Author’s Note:
Over the weekend, I commented that I see artificial intelligence as a bigger priority than climate change. I stand by that, but the statement was too broad, so I want to be more specific.
I’m not saying climate change is unimportant or that it should take a back seat. What I mean is that AI is developing faster than our laws and institutions can keep up, and the decisions we make in the next few years could shape how we handle every other challenge, including climate. Both deserve serious attention. My concern is that AI isn’t yet getting the public urgency it needs.
We Know How to Stop Climate Change and Won’t. We Don’t Know How to Stop AI and Are Racing Anyway.
Two catastrophes. One is a failure of will. The other is a failure of knowledge. Only one of those responds to pressure.
The Jack Hopkins Now Newsletter #1,024: Tuesday, September 15th, 2026.
Let me start with a confession about how these arguments usually go.
Somebody says AI is a bigger threat than climate change. Somebody else says that’s absurd, look at the fires. And then both of them spend an hour comparing two things that cannot be compared, because they’re arguing about MAGNITUDE when the real difference is TIME.
So let me make the narrower argument. The one I think actually holds up.
Climate change is more certain. AI is faster. And speed is the thing nobody is pricing correctly.
Climate Runs on Physics. Physics Has a Speed Limit.
Here’s the strange mercy buried inside the climate problem.
It’s slow.
Carbon dioxide already in the atmosphere commits us to warming that arrives over decades. The oceans absorb heat on a timescale measured in centuries. The system is enormous…and enormous things have inertia.
That inertia is horrible news…because the damage already locked in cannot be unlocked.
But…it’s also the only thing keeping the problem legible. There is no version of 2027 where climate change ends civilization. The physics doesn’t allow it. Whatever is coming…it’s coming at the pace of an ice sheet.
You cannot say that sentence about the other thing.
The Other Thing Has No Speed Limit
Dario Amodei, running Anthropic, wrote this out loud on Saturday:
Six to twelve months before a swarm of AI systems could be capable of taking over the entire internet potentially causing hundreds of billions of dollars in damage.
Jacob Coxon, who quit Anthropic last week, told the Wall Street Journal things could be out of control by the end of next year.
Now. You’re free to think those numbers are wrong. Plenty of serious people do.
But…sit with what it MEANS…that credible insiders can construct such a sentence at all.
Nobody…not one climate scientist alive…can write “eighteen months until collapse” and be taken seriously. The physics forbids it.
In AI, the men building the thing say it casually…on a Saturday…in public…and their own safety leads nod along.
That’s the asymmetry. Not severity. Velocity.
And Here’s the Part That Bothers Me Most
We know how to fix climate change.
Read that again…because it’s easy to miss how remarkable it is.
We know the mechanism. Emissions…forcing…temperature…consequences. We know what would stop it. Decarbonize the grid…electrify transport…price the externality. Every piece of that is engineering and politics.
We are failing to do it. That’s a failure of WILL.
AI is not that.
Evan Hubinger, who leads Alignment Science at Anthropic…said publicly…last week that the company does not yet have a plan to solve alignment for superintelligence and is not clearly on track to.
Doesn’t have a plan.
Not “has a plan and lacks the funding.” Not “knows what to do and can’t get Congress to act.”
Doesn’t know how.
That’s a different species of problem. Climate is a solved problem we won’t execute. Alignment is an unsolved problem we’re racing toward anyway.
Now Let Me Argue Against Myself
Because if I don’t, somebody else will, and they’ll have a point.
Climate is CERTAIN. AI is a probability estimate.
Warming is measured. It’s in the temperature record…the sea level record…the insurance markets. It is happening, right now, at a pace we can graph.
AI catastrophe is a forecast. A serious one…from serious people…Hubinger put extinction risk above 10% this decade…but a forecast made largely by people who work in the industry making it.
Those two things are not the same kind of knowledge. A thermometer and a probability estimate don’t belong in the same sentence…and anyone who tells you otherwise is selling something.
So…no…I won’t tell you AI is BIGGER. I don’t think that claim survives contact.
I’ll tell you it’s SOONER…and that sooner deserves its own kind of attention.
The Inversion Nobody Expects
Here’s where it gets interesting, and where I part company with most of the doom crowd.
Speed cuts BOTH directions.
Climate requires coordination among eight billion people and every government on earth…including several that have no intention of cooperating. It requires rebuilding the entire energy infrastructure of the planet.
AI requires agreement among a few thousand people inside roughly five companies, most of them in one country…most of them speaking the same language and posting on the same website.
That is not a hopeless coordination problem. That’s a small one.
And last week, THREE of them…Musk, Altman, Amodei…publicly asked to be slowed down.
I cannot think of another civilizational risk where the people causing it have asked…on the record…for somebody to stop them.
The faster threat may well be the more fixable one. Those are not contradictory. They’re both consequences of the same fact: it’s concentrated.
They’re Also Tangled Together
One more thing, since it’s the piece both camps ignore.
These aren’t separate problems.
The IEA projects global data center electricity consumption roughly doubling from 485 terawatt-hours in 2025 to around 950 by 2030…about 3% of world electricity demand.
AI-focused data centers grew 50% in 2025 alone. In Ireland, data centers already consume around 21% of national electricity.
So AI is making climate marginally worse right now, today, measurably.
Though…I’d note the honest counterweight: data centers currently account for roughly 0.5% of global CO2 emissions…and the IEA sees that reaching perhaps 1% by 2030. It’s real. It’s not the main event.
What I Actually Believe
Not that AI is worse than climate change. I don’t think that’s knowable…and the comparison collapses two different axes into one number…which is how these conversations always go wrong.
What I believe is this:
Climate is the more certain catastrophe. AI is the more urgent one. And…urgency should determine what you attend to right now…because attention is the scarce resource…not concern.
You’ve got room for both. But only one of them has a window that might close inside a year…and only one of them still has a small enough table that a few thousand people sitting down could change the outcome.
That’s where the leverage is.
Leverage is perishable. Certainty…isn’t.
#HoldFast
Back soon.
-Jack
Jack Hopkins
P.S. The single most useful distinction in this whole argument is one I’d ask you to steal and use everywhere: we know how to solve climate change and won’t, versus we don’t know how to solve alignment and are racing anyway.
Those are opposite failures. One is a failure of will, which is a political problem. The other is a failure of knowledge, which is not. Political problems respond to pressure. Knowledge problems respond to time…and time is precisely what nobody in this race is willing to spend.
P.P.S. Somebody will read this and hear “so climate doesn’t matter.” That’s not what I said and it’s the misreading I’d most like to head off. The fires are real. The water is rising. Both things are true at once…and a person who can only hold one catastrophe in his head at a time is going to be badly surprised…by whichever one they set down.
Sources
Business Day’s run of the Reuters report, including Amodei’s framework and his 6-12 month warning: https://businessday.co.za/world/2026-09-13-trump-dismisses-very-negative-forces-warning-over-ai-risks
Reuters via Yahoo, on Trump’s response and the Musk/Altman/Amodei alignment: https://www.yahoo.com/news/politics/articles/trump-says-very-negative-forces-142843614.html
IEA, “Key Questions on Energy and AI” — data center demand doubling from 485 TWh (2025) to 950 TWh (2030), AI data center growth of 50% in 2025: https://www.iea.org/reports/key-questions-on-energy-and-ai/executive-summary
Carbon Brief’s context piece — data centers at roughly 1% of global electricity and 0.5% of CO2 emissions today, reaching about 1% of emissions by 2030: https://www.carbonbrief.org/ai-five-charts-that-put-data-centre-energy-use-and-emissions-into-context
Brookings, on US data center demand growth and the Google emissions paradox: https://www.brookings.edu/articles/global-energy-demands-within-the-ai-regulatory-landscape/
Sad but True 😞
Yes I'm keeping my eye on Kash Patel's hearing and this guy has zero class or decorum, and wonder if he even knows where's he's at?!!