This is a story running in the news, even on the BBC website, which is scaring a few people. There isn't any prediction of an accelerating population collapse. Indeed our population is increasing and is expected to level off some time between 2050 and 2100 and AFAIK nobody is saying our population will collapse in 2050 except these two CEOs in their interview.

This is another article I'm writing to support people we help in the Facebook Doomsday Debunked group, that find us because they get scared, sometimes to the point of feeling suicidal about it, by such stories. Do share this with your friends if you find it useful, as they may be panicking too.

skip to Aging populations are a worldwide issue

We should rely on the papers by population experts not the hunches of wealthy CEOs of technology companies. They gave no sources, and a paper that said such a thing by reputable authors would be headline news and astonishing and bizarre. Also - can I suggest to these CEOs themselves, if they see this, that they consider the effects of their remarks on scared people who are likely to believe them just because of they are famous and wealthy. I think that there is a greater need to be responsible and careful about the effects of ones comments if one is famous.

The BBC article, which just quotes them without comment, is here

This is what they said, my own transcript from the video:

Elon Musk

Most people we have too many people on the planet but actually this is an outdated view …

I think the biggest problem the world will face in 20 years is population collapse.

[repeats this several times] not explosion, collapse.


Jack Ma:

1.4 billion people in China seems a lot but in the next 20 years this thing will bring big trouble to China and the population decreasing of the whole [Earth?]

The speed of population decreasing is going to speed up. You call it collapse.

Elon Musk: Accelerating collapse. The common rebuttal is what about immigration - from where?

Perhaps they are thinking about the "toy models" that sometimes hit the headlines with stories like this. If so these are not meant for projections. Only 4 or 5 numbers to describe an entire "toy" world! See toy models (below). They are just for exploring ideas and are not meant to substitute for the detailed projections of the UN Population Division and the various researchers working on projections of our population through to the rest of this century.

The rapid decrease in China is due to its one child policy and is a special case. From 1980 to 2016, it was illegal for parents to have more than one child.

One-child policy | Definition & Facts

It recently stopped its one child policy. It is still rising but is expected to decrease by the end of the century, perhaps to a bit above 1990’s levels, and after that, it’s still expected to decrease, but more slowly. There is no way it can get down as far as 1970s levels by 2100 and even after a one child policy for 36 years it does not face an accelerating population decline in projections

Here are some of the projections by the UN Population Division for individual countries all on one graph - you can see that China is a special case

From: World Population Growth

Yes, China has a projected rapid decrease, but not in the next 20 years. It may peak some time around 2050.

In these projections, India also has a significant decrease by the end of the century, but not as much as China and not going to come down as low as it is today.

Africa has a rapidly increasing population. Nigeria for instance is still rapidly increasing and this is expected to continue through to 2100, though as we’ll see, it could level out before then.

Many countries have slowly decreasing populations but nothing that remotely resembles an accelerating population collapse. Japan has one of the fastest decreases but you can hardly call it an accelerating decrease either.

It's important to realize here - these declining populations are because of prosperity, not because we are running out of resources. Japan is one of the wealthiest countries with the fastest decraesing population. As we become more prosperous we have fewer children, and invest more in each child we do have. This is true for all forms of government, all economic systems, and all religious groups.

These are the projections for Europe and North America, along with Asia and Africa.

Figures from 2014. Most of the population growth is in Africa by the end of the century by these figures, with everywhere else leveling off by then, the least developed countries are the ones that grow most rapidly, so that's a reflection of the situation in Africa

It is just nonsense like his plan to continuously explode nuclear weapons over the poles of Mars to warm it up. What he says is not peer reviewed, not checked by anyone and highly imaginative and with a strong element of science fiction in it.

It is true that we can theoretically feed the world with even ten times the current number or more, but the faster our population increases, the more pressure on food supplies and the forests, and ecosystems.

We could do it sustainably too but that surely comes first, to increase rapidly while not solving sustainability would be to make the situation far worse.

As for the best way to have more bright wise people in the world, then education is the key not population.

We already have 7.5 billion people, and many of them don't even have a decent education. How many people are there in Africa, say, just getting by with extreme poverty, who might have solutions to help with the world's problems if they had a higher education and access to more funding to try out their ideas?

We can increase the number of well informed educated people by helping the poor and deprived people to achieve a better education, reduce child mortality and achieve better health.

Just increasing numbers without working to improve education worldwide could easily REDUCE the number of educated people.

The population is expected to level out by 2100 and possibly by 2050

The biggest projected population increases are in Africa where there's a projected increase from 1 to 4 billion by 2100 and the best way to reduce that increase is to help the Africans with sustainable development, education especially of women, and higher education?

There are signs already of the younger population having lower birth rates. If you factor in changes due to higher levels of education, the result can be anywhere between 6.9 and 12.6 billion by 2100 according to another study. See The human core of the shared socioeconomic pathways: Population scenarios by age, sex and level of education for all countries to 2100

Their lower estimates have the population peak at less than 9 billion by 2050, and their conclusion was:

  • Future fertility and hence population growth will depend on female education.
  • In the median assumptions scenario (SSP2) world population will peak around 2070.
  • By 2100 world population ranges from 6.9 (SSP1) to 12.6 billion (SSP3).

It’s interesting to notice how much of an impact tertiary education (dark blue) has. Education of women is also important.

From that paper

Since almost universally more educated women have lower levels of fertility—an effect that is particularly strong for countries in the midst of demographic transition—the changing educational composition of young women alone is a major factor influencing population growth. Lutz and KC (2011) have recently shown that even when assuming identical education-specific fertility trends, different scenarios about future female education levels already can lead to a difference of more than 1 billion in world population by 2050.

So, more education for women can knock a billion people off the projection for 2050, that's really near future.

Either way, things are actually looking brighter than one might think. Which isn't to say it will be easy, but there is no reason why we have to ruin ecosystems on Earth.

We can feed everyone through to 2100 and beyond on all scenarios but it is far easier if we have a healthy and well educated population in Africa and their population levels off earlier.

We can grow enough food for everyone through to 2100 and beyond on all scenarios


skip to Neural lace

There are issues with an aging population in China and Japan yes. But there are worldwide. In the UK we have an increasing aging population too and our population is not yet declining significantly.

Worldwide people are living longer and that's as much an issue as declining birth rate. People in many regions of the world have an average life expectancy in the late 70s or 80s. Today’s average life expectancy is 71 years worldwide. In 1950 it was 46, an increase of 25 years.

Back in 1800 the average world life expectancy was 30, now it is over 70!
Twice as long – life expectancy around the world

The UN population division projects a world population by 2100 with nearly 2.5 billion over 65 compared with 700 million today, and 0.9 billion people over 80, compared with 150 million today. They are projecting a similar proportion of the population over 80 by 2100 to the proportion that's over 65 today. In the world as a whole 80 will be the new 65 by 2100.

Meanwhile they project around 5.9 billion aged 20 to 64 compared with about 4.5 billion today.

You can explore their projections here.

In the UK we have already increased the retirement age and they are doing the same in Japan. Promoting health of the elderly is an important part of this, so that they can work for longer.

It’s particularly acute for Japan and China but we all face this, how do we make sure that we support a population with many more elderly people than for any previous generation? Japan is leading the way here, as it finds solutions.

Population ageing will be a formidable, though not insurmountable, force to contend with in the coming decades. Japan is confronting the issues of an ageing society earlier than other countries, affording it a unique leadership opportunity to devise and implement effective responses to the socioeconomic challenges of a rapidly ageing population. …. There is scope to launch a system of complementary initiatives that can lead to sustainable population growth for the future. Promoting health will naturally be a key element of such strategies.

Japan’s age wave: Challenges and solutions


skip to toy models

The actual interview is here where he then goes into lots of wild science fiction ideas such as claiming that his “neural lace” will permit “mind uploading” - that is a science fiction idea. His neural lace is only likely to be approved for treating Alzheimer's’ patients in the near future.

Live: Jack Ma and Elon Musk's AI debate in Shanghai

The recent research for Alzheimer's' patients was done by recording neural signals that fired when you store short term memories. These are not thought to be the memories themselves and they were not playing back the memories. Rather it is a signal that seems to help our brains to store short term memories.

If you play back this pattern in the future, then it helps you to store memories. Not to recover the memories stored when the pattern was recorded but to store new memories.

(click to watch on Youtube)

This is recent research from spring of 2019:

If you read the paper itself they go into detail.

First, they did 80 to 100 memory tests (Direct Match to Sample or DMS in the paper). The way the neurons fire when they were successful in this short term memory was recorded and then analysed. Based on this they worked out a signal to fire back to them and which neurons to stimulate to encourage short term memory.

Then in a second session, they did the memory tests during which they stimulated their brains with this pre-recorded signal which then helped them to lay down new short term memories.

From this you can see that the signal they stimulated them with did not encode the memories in any way at all - they were based on previous tests with different images.

This stimulation lead to a 35% improvement in performance in the memory tasks.

This is their technical description of what I just paraphrased:

Each subject provided 80–100 DMS [Direct Match to Sample] trials with hippocampal neural activity recorded from each microelectrode location. Recorded neuron data was sorted by location and firing rate from the first recording session, and was subject to analysis and characterization of task-related hippocampal cell firing via a customized MIMO [Multiple Input Multiple Output)-] nonlinear mode developed by Drs Song and Berger et al USC. The MIMO model computed nonlinear firing relationships between input–output neuron pairs that occurred during the sample phase of the DMS task associated with correct choices in the subsequent MP of the same trials. The MIMO model parameters for each subject were derived and returned to the study team within 48 h, which allowed re-testing of subjects in a second recording session with online application of the derived MIMO model for electrical stimulation of the same neural substrates recorded in correct trials in the previous DMS memory task.

Developing a hippocampal neural prosthetic to facilitate human memory encoding and recall (scroll down to Short-term memory: DMS task)

This is not useful for most people because we already store short term memories fine. You'd only need that if you have serious problems with short term memory.

It does not mean we know how memories are stored or how to “read” them. The brain is not organized like a computer. It has no separate memory, no computer chip, no sequences of 0 and 1 bits, no instructions, no computer code, nothing.

It is roughly as similar to a computer as it is to a clockwork automaton. I.e. not at all.

There is just nothing at present to suggest we can read any thoughts from a human brain.

Some think that our brains may work like neural nets. But the evidence for this is weak, and the blue brain project did not come up with convincing results and is widely regarded as having failed. See my answer to Will the Blue Brain Project lead to the creation of a conscious artificial intelligence?

Even if that were right, there is no way to recover "memories" from a neural net. You can't get google's "DeepDream" to recover the images it was exposed to during training.

Neural nets have the opposite problem of conventional computer programs. There is nothing there that corresponds to reasoning or logic or separate memories. It's like constant unreasoning imaginative leaps. You get good results by combining neural nets with conventional programming but while our brains may possibly use something similar to neural nets there is nothing there that remotely resembles conventional programs, separate memories, bit streams, bytes, bitwise operations etc.

Some (myself included) think we will never be able to have artificial intelligence either computer programmed or based on neural nets or similar. That our brains are essentially non computable in their processing (rather like radioactive decay but more complex, not just purely random non computability).

Technically the hypothesis is that an intelligence that is able to recognize mathematical truth (and probably truth generally) is not Turing machine equivalent - this rules out neural nets, digital computers, quantum computers (the designs for them currently can only do the same things as ordinary computers but faster), use of randomness, parallel computing etc. All modern computers are "turing equivalent" - if something can be calculated by one of them it can be calculated by any of them going back to Babbage's original designs for a programmable clockwork computer in the nineteenth century. They are all the same, apart from limitations of speed and memory.

Whether that is true or not, there is currently no evidence at all that we can build a computer program that is able to recognize truth for itself. A self driving car could be reprogrammed to find the first traffic light and drive into it, Alpha go could be reprogrammed to lose every game as fast as possible and there is nothing there that even knows it is doing anything different never mind care about it.

It might well be that the easiest way to reach super intelligence is through genetic modification, as in the David Brin stories, genetic uplifting, e.g. of whales, dolphins etc say. And a super intelligence is not necessarily wise. Our cleverest people are sometimes also blinkered, focused on some particular thing and not knowledgeable outside of it. Or they develop strange ideas of their own. They are able to build up bewilderingly complex patterns of ideas but these can still be misleading and confusing even to themselves. Also there is no guarantee that it has a healthy mind, it might be depressed, wonder why it exists etc. Marvin the paranoid android in HitchHiker's guide to the universe may be a good example of what a future strong AI would be like if we ever build one.

Here is a compilation of some of the best bits from Marvin the paranoid android :)

(click to watch on Youtube)


For more about this see my


We can develop interfaces that let us speak without using our voices like the way that Stephen Hawking could “speak” by twitching a muscle in his cheek. We could have ways to do that by just thinking in particular ways - but it is not reading our thoughts, rather we would learn to think in particular ways to control motion of a mouse on a screen, selecting an item and so on, like typing but with your mind. Or we can write clever programs to detect subvocalization and things like that.

It's the same for his other remarks. He is inspired by what most would consider far out and wild science fiction ideas. Most are flimsy and fall apart as soon as you poke at them and do not have much detailed thought behind them, at least nothing published.

Somehow he also manages to run a successful rocket company and an electric car company. They are based on his ideas and insight, he clearly has plenty of moments of brilliance. His Falcon Heavy lift off was a remarkable achievement for instance. As a space geek I'm cheering him on for his rocket technology accomplishments.

(click to watch on Youtube)

But it doesn’t mean you should listen to what he says on all topics, and accept everything just because he said so.


skip to Can feed everyone

You should NEVER use a toy model for prediction. This sometimes hits the news, stories where an academic sets up a ‘toy model’, a simplified model, the entire Earth with all its people, the land, oceans, atmosphere, our resources, all reduce to a few numbers and equations governing them. They are just for exploring ideas in a thought experiment simplified world - hugely simplified.

A rather fun example I remember from when I was at university in the 1970s studying maths was a seminar I went to on the mathematics of traffic flow in cities. The lecturer said something like

“Let’s model our city as a flat concrete disk. The traffic can drive around this disk along any circle centered on the center of the disk and it can also drive radially towards the center or away from the center”

He then used this “toy model” to discuss traffic flow in cities with ring roads.

Then there’s the Spherical cow joke:

Milk production at a dairy farm was low, so the farmer wrote to the local university, asking for help from academia. A multidisciplinary team of professors was assembled, headed by a theoretical physicist, and two weeks of intensive on-site investigation took place. The scholars then returned to the university, notebooks crammed with data, where the task of writing the report was left to the team leader. Shortly thereafter the physicist returned to the farm, saying to the farmer, "I have the solution, but it works only in the case of spherical cows in a vacuum".

Spherical cow - Wikipedia

Spot the cow

Except it should be in a vacuum:

Spherical cow - Wikipedia

As a joke in Big Bang theory:

There are many similar toy models applied to the whole world and they normally do not model social responses that can make our future sustainable. But you can also do a toy model with a sustainable future. They can't be used to predict anything about our real world.

You might as well use a toy model of a spherical cow in a vacuum to predict milk yields Holstein Friesian cattle - Wikipedia

Cow female black white

Modeled as:

That of course can't be used to predict anything about milk yields and how to feed your cow for optimal milk.

In the same way these toy models of population are not in any way a substitute for the detailed studies of the IPCC, IPBES, world population division etc.

This is an example, a toy model with only four numbers to describe it at any time. The variables are

  1. Nature which can be made into Wealth (but eventually regenerates)
  2. Commoners who convert the Nature into Wealth
  3. Wealth (from Nature)
  4. Elite who take the Wealth from the commoners but don't do any extracting themselves.

It has a fixed birth rate for the commoners and elite, the same for both. The commoners and the elite vary in lifespan. And that's it. They then set it running with initial numbers for those four things and see what happens.

This hit the headlines as

(It’s not actually a NASA study as that article says)

This can collapse, can oscillate or reach a stable state. This is an example of a collapse.

“Collapse can be avoided, and population can reach a steady state at maximum carrying capacity if the rate of depletion of nature is reduced to a sustainable level and if resources are distributed equitably.”

Human and nature dynamics (HANDY): Modeling inequality and use of resources in the collapse or sustainability of societies

Remember this is a toy world that is completely described with four numbers and some equations connecting them.

They run it forward in time and find that its ecosystem collapses, and civilization collapses. That’s no surprise as it is often built into their assumptions behind the model, with usually no social response, or some simplified inadequate response. This then becomes a news story.

But it is of the essence of a toy model that we do not believe that numerical outputs will be accurate.

“… a toy model is a consciously over-simplified model of some real-world phenomenon that typically attempts to study the effect of only one or two of the factors involved while ignoring many complicating real-world factors. It is thus“qualitative” in the sense that we do not believe that numerical outputs will be accurate.”
Toy models in Population Genetics: some mathematical aspects of evolution

Here is another toy model that does model social processes. The social processes are migration, environmental awareness, social learning, and voting

Without social processes, you see the atmosphere rise to above 1000 gigatons and the upper oceans to a similar level of sequestered CO2. Meanwhile the plants and soils are devastated, losing more than half of their natural carbon content and the fossil fuel reserves in this toy model are reduced to a tiny fraction of the original amount with diminishing returns.

In the right hand side you see the gigatons of carbon in the upper oceans dark blue and atmosphere, light blue, decreasing and the amounts in the fossil fuels, gray, staying the same once environmentally friendly individuals are in the majority, and the CO2 in plants and soils staying steady.

In short, it models a sustainable world due to social processes while without them the CO2 in the atmosphere and sea goes up, the carbon in soil and vegetation plummets and the fossil fuels are all used up.

It is not plausible that we would do nothing when faced by such a scenario.

This is an example simplified toy model but gives a way to model social responses to climate change in a qualitative way. It is meant as a framework that can be used by other researchers to explore these questions.

This is the paper:

This is their website:

I talk more about toy models here

Simplified ‘toy models’ of Earth’s future without social processes often bleak - can be sustainable with zero emissions if you add them in

This is my best guess about what may have lead Elon Musk to say such an absurd thing about our future prospects, perhaps he read a paper or saw a presentation on a toy model?

He clearly does not have time to research these things, perhaps because he is so occupied full time running his businesses?


Skip to See also

So, the world population is expected to level off some time between 2050 and 2100 but because of prosperity not because of running out of resources. We have already reached peak child, almost. We have nearly the same number of children in the world now as we had ten years ago and the population is growing rapidly because of the increases in life expectancy.

We can grow enough food for everyone through to 2100 and beyond on all scenarios

Also we can continue with GDP growth. With a circular economy, we can have sustainable GDP growth. To see how that's possible - with the right technology you can create value from sand air, and sea water with saltwater greenhouses growing tomatoes, say. There is no need to use up the world's resources to continue with GDP growth.

For more on this:

The IPBES report in early 2019 outlined a sustainable future for our planet, also preserving biodiversity. See


Skip to Doomsday debunked

And on the Paris agreement:

See also

Doomsday Debunked

back to top

Our petitions

Seven tips for dealing with doomsday fears

If you are scared: Seven tips for dealing with doomsday fears which also talks about health professionals and how they can help.

If in the middle of a panic attack, see

Useful links to bookmark

Tip, bookmark those links to search for debunks more easily. Here is a screenshot of my bookmarks

Facebook support group

Facebook group Doomsday Debunked has been set up to help anyone who is scared by these fake doomsdays.

Wiki Doomsday debunked wiki

If you need help

Do message me on Quora or PM me on Facebook if you need help.

There are many others in the group who are available to support scared people via PM and who can also debunk fake Doomsday “news” for you if you get scared of a story and are not sure if it is true. See our debunkers list

If you are suicidal don’t forget there’s always help a phone call away with the List of suicide crisis lines - Wikipedia