The quiet after

Amidst the hype of pessimistic doom-and-gloom mixed with ultra-optimism in the future of AI, I decided to read through my book notes again from ‘What Technology Wants’ by Kevin Kelly¹. I’m reminded that there exists a much more calm and objective perspective of the technology, and thought it could be helpful to write my thoughts down.

There will come a point in time when the hype dies down and things enter the ‘slope of enlightenment’. It’s a neverending cycle, and the prevailing narrative will always make the pendulum seem to swing farther than it actually does. But, it will eventually become invisible to us and we’ll treat it like a natural extension of our consciousness. If you are feeling a bit of anxiety over this topic and the seeming uncertainty in the future, I hope this article brings some value to you as we look into what happens in ’the quiet after’—a future where people completely take it for granted.

What is technology?

Let us first examine what technology does. It serves primarily to open up new opportunities. Our ability to make warm clothing opened up possibilities that were closed to us before—the arctic regions. Harsh weather and climates are barriers, but also stand guard to a treasure trove of information. Winter clothing gave us access to those regions, unlocking its secrets and resources.

The word ’technology’ nowadays is almost synonymous with software. It almost always conjures up images of the digital world that we live in, thanks to the mega-cap companies that have captured the entire world’s imaginations. Technology is so much more than that. Let’s take a step back.

Take double-entry bookkeeping for example. It was first invented in 1494 by a Franciscan monk, and it enabled companies to monitor their cash flow and steer complex businesses. The fact that it initially manifested in the form of pen and paper does not take from away how revolutionary it once was. Today, it underpins our entire financial system and facilitates all the buying and selling in the world.

Now, let’s peer into the nature of AI.

Stochasticity

“the quality of lacking any predictable order or plan”

The golden child of AI today is the Large Language Model (LLM), whose true nature is rooted in stochasticity. That is both a strength and a weakness. For better or for worse, it almost seems like the entire world is distrustful of it from the get-go, and are desperately trying to force determinism into the picture. Just like you wouldn’t wear winter clothing all day in sunny Singapore, we should probably stop trying to stuff it into every nook and cranny.

If not everywhere, then where? Thankfully we already use some form of this randomness attribute in software development. Methods like Chaos Testing, Red Teaming and Bug Bounty programmes are in essence subjecting a system to unknown stressors. The goal is to improve overall resiliency and robustness of the system. We don’t try to make Chaos Testing more ordered or predictable, neither do we replace deterministic tests completely. It is used in conjunction, just like an additional tool.

AI is indubitably powered by electricity and exists as bits and bytes. Don’t be fooled by its current form which could change in the future. Keep its quality of “randomness” in mind and examine what AI can do for us with this strength.

Breaking invisible barriers

Our physical brain is incredibly powerful, but not without flaws. We are sometimes our own worst enemies because of cognitive biases, stereotypes and heuristics. I know what you’re thinking. AI can be used to reinforce cognitive bias. Sure, but that’s hardly a problem with the technology itself—you can amplify your bias just as well with search engines and joining like-minded communities on social media. People hear what they want to hear.

Cognitive biases are invisible to the naked eye and more insidious than the harsh climate of the arctic region. They block out entire regions of ideas and shrink down potential solution spaces. Worse, they make you think that you’re right and give you a huge blind spot. One of the ways AI can help then is by challenging assumptions in an unexpected and non-judgmental way, which is exactly what we need to overcome biases—free-association.

Free-association

AI itself cannot produce novel work. Heck, we cannot even rely on it to generate useful ideas. But what it can do is help encourage divergent thinking simply by slinging shit on the wall. The goal of using AI is to stir our imagination, not do the thinking for us. This enables more free-association that tends to lead to unusual combinations between ideas and ultimately has a better chance of yielding creative solutions. This concept is fundamental to Edward de Bono’s Parallel Thinking² technique as well.

Great design teams try to achieve an environment conducive for creative work by hiring and assembling a diverse pool of talent with varied experiences. The hope is to prevent groupthink and narrow-mindedness. Even so, humans are still susceptible to self-censorship. AI is less so.

One common complaint about AI is that it fundamentally can’t “think”. Well, isn’t that exactly what we need—a sort of brain that’s free of cognitive bias? I know, I know. LLMs are trained on a subset of information available in the world and have biases in the form of model weights. Focus on the technology itself, and not what the megacorps are offering us in the form of ChatGPT, Claude, Gemini or whatever flavour of the month is. Imagine a world where we train our own models and control its weights. It becomes just another tool, and all we have to do is to configure its parameters and wield it responsibly.

Shrinking the distance

Safe air travel shrunk the physical world. The internet compacted it even more. The steam engine mechanised locomotion and all sorts of machines in the industrial revolution. It was as if humans gained a new exoskeleton to impact the physical world and even move mountains. AI is like an ’exo-brain’, an external brain we can use to move much more rapidly in the information world. AI takes it one step further and brings domains even closer to ourselves.

Domain expertise can be challenging to access. To even find the right topic, area or person, you first need to know the right words and some form of initial knowledge to search effectively. If we don’t know what we don’t know, we can’t even begin to look for it. AI fills this gap by fleshing out the unknowns to get us started. In some sense, this brings domains closer within our reach, because many experts are busy spending their time working, honing their craft and practice. They’re not sitting around waiting to answer questions like LLMs.

One great example is Luke Wroblewski, an internationally recognised digital product leader with a wealth of experience designing and building software used by billions of people worldwide. You might try DMing a question to him on LinkedIn, but a busy man like him is unlikely to have time to answer you no matter how well-intentioned he may be. He recognised this problem, and trained his own chatbot on all of his articles so that people all around the world can tap into his experience and ask away for free. The results were massive. Luke claims that people ask his personal AI 40+ questions a day, where each question would have taken him 5 minutes to answer manually³.

My point is, there’s a better way than framing AI solely around productivity gains: we now have expert knowledge right at our doorstep, anytime, anywhere.

Overcoming psychological resistance

Getting started on something new is probably one of the hardest things to do, despite having all the technology and information in the world at our fingertips today. This cognitive barrier and initial resistance prevents many people from realising their ideas and imaginations. AI helps us get over this hump because of its generative powers. We all know what’s happening in the software world—AI is capable of vomiting out so much code at a faster rate than humans can.

Sometimes all we need is just a little prod to get something going, and that tiny push can be in effect a huge buff to enterpreneurs around the world. The effects of Vibe Coding is proof enough of its success. Many designers are now taking it upon themselves to create entire apps and businesses without hiring a single software developer. Sure, the generated code and systems can be insecure and need further expert review. But it did get ideas off the ground and over the line.

People look at this phenomenon and say, oh wow, AI can do all the coding for us. I think what’s more powerful is the ease that we can get over the initial hump more so than building the whole damn thing by itself. Theoretically, this should improve economic activity, encourage experimentation, increase rate of technological advancement and make new discoveries. Great for mankind.

A great blessing to the curious

Personally, the biggest gain for me is how it helps with learning things. AI is one of the greatest boons to ever happen to the curious. Autodidacts all over the world know this, and I’m sure we are enjoying the technology a whole lot more comparatively.

I recently tried to learn music theory on my own, alongside practising the piano again. Being a self-taught pianist, I focused on musical scales and learning pieces. I’ve never touched theory. I’m unfamiliar with a lot of musical terms. It’s hard to even form a proper question even if I have an expert on call because I don’t have the right words. In fact, I have so many dumb and stupid questions that only the most patient tutor invested in me as a person would bother answering. Conversing with a LLM helped me reach an amateurish level of understanding that’s just enough for me to get over that brick wall and continue learning.

What awaits us

Technology never reverses. It eventually become invisible and fades into the background. Can you imagine life without electricity or the internet today? We care more about bars of mobile network connectivity or Wi-Fi access than the actual internet itself. We’ve fully taken it for granted, and in time to come the same would happen to AI. I would make sure that by the time that happens, I’m already well-acquainted and have mastered the use of it.

AI is definitely over-hyped now, but I implore you to ignore the noise and understand how it opens up opportunities for you personally. Megacorps should not be the ones telling you how you can benefit from it or what the future looks like. Take control of it and harness the power of AI’s stochastic nature rather than fighting or fearing it.

“Can the human mind master what the human mind has made?"—French poet and philosopher, Paul Valery