: “Open AI is currently on pace to generate more than $1 billion in revenue over the next 12 months from the sale of artificial intelligence software and the computing capacity that powers it. That’s far ahead of revenue projections the company previously shared with its shareholders, according to a person with direct knowledge of the situation.
The billion-dollar revenue figure implies that the Microsoft-backed company, which was valued on paper at $27 billion when investors bought stock from existing shareholders earlier this year, is generating more than $80 million in revenue per month. Open AI generated just $28 million in revenue last year before it started charging for its ground-breaking chatbot, ChatGPT.”
All of which only reinforces last week’s quote about ‘if you think you know what’s going on…
I’m very aware I have yet to broach the issues of AI ethics, a shame because the so called Gorilla Problem of AI is a doozy but it will have to wait as this week I’ve been keen to launch AI-M the programme I’ve meticulously developed for fully onboarding any business, and any one, into the world of AI.
Do get in touch to find out more.
In other news as if AI wasn’t enough of a bundle of frenetic, maddening beauty, we’ve just welcomed a golden doodle puppy into our home, luckily for now she and the AI are getting along in the time honoured fashion of completely ignoring one another...wonder how long that will last?
Ok off with you, I’m sure at least one of your helicopter’s needs some new fluffy dice or some such.
Vincent