The AI Transition —
What Governments Need to See
Essays on displacement, economic restructuring, sovereign compute, and what happens when the models get good enough to matter.
Why This Page Exists
I'm George Pu. I've built $10M+ in businesses without venture capital, and I've spent the last two years writing about what AI is actually doing to the economy — not what conference panels say it's doing.
This page collects my essays on the questions that matter most for governments and policymakers: How many jobs are actually at risk? What does sovereign compute mean in practice? Where does the displacement hit first, and what does restructuring look like?
These aren't policy papers. They're observations from someone who builds with AI every day and sees the gap between what founders know and what governments are planning for. I'm not selling consulting. I just think the people making decisions should have better inputs.
If you work in policy and want to discuss any of this, reach out.
The Information Gap
There's a growing disconnect between what's happening in AI labs and what reaches policy desks.
What Most Governments Hear
- •“AI will create more jobs than it destroys”
- •“Reskilling programs will bridge the gap”
- •“We have 10-15 years to prepare”
- •“Regulation will slow this down enough”
What Founders Already Know
- •AI is already replacing roles, not just tasks
- •Reskilling can't outpace capability doubling every 6 months
- •The window for structural preparation is 18-36 months
- •Countries without sovereign compute become dependencies
“The gap between what AI can do today and what most governments think it can do is the most dangerous information asymmetry of our time.”
— The premise behind every essay on this page
Four Lenses
Each theme represents a category of essays — a different angle on the same transition.
The Displacement
Who loses work, how fast, and what the second-order effects look like.
Not hypothetical. Companies are already cutting headcount and citing AI. The question isn't if — it's which roles, in what order, and how fast the cascade moves through the economy.
The Economy
Revenue concentration, wage compression, and what GDP stops measuring.
When 10 people do what 100 did, GDP might hold steady while employment craters. Traditional economic indicators will mislead. New frameworks are needed.
Sovereign Compute
Why compute access is the new geopolitical resource — and who controls it.
Three countries control the AI stack. Everyone else rents access. For nations, this is the new oil — except the supply chains are even more concentrated.
The Projection
What the trajectory looks like at 12, 24, and 60 months from now.
Extrapolating from current capability curves, cost reductions, and adoption rates. Not predictions — projections based on what's already measurable.
The Numbers That Matter
Data points from real research, not conference optimism.
The Window Is Shorter Than You Think
The Quiet Phase
Individual roles eliminated. Hiring freezes disguised as “efficiency.” Companies building internal AI but not announcing headcount impact. The data doesn't show it yet — but the decisions are already made.
The Visible Shift
Entire departments restructured. Mid-career professionals displaced at scale. Tax revenue from knowledge work begins declining. The public narrative catches up to what builders already see.
The Restructuring
New economic structures emerge — or don't. Countries that invested in sovereign compute and workforce transition frameworks pull ahead. Everyone else imports capability and exports dependence.
Get Essays on AI & Government
Long-form thinking on displacement, restructuring, and what governments should actually be planning for. No policy jargon. Just observations from the build side.
No spam. Unsubscribe anytime.