On January 1st, 2020, I published “30 Predictions for 2030” — a long list of unfalsifiable public bets about how the world would evolve over a decade. I framed it as a scorecard for my own thinking: a way to expose my assumptions, invite disagreement, and have something concrete to revisit years later.
We’re now halfway through that decade.
Five years is not enough to close the book, but it is enough to measure direction, momentum, and magnitude. Enough to separate noise from signal. Enough to see where my 2020 worldview was sharp, where it was naive, and where reality surprised all of us.
What follows is my mid-decade audit — a try at an honest reflection on how the world unfolded versus what I expected, and what this tells us about the next five years.
Note : this was 50% written with AI, but the thinking is (still) human.
1. Climate, Geopolitics and Global Power (P1–P9)
Climate as a structuring force — right, but too gentle.
In 2020 I argued climate would become the gravitational field around which energy, migration, conflict and economics would orbit. This was correct, but I underestimated one thing: the speed. Climate shocks moved faster than institutions, and the biggest accelerators were not moral awakenings but price signals and risk exposure in supply chains.
China becoming #1 — wrong on timing.
The math still points up, but the demographic cliff, real-estate implosion and domestic slowdown delayed the crossing point. My logic was sound; my calendar was optimistic.
India rising — strongly correct.
India’s ascent is one of the defining arcs of the decade: talent density, infrastructure leapfrogging, and political will converging at scale. If anything, I under-predicted its momentum.
Africa — too early to call.
I wrote that Africa would be a high-variance story with two possible futures. Mid-decade, both are still unfolding in parallel.
The US declines — completely wrong so far.
This is one of the predictions that aged the worst. I underestimated the compounding effect of AI infrastructure, capital markets, and talent attraction. The United States re-centered global power through technology rather than politics. This was a blind spot.
Europe crumbles — unfortunately accurate.
Not through collapse, but through drift: institutional inertia, regulatory overreach, strategic ambiguity. Europe is not falling; it is dissolving around the edges.
A bloodier decade — tragically correct.
I expected more proxy conflicts. Reality exceeded the prediction.
2. Markets, Institutions and the Economic Engine (P10–P20)
Innovators outperform incumbents — correct.
AI-native challengers are outpacing incumbents across every layer of the stack. Some incumbents fought back harder than expected, but the structural asymmetry — speed, talent, capital — favors the newcomers.
The Bull Run ends — right in theory, wrong in practice.
Value creation didn’t slow; it merely changed shape. 2022 validated the logic. 2023–2025 contradicted it. The jury is still out.
Liquidity crisis and de-dollarization — missed.
Instead of a liquidity crunch, we got inflation.
De-dollarization is still an open bet.
Chaos increases the value of safety, and the dollar is still the safest house on the block.
Food gets disrupted — half-right.
Consumer habits shifted more than industrial infrastructure did. The physics and regulation of food move slower than startups.
Space explodes — underestimated.
Space became an infrastructure layer. SpaceX is turning into the AWS of orbit. I was optimistic; reality was even more so.
Goods transportation shrinks — early.
The logic (automation → reshoring → local production) is unfolding, but hasn’t yet bent global logistics curves. This will likely be a 2027–2033 phenomenon. The rate at which the US is reshoring points towards directionnally right, but jury’s still out.
Banks disrupted, inequality taxed, GAFAM crumble — all wrong.
Banks digitized faster than I expected.
Redistribution barely moved.
GAFAM did not crumble; they metastasized and leveraged their access to capital to bank on AI.
These predictions all underestimated a now-obvious force: AI amplifies institutions with distribution and capital, and the biggest platforms started the decade with the widest distribution and war chests in human history.
3. Technology, Automation and the Nature of Work (P21–P30)
Crypto becomes mainstream — wrong mechanism, right intuition.
Value started flowing back to users, but not through blockchains. It happened through AI-mediated digital economies. The thesis was right; the implementation layer was not.
Dematerialization accelerates — strongly correct.
The world shifted decisively toward information-first value creation, modular systems, and automated reasoning. My later work on “factory-as-a-computer” came directly from observing this from the field.
Protocol uniformization drops — wrong.
Fragmentation exploded.
LLMs didn’t unify protocols — they simply learned to translate between them.
Ownership becomes optional — correct in spirit.
Not through the Airbnb/Uber model I had in mind, but through AI-optimized access to physical capacity. Autonomous logistics, shared industrial assets, and pay-per-use models are rising faster in B2B than in consumer markets.
Creativity and emotion as the human edge — very correct.
As cognitive labor became compressible, emotional labor and creative judgment became the premium frontier. This prediction aged almost eerily well.
Moore’s law halts — correct in essence.
Transistor scaling slowed.
System performance exploded anyway thanks to specialized architectures.
Moore’s law died; something bigger replaced it.
A Scorecard (So Far)
This is the mid-decade tally:
- Geopolitics (P1–P9): ~6/9 directionally right
- Markets & institutions (P10–P20): ~5/11 right
- Technology & work (P21–P30): ~7/10 right
Total:
- 18/30 right ;
- 9/30 wrong ;
- 3/30 still undecided.
Five years to go. Not enough to claim victory. More than enough to claim useful frameworks and dangerous biases
What Reviewing These Predictions Actually Taught Me
Beyond the accuracy of any individual bet, this retrospective revealed several deeper truths about how I used to think.
1. I systematically underestimated technological compounding.
When a technology curve bends, it doesn’t flatten — it jumps to a new curve. This played out across AI, biotech, and space.
2. I underestimated the United States’ capacity to re-assert dominance through technology.
Political chaos distracts from structural strengths.
Compute, capital and talent are structural strengths.
Long US and democracy
3. I was early — sometimes very early — on automation and industrial AI.
The predictions around dematerialization, information-first factories, automation, and the rise of human-in-the-loop systems are now the core of what we build every day.
4. I overestimated crypto and underestimated AI.
In 2020, everyone did. But that’s no excuse. Technology plays on the long term and the next wave is hard to predict.
5. I understood the shift in human labor better than I realized.
Ownership decline, emotional labor, creative value, and the splintering of career paths are now mainstream conversations. In 2020 they were contrarian.
Why Do This Exercise at All?
Because predictions are not about accuracy. They are about clarity.
Revisiting them forces me to:
- expose where I was anchored to the wrong priors,
- recalibrate my understanding of compounding forces,
- and update how I build, invest and bet going forward.
And, perhaps most importantly: it forces intellectual humility.
Nobody “predicted” the last five years.
We all got punched somewhere — I sure did.
But some frameworks aged well — and those are worth carrying into the next decade.
Five Years Left
If the 2020 rule was “20 correct predictions out of 30 is a win,” I am still in the race.
But more interesting than the score is the process.
Five years from now we’ll know:
- whether the US peak is structural or temporary,
- whether China breaks through or stalls,
- whether automation rewires the industrial economy faster than expected,
- whether the great compression of cognitive labor leads to a creative renaissance or a societal shock,
- and whether AI becomes a stabilizing force or an amplifier of chaos.
I’ll revisit this list again in 2030.
In the meantime, this was the score at halftime.