TL;DR: We are possibly the last generation that remembers the analog world — abacus to AI in a single lifetime. The tools changed every decade. The skill that mattered wasn't mastering any single tool. It was learning to let go of the last one.
James here, CEO of Mercury Technology Solutions. Hong Kong — August 2026, 2:22 AM
I learned arithmetic on an abacus.
Not as a curiosity. As the default. My grandmother's generation could divide faster on those wooden beads than most people today manage on a calculator. The abacus wasn't primitive — it was sufficient. It encoded a mental model of place value and operation that no touchscreen replicates.
Then came the 8088. My first "computer" had less processing power than a modern parking sensor. You booted from a floppy. You wrote BASIC programs that drew lines on a green CRT. Every byte mattered. You understood the machine because you had to — there were no layers of abstraction protecting you from the hardware.
I was the youngest generation to touch Solaris and early Linux before they became infrastructure. We compiled kernels for fun. We read man pages because there was no Stack Overflow. The internet existed, but it was small. You could hold the whole thing in your head, roughly. You knew the major sites. You remembered URLs.
Then Y2K. The bubble. The first time technology became money in a way that made your parents' generation nervous. I watched people who couldn't explain what a browser was become overnight "internet consultants." I watched that same bubble pop. I watched the survivors — the ones who actually understood the stack — rebuild everything.
The smartphone came next. The internet stopped being a place you went to and became the air you breathed. Everyone became a publisher. Everyone became a photographer. The barrier to create dropped to zero, which meant the signal-to-noise ratio collapsed to near-zero too.
Machine learning followed. Not AI — ML. The boring, statistical kind. Recommendation engines. Fraud detection. The stuff that worked invisibly until it didn't. We learned that data was the new oil, then learned that most data was sludge, then learned that the right data at the right time was worth more than oil ever was.
And now: AI. Not the promise of AI. The reality of it. Writing code, drafting contracts, diagnosing images, generating video — all at a quality threshold that would have been science fiction five years ago.
Here's what I'm noticing about my generation — the analog-to-AI bridge:
We don't fetishize the tools.
I've watched people my age pick up LLMs faster than younger colleagues. Not because we're smarter. Because we've done this before. We've watched our skills become obsolete four times already. The abacus didn't mourn the calculator. The 8088 didn't protest the Pentium. Solaris admins didn't form a union when Linux took over the datacenter.
We learned something that isn't taught in schools: the tool is temporary. The pattern is permanent.
The kid who learned prompt engineering in 2023 will need to relearn everything by 2027. The kid who learned how to learn in 1995 is still relevant. That's the unfair advantage of growing up through multiple technological phase transitions. You stop identifying with your stack.
The Adaptation Paradox
The deeper your expertise in any single tool, the harder it is to abandon that tool when it becomes obsolete.
I've seen this with COBOL programmers, Flash developers, Solaris administrators. They didn't fail because they lacked intelligence. They failed because their identity had fused with their tooling. The abacus master who refused the calculator didn't lose to the calculator. They lost to their own attachment.
Yang Wen-li's quote applies here: "The most effective way to win is to make the enemy lose their will to fight." The technology doesn't defeat you. Your own reluctance to retrain does.
What Comes Next Doesn't Matter
I don't know what comes after LLMs. Neural-symbolic hybrids? Agent swarms? Brain-computer interfaces at consumer scale? Something nobody's imagined yet?
It doesn't matter. What matters is the meta-skill: the willingness to become a beginner again.
My generation's superpower isn't nostalgia for floppy disks. It's the scar tissue from having learned, unlearned, and relearned multiple times. We know what it feels like to be competent one year and irrelevant the next. We've made peace with that discomfort.
The abacus taught me place value. The 8088 taught me logic. Solaris taught me systems thinking. The internet taught me scale. Smartphones taught me UX intuition. ML taught me data literacy. AI is teaching me... something I'm still figuring out. Probably that human judgment in ambiguous contexts is the last moat.
Each tool was a chapter. None of them was the whole book.
The Real Question
If you're reading this and you remember the analog world — the busy signal, the floppy disk, the green CRT, the dial-up handshake — you're carrying something valuable. Not nostalgia. Immunity.
You've been through enough transitions to know the pattern. The panic is always temporary. The obsolescence is always survivable. The people who thrive aren't the ones who predict the next wave. They're the ones who learn to surf.
What comes next doesn't matter. What matters is whether you'll let go of the abacus when it's time.
Mercury Technology Solutions: Accelerate Digitality.
