In 1984 the BBC broadcast footage of the famine in Ethiopia, and within weeks the Western world was singing along to Feed the World. What I remember most is that nobody seemed surprised. We’d been told this was coming.

Previously, in 1968, Paul Ehrlich opened The Population Bomb with a line that set the mood for a generation: “The battle to feed all of humanity is over.” Hundreds of millions would starve in the 1970s and 1980s, he wrote, and nothing could prevent it. By the time the Ethiopian images arrived, the West had spent sixteen years expecting them. The famine was in fact driven largely by civil war and government policy, not by a planet that had run out of food. But that nuance didn’t stand a chance against a story everyone already believed. I think about this a lot when I listen to the current wave of AI doomerism.

We didn’t starve

It turns out the world tried two ways of defusing the bomb.

The first was regulation. China introduced the one-child policy in 1980 and kept it for 35 years. It worked in the narrow sense that births fell, with the fallout of a skewed gender ratio and one of the fastest-ageing populations on earth. The awkward footnote is that Chinese fertility had already roughly halved during the 1970s, before the policy began. The law took credit for a trend that was already running, then added its own side effects.

The second way was science. Contraception reached women all around the world and, given the choice, most chose to have far fewer children. Rather than The Law taking action, many Western nations relied on accessibility to science and personal responsibility with choice. At the same time the Green Revolution (high-yield wheat and rice, fertiliser, irrigation) multiplied the food supply, and from the 1990s genetically modified seeds pushed yields even further. Dying of starvation because there are too many people now seems ridiculous as a forecast.

I’m a firm believer that science and innovation solve the world’s big problems, and that law and regulation mostly don’t. There are exceptions. But as a first order approximation, it has served me well.

Pandora’s box is open

AI doomerism has the same shape as the population bomb. It’s a real problem that has been extrapolated to catastrophe and is now facing the demand that someone stop it.

The difference this time is that stopping it isn’t on the table. Pandora’s box has been released, and we are both unable and (for some) unwilling to shove it back in. The use of open-weight models has opened the opportunity for people to study and adjust accordingly, and most models are already good enough to do what many people worry about. At a global level (and it is very much a global issue), it is impossible to have the world’s 200-odd countries agree on anything, let alone something as complex as AI. It has become a geopolitical and strategic issue, and no government that believes its rivals are racing will volunteer to slow down.

The hard truth is that we simply need to live with it. I keep coming back to the framing I took from Kerr Neilsen, BT’s star fund manager when I was there, which I wrote about in “Don’t be idealistic”. You can’t solve these problems or turn back the clock. Work out where you are and how to make money from it. The doomers have drawn up a long list of problems, and that’s a good thing because I invest in problems.

What regulation is for

None of this makes regulation useless. It does a different job from the one people assume. Regulation rarely solves a problem. What it does, very effectively, is create a buyer for the people who will.

I made this argument in Problems, problems, problems, where regtech was the elegant exception to the rule that angels should avoid society-sized problems. Avarni didn’t come about because the Australian Sustainability Reporting Standards fixed anything about emissions. It came about because the standards turned every large company into a customer for audit-ready reporting.

Social media is the closest precedent for where AI is heading. It spent a decade creating a pile of societal problems, which the psychologists and sociologists catalogued long before the politicians caught up. Then the regulation arrived, including Australia’s minimum age for social media from December 2025. With it came a market for age-assurance technology that barely existed a few years earlier. The law didn’t fix teenage anxiety. It created a compliance problem, and entrepreneurs moved in to sell the answer.

AI regulation is going to follow the same path, some of it good and some of it bad, and it will spawn its own generation of regtech companies.

The Mythos moment

Cybersecurity is the clearest case, and it has changed character in the last six months.

For most of my career, cybersecurity was an irritating thing in the background that you sort of had to do. The IT team asked for budget, the board nodded, and nobody outside the security team thought about it much. Then in April 2026 Anthropic announced Claude Mythos Preview. It was a model that could find, chain together, and exploit previously unknown vulnerabilities in major operating systems and browsers at a speed no human team could match. Anthropic chose not to release it publicly. Instead, it formed Project Glasswing, a restricted group of large technology and infrastructure companies using the model to find and patch flaws in critical software.

That put cybersecurity on the agenda of every CEO in the world. The detail that interests me most came from the analysis that followed: finding vulnerabilities is no longer the hard part. Patching them fast enough is. When discovery becomes cheap, the bottleneck moves, and a bottleneck is a business.

I expect cybersecurity to end up treated the way we treat insurance. Nobody debates whether a business needs insurance; it’s a cost of doing business, a slice of revenue that everyone pays without a second thought. My guess is that cybersecurity settles at something like 1% of revenue, and I’ll flag that as a bet rather than a forecast. Insurance went through the same evolution. Fire insurance only became a proper industry in London after the Great Fire of 1666. Every new hazard eventually becomes a line item, and line items get paid by every business, every year.

The world angels live in

I’ve written about why optimism is the right seat for an angel. This is the practical version of that argument. It’s a whole different approach when you read AI doomerism as an angel, and the list of threats becomes a list of pitches that haven’t been made yet. Here are some of the problems I expect founders to bring to Brisbane Angels over the next few years.

The hard problems

• Patch and remediation automation. If an AI can find a thousand holes overnight, the scarce skill is closing them.
• Watching the agents. Permissions and monitoring for autonomous AI agents acting inside company systems.
• Identity beyond the voice and the face. The deepfake call from the “CEO” instructing the finance team to pay.
• Defending infrastructure against drones. Critical sites such as airports, ports, energy sites and stadiums went from low-risk to critical overnight.
• Biosecurity. Screening DNA synthesis orders and spotting new pathogens early.

The soft problems

These are the ones the -ologists will identify.

• Provenance. Authenticating photos, recordings, or documents for courts, insurers, and elections.
• Proof of personhood. Knowing whether you’re dealing with a human online.
• Assessment integrity. How a school tests anything when every student has an AI.
• AI compliance. The regtech layer for whatever rules each jurisdiction lands on.

Some of these will have commercial buyers from day one. Others will wait for a regulator to create the buyer. A founder who can tell me which kind they’re building is already ahead of most.

One problem at a time

The entrepreneurs will find the problems and propose the solutions. That’s what they’ve always done, from fertiliser to the pill. I don’t see why AI should be any different, especially now the human race has AI tools helping it work faster.

What I don’t know is which of these problems becomes the next insurance, the line item every business pays forever. Some will turn out to be features the big platforms absorb for free. Nobody knows which is which yet, but I would rather see many of these pitches, and back the ones where I can’t articulate why they will fail, than try to pick the one problem that matters.

In 1984 the consensus was mass starvation. It was wrong, and the problem was real enough. What the forecast left out were the people who would solve it, and AI doomerism is doing it all again.

Richard Moore — MooCoo Ventures

Richard Moore is co-founder of MooCoo Ventures, an angel syndicate that co-invests alongside Brisbane Angels, one of Australia’s most active angel groups. He has made over ninety personal angel investments since 2013.

moocoo.vc

DISCLAIMER: The information provided is for general informational and educational purposes only and should not be construed as financial, investment, or legal advice. Any decisions you make based on this information are your sole responsibility. You should consult with a qualified financial advisor or other professional before making any financial decisions. Past performance does not guarantee future results.