Published: |
Walking vs Driving
We are six month into the November 2025 inflection point, from summarising it in 5 mins to Pope’s Magnifica humanitas, I’ve seen opinions and advices on all spectrum, but I’ve still not seen analogies where we can explain to everyone, why is AI here?
This started out as a LinkedIn post, here I’ve added more nuances to it.
Why is AI here?
From re-writing engines in a week, to automated legacy applications upgrades. The question is almost never “why is AI fast?”, but rather “how to drive AI safely?” And that is the key element to my analogy: walking vs driving.
The invention of motors in cars, trains, airplanes forever changed our existence. These machines accelerated our experience on this planet and beyond, it’s hard to imagine a life without them now.
But, despite the speed, convenience, and mass adoption of these machines that helps us move, we still reserve some experiences organically: a short walk to the supermarket, a stroll in the park, 5k running, a hike in the mountains…
If everything defaults to AI, that’s like saying let’s drive a car for hiking, or use a scooter for running.
We understand how efficient these machines can get us from A to B, or carry heavy loads, but they cannot replace every experience where we seemingly are trying to achieve the same goal.
What should we focus on?
To extend on this analogy, we do heavily rely on these machines, but we’ve also built the most extensive, regulated, and maintained infrastructure to support them.
Every driver must pass theory and practical tests, every vehicle was designed and built to follow rigorous safety standards, the aviation industry is the safest industry because it treats every failure, near-miss and minor technical anomaly as an opportunity to re-engineer the entire system.
You wouldn’t drive a car without a seatbelt, take a plane without a certified pilot, or even going into an airport if it didn’t have a functioning security checkpoint.
Without these multi-layered approach across engineering, culture, and regulations, we would not be able to enjoy the safety and convenience these machines bring to us.
These are the security, infrastructure, service layers you can build around the ecosystem. You can build an airport with little or no knowledge of aerospace engineering, or a car-sharing service without owning a car or driver.
The AI slops, they are air pollutions. Its damage and effects varies in degree, from individual to industry level. Some people, deemed they are unavoidable for growth. But without proper regulations and controls soon (at the time of writing, AI industry are “less regulated than sandwiches”), they will grow to be a significant risk factor and contribute to “climate change”, which may eventually be the bane of our digital exisitence. Wars on pollution are necessary to us physically and digitally.
Despite the mainstream headlines of “AI taking over jobs”, I believe there are still plenty job opportunities for humans, but they’ll look different from previous generations. Think of how farmers, nurses, teachers, lawyers, etc adapted to cars, internet, smartphones, social media, now there’s just another requirement. But some jobs will disappear, like how alarm clocks replaced knocker-uppers in the Industrial Revolution (what will be replaced in the AI Revolution?).
In the age of cars, don’t move on foot or on horses (if you do, make it a niche). Most people will transport by car or metro, but still keep their hobbies of running on the weekend. You can have both.
Notice the extensive integrations of software around us, and build emotional experience too. When Bob Baxley had the opportunity to redesign the Apple retail POS system, he had the image of an early-onset Parkinson’s employee who struggled to click the small buttons in mind. There are two bridges in San Francisco, both do great jobs at being a bridge. But only one bridge stands out, and it is the better designed one (it cost less too). Design better only requires more care, not necessarily more resources.
How will this impact us?
For the workers, they are expected to upskill their AI capabilities now, or in the near future. This is not caused by a singular demand, but a combination of demands from the business, customers, markets, and even governments.
Imagine you are a postal worker, in the old days, you would walk around the neighbourhood, delivering letters and parcels to the doorsteps, sometimes even stop to have a quick chat with the residents, some offered you tea and biscuits, and you helped a few as their handyman, search party, and friendly neighbour.
Fast forward to today, you are required to drive cars, trucks, scooters, onewheels, whichever machinery to get the job done fast. Time and efficiency are the key metrics. Your performance metrics were never “how many catchups did you do”, it is now “how many X delivered on Y” on steroids. Soon, you are required to learn drones too, because “customers enjoy their 2 hours delivery promise”.
Businesses that do not adopt the latest and fastest technologies will be out-competed, and you don’t want to lose your job. It may not be your choice which machinery to use at work, but it is your choice whether to walk or drive to the supermarket, or go running on the weekend. If work makes you sit all day, you have to do yoga or stretches whenever you can. If work makes you think less, you have to put down the Reels and pick up the books again…
For business and organisation decision-makers, they are expected to make decisions that prioritise machines or humans.
The trend of “tokenmaxxing” or “agentic-madness” focuses on how to make the machines “smarter”. While we are busy building context knowledge base or guardrails, we are focusing on improving the efficiency of the systems.
But systems are often short-lived compared to humans. What happens when the systems goes down, affected by acquisitions, or its goals were too rigid?
The modern Amercian cities are centered around cars and highways, and many years later, people started to realise the European walk-able cities are healthier and more community-friendly. Unfortunately, the infrastructure and culture are too well established, undoing or changing the existing landscape will feel like an impossible task.
Often, the friction is where the magic happens. Don’t think “cars can replace marathon runners”, the same applies to AI and people.
The surface goal was never the point. It’s not the distance, nor the productivity, nor the cost. Maybe the untapped potentials are unquantifiable and unknown, like the “freedom”, “aspiration”, “satisfaction”, “belonging”, or any other words you shall find in a minimalistic marketing poster to make you wonder what they are selling.
For a more practical guide on slowing down with AI, see #HUMANTOODO