Apparently, I’m Working on the Wrong Existential Crisis

It’s the End of the World. I Have Work to Do.

Yesterday afternoon, I got on a call with someone who wanted to work with me.

I’m still not entirely sure whether he was looking to partner with me or sell me something. But early in the conversation, he mentioned that he had been working in artificial intelligence for more than a decade.

And then the conversation took a turn I was not expecting.  He told me, with a sincerity and conviction few possess, that protecting ourselves from AI was the problem we should all be focused on. 

Not job losses. Not misinformation campaigns. Not even the extraordinary amount of energy and resources AI is beginning to consume.

His concern was considerably more fundamental.

AI was going to wipe out humanity…within five years!

Which led him to a question I wasn’t expecting. Actually, it shook me. Why was I wasting my time on climate change when AI would end humanity long before climate change became the issue we feared?

Okay, yep. You’ve got my attention, I thought.

I have spent over 25 years working on climate change. I invest in technologies intended to address it. I write about it. I speak about it. I genuinely enjoy trying to understand and solve pieces of what will be the largest economic transformation of our lifetimes.

And here was someone who had spent more than ten years working in AI effectively telling me, “You’re working on the wrong existential crisis.”

He then pointed me to Anthropic.

It turns out he wasn’t making this stuff up.

Anthropic’s IPO prospectus warns that increasingly powerful AI systems could create “catastrophic or existential risks to humanity.” According to Reuters, about 80 of the 261 pages in the main body of the prospectus are devoted to risks. The filing describes the possibility of models behaving in unexpected ways, resisting shutdown or manipulating information. 

Consider how extraordinary that is.

A company preparing what could be one of the largest IPOs in history is effectively telling investors that its technology has tremendous potential to transform the world and, under some circumstances, could help end it.  And people are lining up to invest.

Welcome to 2026.

So I spent much of the rest of yesterday thinking about the conversation.  Am I working on the wrong problem?

The more I thought about it, the more I realized I disagreed with the premise behind the question.  Not necessarily his concern about AI, though I am much more optimistic than he, clearly.  The idea that because his problem might be bigger, more immediate or more existential, mine somehow mattered less.

Is the AI-fuelled apocalypse real?  I don’t know.  I sure hope not.  And I am certainly not qualified to tell you that the people deeply concerned about it are wrong.

There are very smart people taking the risk seriously. The 2026 International AI Safety Report brought together more than 100 experts from over 30 countries and international organizations. It concludes that AI capabilities are improving faster than many experts anticipated. There is growing evidence of AI being used in cyberattacks. Developers have added safeguards to some models because testing could not rule out their potential usefulness in developing biological weapons. And as models become more sophisticated, reliably testing what they are capable of has itself become tougher.

These are not concerns we should dismiss.

But uncertainty cuts both ways.

A survey of 2,778 researchers who had published in leading AI platforms found something fascinating. Between 38% and 51% assigned at least a 10% probability to advanced AI producing outcomes as bad as human extinction.

That certainly makes the hair on the back of my neck stand up. Yours?

Yet 68.3% believed good outcomes from superhuman AI were more likely than bad ones. Even among the optimists, many assigned a meaningful probability to disaster.  I actually find that reassuring.  Not the extinction part, obviously. What I find reassuring is the ability to recognize a potentially enormous risk while still believing in the possibility of an enormously positive outcome.

That feels like a much more useful way to approach the future than deciding in advance that we’re doomed.

Many of us have our own problem - or problems.

Mine is climate change.

The evidence there isn’t particularly comforting either. 

The World Meteorological Organization reports that 2015 through 2025 were the eleven hottest years ever recorded. Last year was approximately 1.43°C warmer than the 1850–1900 average. The oceans continue to accumulate heat, greenhouse gas concentrations continue to rise and the impacts are showing up around the world. Fire. Flood. Drought. Famine. War. Refugees.

After 25 years of working on this, none of that makes me want to give up.

Quite the opposite.

I have had a front-row seat to what happens when enough people decide a seemingly impossible problem is worth working on. 

Global energy investment is expected to reach $3.4 trillion this year. About $2.2 trillion of that is going into clean energy, almost twice the capital flowing into fossil fuels.

Is that enough? No.

Are we moving quickly enough everywhere? No.

Have we figured out how to decarbonize everything? Not remotely.

But compare where we are today with where we were when I started working on this problem.  Entire industries now exist that barely existed then. Technologies have moved from laboratories to massive global markets. Investors who once regarded climate as philanthropy now understand that the transition involves trillions of dollars of capital formation, infrastructure and industrial transformation.

We haven’t solved climate change.

But we have demonstrated that we can do extraordinarily difficult things when enough people decide to work on them.

There is a difference.

And here is the strange part. I love it.

To be clear, I don’t love climate change. 

I would be delighted to wake up tomorrow and discover that we had solved it.

What I love is working on the problem.

I love its complexity. I love that solving it requires us to rethink energy, transportation, agriculture, buildings, materials and entire industrial systems. I love meeting entrepreneurs who look at something everyone else regards as impossible and decide to build a company around proving them wrong.  I love trying to understand where capital can change an outcome.

There are days when working on climate change is infuriating. There are setbacks. Technologies fail. Policy changes. Markets move. Companies die. Money is lost.

And then someone solves something.

That’s addictive. 

If we solved climate change tomorrow, I don’t think I would retire and spend the rest of my days congratulating those that solved this existential crisis.  I’d probably go find another big problem.  Because I don’t think human beings thrive in the absence of problems. I think many of us thrive by finding problems worthy of us.

That is a very different thing.

That's what bothered me about yesterday's conversation.

My concern wasn’t that this guy was worried about AI.  Maybe he should be. Maybe I should be. Maybe we all should be.

What bothered me was the implicit hierarchy: My existential problem is bigger than your existential problem, so you’re wasting your time.

I don’t buy that.

We don’t need everyone working on climate change.  We need enough exceptional people working on climate change. We need people obsessed with making AI safe. We need people obsessed with curing cancer. We need people trying to solve food security, biodiversity loss, poverty, energy access and problems I haven’t spent five minutes thinking about.

This is how humanity works. It how a civilization progresses and improves.

Progress doesn’t require eight billion people to agree on the single most important problem and then march toward it in formation.  It happens because different people become almost irrationally committed to different things.

I have met founders who have spent a decade thinking about a molecule, a manufacturing process or a piece of equipment most people have never heard of.  Thank goodness they have.

The question shouldn’t be, "Is my problem more important than yours?"

It should be, "What problem moves you enough that you are prepared to do something about it?"

I say we start with hope.

My friend Afdhel Aziz has built much of his work around an idea I love. He says, "Start With Hope."

There is something powerful in that language because hope isn’t the same thing as believing everything will be fine.  I don’t believe everything will be fine.

The climate will continue to change. Some companies I invest in will fail. Some have. Governments will make bad decisions. Technology will create problems at the same time that it solves others. And perhaps AI presents risks we have barely begun to understand.

Hope, to me, is something more demanding.

It is the belief that our actions still matter.

If you convince yourself that climate change is inevitable and nothing can be done, there isn’t much reason to do anything.  If you convince yourself that AI will kill everyone in five years, why build a business? Why cure a disease? Why educate your children? Why plant a tree?

Why do anything?

I refuse to accept that as a useful way to live.

And what if AI doesn't kill us (queue Pascal's wager for the existence of God)

Here’s the possibility I would rather spend some time on. 

What if AI helps us become better at solving the problems we already have? 

I’m not suggesting we throw caution to the wind and hope for the best. Quite the contrary. If there are serious risks, I want brilliant people devoting themselves to understanding and mitigating them.

But think about the other side of the equation.

We are developing systems capable of analysing quantities of information humans simply cannot process. They are already becoming extraordinarily capable in mathematics, coding and science. 

What happens when we apply that capability to chemistry? Materials discovery? Grid optimization? Agriculture? Medicine? Energy efficiency? Climate modelling?

We spend a lot of time imagining an AI that looks at humanity, concludes we are inefficient and finds a way to get rid of us.

Maybe.

But there’s another possible outcome.

Perhaps intelligence powerful enough to recognize all of our frailties can also help us overcome some of them.

I’d rather work toward that future.

Where does that lead me?

Twenty-four hours after that conversation, I still don’t know whether the guy I spoke with is right about AI.

I don’t think he knows either.

But I do know this...

Don’t tell me my problem isn’t worth solving because yours might be bigger. Tell me what problem moves you enough to spend a meaningful part of your life trying to solve it. Then get to work.

Mine is climate change.

For someone else, it may be AI safety. Or cancer. Or education. Or homelessness. Or something very specific that the rest of us haven’t realized is a problem yet.

Find the one that gets under your skin.

Learn everything you can about it.

Build something. Invest in something. Change something. Bring people with you.

And if the scale of the problem occasionally overwhelms you, look around at all the other slightly unreasonable people doing the same thing in their corner of the world.

That’s what gives me hope.

Maybe AI is going to kill us all in five years.

I certainly hope my new acquaintance is wrong about that.

In the meantime, I have work to do.

---

To read more from Nelson, you can purchase The Gigacorn Hunter: Seven Principles for a Climate Investor here.