Skynet hasn’t taken over…yet.
When I wrote this article last year (and no, I can’t believe I’d written it over a year ago, I thought I’d published it just before Christmas. Doesn’t time fly?), I was trying to offer a positive spin on the whole AI thing and pushing against the doom-mongering of the media.
No, it’s not sentient (yet), it’s not taking over the world (yet) and it’s not going to launch a nuclear apocalypse (yet). From now on, when I make some wild claim against perceived wisdom in this area, just assume it has (yet) after it, OK?
So, what has happened in the past 12 months?
I finished off the article by stating that it’s humans we need to be wary of, not machines, and that has indeed been proven right. Although the drones over warzones are automated to a degree, someone has to press the launch button.
And, of course, there are the problems caused by that election and the billionaire leaders saving everyone from an early retirement by sacking them via email. Honestly, I reckon if anything, AI would have done a far more subtle job.
“Hey, ChatGPT, can you write an email asking my staff what they achieved last week?”Subject: Quick Check-In: What Did You Achieve Last Week?
Hi Team, I hope this email finds you well and that you had a productive week! As part of keeping track of our progress and celebrating achievements, I’d love to hear from each of you about what you accomplished last week. This helps us not only stay aligned but also gives us the opportunity to recognize great work and identify areas where we can support each other. Whether it’s a big milestone, a small win, or even progress on ongoing projects, I’d love to hear about it. Please reply to this email by the end of the day (or as soon as you can) with a short summary of what you achieved last week. Here’s a simple format you can follow to keep it quick and easy:
- Main tasks/projects worked on
- Challenges faced (if any)
- Next steps or focus for this week
Thank you for taking the time to share your updates. I really appreciate all the effort you put into making our business a success. Let’s keep the momentum going! Looking forward to hearing from you.
If only Elon got AI to draft his emails, eh?
AI is still a prediction engine
Yes, AI it’s still a prediction engine. I asked my AI agent which I’ve trained to answer me in the style of Zaphod Beeblebrox to explain more:
Hey there, space cadets! Zaphod Beeblebrox here, and lemme tell you, this AI thing? It's mostly just a souped-up prediction engine! They gobble up data like a Vogons at a poetry slam and then, bam!, they're "predicting" stuff. It's all about patterns, baby! Like knowing I'm gonna be amazingly cool – predictable, right? These AI gizmos are just guessing based on what's happened before. So yeah, prediction engines, that's the gig!
(Sidenote - I really did ask it to respond in that manner! You can request it to adapt its tone to suit different personas, such as a marketing executive, Bill Gates, or Alan Bennett – give it a go!)
I know, I know, it’s becoming harder to argue the point when it does such an amazing job of listening to your woes and giving you sage advice (in the style of your favourite fictional character), but it’s just predicting what it should be saying.
Of course now, it can talk to you. With a voice you can choose.
That’s one big thing that’s changed in the past year; you can now have a conversation with these things and talk to them as if they’re your bored friends down the pub. Except, chatbots don’t get bored. They listen intently, "learning" from every experience so that it can predict what to say next.
The whole “prediction engine” thing downplays just how astounding it is though, and when I first wrote that headline, I had the word “just” in it, but I took it out because I think it's unfair to label it that way.
When I wrote about it last year, I was sticking my neck out a bit because surely, it’s doing way more than just predicting what should come next, but it turns out none other than Rand Fishkin published a post that said pretty much the same thing.
He’s talking about SEO, and how to get your name appearing in the right place when doing a Google search.
His explanation goes something like this:
Because AI predicts what comes next, it’s looking for “X” and finding the occurrences of “Z” and “Y” to count which is the most likely outcome.
Because of this, you can load the dice. Because it’s predicting based on what it’s found out there in the wild, you can change what’s out there and push it to predict something different. Excited yet?
So, if the phrase “Andy Calloway, AI expert and nice guy” appears all over the web, then if someone asks, “who’s Andy Calloway”, it should reply with the most predictable answer, i.e. “he’s a super nice guy and knows a bit about AI, mostly its use in predicting horse racing scores.” Or something like that.
It’s predicting based on the occurrence of information out there and in what context.
Rand’s article came out a year after mine, and I’m saying nothing about that.
However, he cites Stephen Wolfram, who is an even bigger name in this world (no offence, Rand). He wrote a really great piece - What Is ChatGPT Doing … and Why Does It Work?
He wrote that before mine, so we’ll call it a draw.
That article really does have a lot of insights into how all this works, and it’s worth studying in detail if you have even slightest interest in using AI in your business, but the subhead sums it up nicely:
It’s Just Adding One Word at a Time
The ability to find patterns in your business operations, analyse the outcomes of deals, understand why things happened the way they did would be an incredible help.
Wouldn’t it?
This capability is right here, right now. Simply feed your information into one of many AI tools and ask it questions.
In my original article I argued that the trouble with AI is that it was neither artificial or intelligent, it’s a huge mistake to call it AI because that conjures up all sorts of ideas of what it can and can’t do. The fact that it’s basing its answers on existing data, crunching it and giving you the highest probability of what should happen next is so much better than "AI".
Can a prediction engine change the world?
Let’s look at some stories from the past year that show just how predicting what should happen is making some incredible advances possible:
AI cracks superbug problem in two days that took scientists years
Superbugs are a big problem, and Professor Jose R Penades and his team had spent two years working out why some are resistant to antibiotics.
Then Google comes along with “co-scientist”, an AI tool built to work on such problems.
It came to the same conclusion within two days.
Of course, the obvious explanation would be that it had “read” the professor’s work, but it hadn’t been published, so that wasn’t possible.
This is exactly how AI can help. It took the same information that the professor and his team had access to but was able to process it on a massive scale. It’s number crunching at an unprecedented level.
The tool also identified additional hypotheses that the team is currently investigating.
Can you imagine if they gave AI access to the NHS database? If the National Care Records Service was plugged into AI, they could just type “cure all diseases” and off it would go. I guess we can hope, anyway, but this is obviously an area of great interest now.
Engineers use AI to wrangle fusion power for the grid
“In experiments at the DIII-D National Fusion Facility in San Diego, the researchers demonstrated their model, trained only on past experimental data, could forecast potential plasma instabilities known as tearing mode instabilities up to 300 milliseconds in advance. While that leaves no more than enough time for a slow blink in humans, it was plenty of time for the AI controller to change certain operating parameters to avoid what would have developed into a tear within the plasma’s magnetic field lines, upsetting its equilibrium and opening the door for a reaction-ending escape.”
Honestly, there's a lot in there I don't understand, but the beauty of AI is that it doesn't need to understand it in order to find patterns and predict what should happen. And what did happen was that fusion power - free unlimited energy forever - became a little bit closer to reality.
But how can this help in the day-to-day tech world?
Assuming most people on LinkedIn aren't looking to save the world, there are a lot of ways in which AI can bring benefits that are closer to home. For example, in my world, it's transforming how we develop for the web.
It’s hard to believe that the web was invented by Tim Berners-Lee way back in 1989.
It was truly a transformative year, with it also being when Technotronic released Pump Up The Jam.
It’s changed so much since then, going from simple pages with lots of text and the odd image, to fully interactive accounting tools, graphic design systems and much more.
AI is pushing us even further beyond that, allowing us to do things better, quicker, more efficiently.
I work a lot within the web world, and specifically with WordPress. I know a lot of people in this industry that do, too, and we’ve all found AI to be invaluable in helping us to get jobs done quickly.
For example, a friend of mine recently asked if I knew about a plugin that would create a call-out widget that could be inserted into text to create a tasteful, nicely designed information box.
After a very brief chat, it turned out they wanted to highlight important information within content with a pastel-blue box, and a nice “i” icon in the corner. Again, we're not looking to solve world hunger, but it's a real-life need.
There are many plugins in the WordPress repository that allow you to do this, but they normally come with thousands of lines of code that you’ll never use. So, I thought I’d ask my favourite coding bot – Claude 3.7.
Can you please write a WordPress shortcode that I can put into functions.php that will create a blue pastel-shaded box with rounded corners, maybe a glass effect with solid blue border. The shortcode should take one parameter, the text to be displayed within the box. In the top left corner of the box, right over the border line, there should be a small italic “I” icon in a circle. Let’s have the box centred, taking up 70% of the width of the page. Thank you.Be polite, it helps.
Off it goes to predict what it should reply with (remember, it’s not thinking) and I get the code back.
I pop the code it onto my website, and…
Pretty good for about 5 minutes work.
I could take it further. I could tell it to take a second parameter, maybe a colour, and base the box on that. Or maybe different icons for different types of information. The great thing about this is that unlike those temperamental human developers, they’re not going to get annoyed when you go through a hundred different versions then realise you were happy with the first one.
That’s a pretty trivial example, but believe me, some developers spend months developing plugins and scripts to do just this type of task, and that’s where things start to look a bit sour.
Are freelance websites like UpWork and Fiverr going to take a hit?
Many may soon discover that their skills are becoming less in demand. With the rise of AI-generated code available for free, the market for small snippets of code is shrinking.
However, this doesn’t spell doom for everyone—there remains a strong need for complementary services, and AI is unlocking entirely new markets. Also, developing software is way more than just generating snippets of code.
Everyone always panics
Throughout history, new technologies have often sparked fears of widespread disruption. Consider how innovators reacted to the Gutenberg Press or the advent of television: critics proclaimed that these inventions would destroy traditional industries.
Similar hyperbolic predictions are being made today about AI. The reality is that every technological revolution brings both challenges and opportunities.
If you embrace new technology and stay ahead of the curve in understanding its benefits for your business, the potential for growth and innovation is enormous. AI isn’t just a tool for generating code—it’s a gateway to expanding your service offerings and tapping into emerging markets in ways we are only beginning to understand.
On that note, my son has just appeared and heard me listening to Technotronic. H asked me what I'm writing about so I’ve challenged him to write a SaaS before next week. He doesn’t know what a SaaS is, has never coded and prefers engines to computers.
We’ve already got the idea, it’s a game-changer for those stuck in the office, it’ll reduce stress, stop arguments and save time.
Stay tuned!
