Reality, Probability, and Perception

Note: All of the graphs in this essay are conceptual except for the IPCC climate projections and nuclear fallout projections from the papers referenced.

This essay is adapted from a previous Frankly video titled “Reality, Probability, and Perception.”

---

Imagine being in a room with 20 or 30 of your friends, family, and neighbors – and asking them these questions:

What do you think the global temperature will be in 2100, relative to pre-industrial levels?

How big will the global human economy be by 2100, relative to today?

Of the 6,500 species of mammals alive today, how many will still exist in 2100?

You’ll likely hear a wide range of answers – even in a group this small. All of us have discrete views on this, ranging from ‘I have no idea, I don’t even know about next week!’ all the way to very specific predictions.

Eventually, there will be a concrete answer to these questions about the future, whether that be for your city, for your country, or for the planet.

Now, consider your own answer to those questions. You may articulate one number or scenario out loud – but in your mind, you’re not picturing just a single outcome. You’re likely envisioning a distribution of futures. You voice a discrete number you think is most likely, but it’s subconsciously resting on a mental curve of possibilities. But what exactly does this mean, and why does it matter?

Playing the Odds

The future of the global climate or projected growth in GDP are rather complex topics – so let’s start with a simpler example of how probability works in the real world. Let’s talk about NFL football.

Who played in the most recent Super Bowl? Since this was written in the summer of 2025, we now know it was the Philadelphia Eagles versus the Kansas City Chiefs.

Now imagine if I had asked you the same question in the summer of 2024 – when you knew nothing about the upcoming season – how would you have answered?

You might have started with the fact that there are 32 teams in the NFL, 16 in each division, and one team from each division will end up in the Superbowl. Statistics tells us that if you multiply the 16 teams in one division by the 16 teams in the other, you’ll get 256 possible matchups. If you knew nothing about football at all, you might assume that all teams had an equal .4% chance of ending up in the Super Bowl.

But of course, not all teams have equal chances. The varying track records, players, and other factors make some teams much more likely to get to the end than others. These differences shape a probability distribution, with the predicted best teams in the fat middle part of the distribution, and the long shots in the thin tails on either end.

Statistics also tells us that all of the probabilities under this distribution curve add up to 1 – in other words, 100% of the possibilities that could happen are within this graph.

This shape of distribution is what we call a ‘normal curve.’ The fat middle with the most likely outcomes comprises about 68% of scenarios. The next tier of contenders account for another 27% and the remaining long-shot teams – unlikely but not impossible – make up less than 5%.

As the football season progresses, that curve changes shape. Some teams are mathematically eliminated. Others overperform or underperform. By the time the playoffs arrive, the number one seeds have the highest odds, and the distribution becomes much tighter and taller as the curve narrows.

Finally, once the Super Bowl is played, the distribution has collapsed into a single discrete outcome, represented by a vertical line in the graph.

In the case of 2025, it was the Philadelphia Eagles and the Kansas City Chiefs – where the Eagles came out on top.

But this line, which is now a concrete event that has occurred with certainty, was but one scenario among 256 along our original distribution. When we think about betting and odds at the beginning of this season, this distribution shapes our expectations.

This is a helpful way to think about the future – not just in football, but in any complex system. Whether we’re talking about global temperatures, economic growth, biodiversity loss, or energy systems, we’re not dealing with certainties. We’re dealing with distributions with ranges of possibilities – some more likely than others – all changing over time.

The Super Bowl is just a simple, familiar metaphor. But it helps us understand the deeper question: how do we prepare for a future that’s shaped not by certainty, but by probability?

The Unknowable Probability Distribution of Reality…

Just like in football, a distribution of possibilities exists for the outcome of nearly any issue or situation at a given date in the future. Importantly, these distributions don’t need to be normal in shape – in fact, they rarely are. They take all sorts of shapes depending on the context and evaluations of the person or entity creating the distribution.

The tail ends of these distributions shrink as we get closer to the date in question – and once we arrive there is no longer a distribution, but a single point (just like we saw in the Eagles vs Chiefs scenario). There will be a specific reality, and one of those possible outcomes will end up being what actually happens. But no one on Earth can say with certainty which it will be.

Let’s take an example. Above is a hypothetical distribution of global temperature rise by the year 2100, as viewed from today’s perspective. The curve spans a range – from a low of +1ºC from pre-industrial baseline to a high of +5.5ºC – each representing a different possible future.

There’s a variety of reasons why we might see this curve tighten and narrow. As we get closer to 2100, the Earth may continue to warm, eliminating some on the lower end. On the opposite side, if (and that’s a big if) serious global policies are implemented, hotter futures on the left may become less likely. A better understanding of the science itself could also rule out some possibilities. The distribution will continue to narrow as the years pass until 2100, at which point there will be one answer for how much global temperature has risen from the perspective of someone living in 2025.

This all makes sense conceptually – but it gets more complicated in practice.

That’s because reality isn’t shaped by just one probability distribution. It’s shaped by many – overlapping, interacting, and constantly shifting. The image above shows a few examples: climate, GDP, biodiversity, plastic pollution, average income, and the amount of remaining topsoil. Each of these has its own distribution. Each can influence – and be influenced by – the others.

Although this layered dynamic applies to all issues in the future, let’s just focus on climate for now. To simplify things, we’ll measure ‘climate’ using the same temperature distribution we’ve been looking at – now in the upper right hand corner. As we look at how these distributions are shifting, the ‘worse’ outcomes will be to the left of the distributions and ‘better’ outcomes will be to the right.

When we think about how we’re approaching the existential risk of global heating, our first instinct (understandably) is often to demand policies that will rapidly reduce temperature-increasing activities. However, any human event or policy that suddenly shifts the distribution of future temperature to the right towards cooler futures will also shift all of the other distributions in response – for better or worse. More benign climate outcomes are not independent of all the other issues humanity faces.

In effect, the graphs of future events move in parallel or inversely with all of these other factors. What is good for poor people may not be good for dolphins or forests. Right now, most policies that help the environment tend to constrain economic growth – and policies that support economic growth tend to increase environmental damage.

This doesn’t mean we can’t intervene. But it does mean we have to accept that the distributions are linked – and that pulling on one thread will tug at the whole system.

We can use this relationship as a rule of thumb, but we also cannot fully know how each of these distributions will affect the others. The number of variables and unknowns adds up to a lot of uncertainty about what the big picture will look like in the decades and centuries from now.

Science Gets Us Close To The Truth

No one can possibly know the future. In most cases, the best humans can do is draw on the collective insights of scientists and experts. The aggregated understanding and probability distributions they produce are our best tools for anticipating future realities.

Directly linked to the issue of what future temperature will look like is the issue of global CO2 emissions – meaning understanding the distribution of likely future emissions can tell us a lot about future temperature. Luckily, scientists have done a lot of research towards modeling such things.

The probability distribution above is from a real scientific paper, which aggregates projections from the IPCC’s climate models, showing expected global temperature increases by 2050 relative to pre-industrial levels. The midpoint of this distribution is about +1.8°C, with over 60% of model outcomes falling between +1.8 and +2.2°C.

Note that there’s close to zero chance that temperatures will be above +3ºC or lower than +1ºC in 2050.

If we extend this analysis (from the same paper) to the year 2100 (the dark green line), now adding another 50 years of emissions, as well as accounting for the lag effects of past emissions. Now, we see the distribution is flatter, wider, with more uncertainty and a higher midpoint at around +2.8ºC. Our least likely points are now around +5ºC at the high end and +1.5ºC at the low end.

As an aside, when seeing these charts – depending on how informed you are and how much you believe in climate change – you’ll think these numbers are either very conservative or highly overblown (or possibly just right). We’ll revisit these reactions later.

These distributions – the aggregate of scientific predictions – are the best projections that we can use for our cultural decisions.

However, there is a caveat. An important one.

Most scientists are experts in one field, which can result in isolation from the effects of events outside of their usual field of study. This is in addition to the political pressures of conventional economic expectations and business-as-usual assumptions which influence long-term projections, which means experts are often operating with a lack of awareness of how other factors could affect their predictions.

Quick disclaimer: I’m not a climate scientist. I hold a PhD in natural resources and I’ve read around 100 climate papers – but I haven’t done in-depth modeling of the climate system itself. What I have done is spend the last 25 years studying complex systems, particularly at the scale of the global human predicament.

From that perspective, I can point to at least three major systemic blind spots that weaken the accuracy of climate distributions based solely on IPCC models.

First, (as often mentioned in my work) most of society, including the average climate scientist, is energy blind. As of 2025, there is not a single integrated assessment climate model from the IPCC that incorporates a scenario with meaningfully negative economic growth in the next few decades due to fossil fuel depletion accelerating.

Viewed from the perspective of the Carbon Pulse, the standard economic and technological outlooks for income and wealth in the coming 75 years are, at best, optimistic. At worst, they’re wildly detached from physical constraints. A decline in economic growth would almost certainly reduce emissions – shifting the future climate distribution to the right, toward less severe outcomes.

But – there’s an important second order effect to consider here, which serves as the second systems oversight in conventional climate models. For individuals, reduced access to oil and natural gas, combined with tough economic times, typically forces a return to energy sources like wood and coal. These are far more damaging for global heating and ecological stability. In aggregate, this can lead to a dramatic increase in the consumption of coal and biomass, as shown in recent economic downturns such as 2009 in Greece and 2022 in Germany. (More on this in my Frankly on the 6th Pool of Carbon).

Currently, there are no major climate models that factor in the potential for increase in deforestation for energy. In such a scenario, we not only burn the stored carbon, but we remove the forest carbon sinks in the process. Incorporating this would actually shift the distribution back toward a worse temperature outcome.

Third, the possibility of a nuclear war remains unaccounted for in the climate distributions. While hard to quantify, several leading estimates place the probability of a strategic nuclear exchange at 20–50% by 2080. In my own mental model, I assign a rough 20% chance, and that would only increase as geopolitical tensions intensify along multiple fronts.

Technically, a large nuclear exchange would be “good” for global temperature. Climate models estimate a short-term cooling of up to 15–17°C, and a longer-term cooling of 1–2°C, shown in the graph above.

This would dramatically shift the temperature distribution to the right.

But of course, this wouldn’t be a hopeful scenario, but a disaster. It would be catastrophic across every other metric – reminding us that there is more to environmental and ecological health than just mitigating rising temperatures. A nuclear exchange would devastate the biosphere and likely trigger a Sixth Mass Extinction. It would reduce emissions – but at an unbearable cost.

These three examples – energy decline, biomass substitution, and nuclear risk – are just a sample of the broader systemic factors not currently captured by mainstream climate models. That’s not a critique of the science. It’s a reminder that the real world is complex, interconnected, and full of feedback loops.

Incorporating the above three things into my personal mental distribution for future temperatures would look something like the above (theoretical) graph.

Of course, this is just from my individual perspective: I could be slightly to completely wrong. That leads us to the most central point of this essay.

Your Individual Perception and Distribution of the Future

Let’s return to our example of asking yourself and a room of your peers what you think the temperature will be in 2100.

What you say out loud will be a specific prediction, perhaps through a number or through painting a mental picture of what it might feel like to live during that time. But what’s actually going through your mind – often without you realizing it – is a distribution of the sort I’ve been illustrating here.

When we say:

“I think temperature in 2100 will be 3ºC warmer than today.”

What most of us really mean is:

“3ºC is the middle of my mental distribution of what the future will be.”

That one number is simply the center of a much wider spread of possibilities you’re imagining, even if you don’t articulate the rest.

When it comes to how we think about the future, there are few evolutionary biases that also shape our distributions.

For example, humans prefer certainty – so we tend to adopt narrow mental distributions that help us avoid the stress of worrying about extreme or unpredictable outcomes.

Another is that, as extremely social beings, our distributions tend to converge around the opinions of people in our accepted social groups – regardless of how much they’re rooted in reality.

Relatedly, we also have strong bias toward binaries – yes or no, this or that, black and white answers. Nuance and complexity can even be physically uncomfortable, especially if the information is outside of what our ‘tribe’ is saying or challenges our sense of identity, making it difficult to break out of mental models once they’re adopted.

There are actually lots of different evolutionary memes that influence our perception of the future – but going through them all here would be beyond the scope of this essay.

How these biases play out in your specific mind also depends on the information available to you and your life experiences. But by definition, each of us believes that our distribution about a certain event or issue is the ‘correct’ one, or else we’d adopt someone else’s view. If someone convinces us that their distribution better describes the most probable outcomes, then our mental distribution will shift in alignment.

Each purple curve represents a hypothetical worldview on some issue by a single person. The ‘width’ of the distribution indicates one’s uncertainty (taller, skinnier distributions represent high confidence/certainty) and the placement of the distribution shows how polarized one’s views are. When we take all this together, you can imagine the vast diversity in opinions among billions of humans when it comes to imagining the future of: AI, global heating, GDP, inequality, politics, and even football. This has always been the case, but there’s a few things in our modern environments that are changing the landscape of these distributions, the biggest one being social media. This change has manifested in (at least) three key ways.

The Social Media Effect

First, we spend much less time together in person via shared physical spaces. Without this time, we’re less able to create a history of interactions to create trust and tolerance. We’re left with fewer intimate relationships to anchor us and are more easily influenced by online dynamics.

Second, people can find reinforcement on the internet for their own beliefs on just about any topic. It’s becoming more and more difficult to discern what information is being manipulated, taken out of context, or all together faked. Without the tools to navigate this, it becomes all too easy to fall into the confirmation bias echo-chamber. Compared to a pre-social-media world, our discourse is missing a lot of the uncertainty that once forced us to reassess our distributions.

Third, social media (in its current form) also pushes us to further extremes. The way the most popular social media sites are built incentivize the loudest, most outrageous voices – which is good for keeping us on the app. The side effect is that we become more siloed from considering or even recognizing the gaps in our worldviews. As a result, our distributions of the future also become increasingly polarized, aka narrower, with less and less overlap and perceived common ground.

We can now see that, compared to a ‘typical’ distribution of beliefs (shown in the first of the graphs directly above), social media leads us both towards more certainty, which is the narrowness of the distribution, and more polarization, which is the left and right extremes of the possibilities (shown in the second graph above).

The result is that – for many of the central issues of our time – most people cannot even talk to each other because there’s very little overlap (or none!) of their perceived distributions of reality.

Why Does This Matter?

Let’s recap the main three points discussed:

1. Predictions and expectations about the future are not binary – they all come together to form some sort of a distribution.

2. There is an actual reality distribution that none of us can know, the shape of which is influenced over time by emergent phenomena.

3. There is an aggregate distribution informed by the best science we have, and there are also a whole lot of individually-perceived opinions about the future that influence our daily actions.

If we want to shift our expectations closer to material reality (and if we hope to make wiser choices as individuals and as a culture), then at least some of us need to converge on a shared, systemic curve. Shown above are two hypothetical mental models of people who might be reading this. Person one might be a well-educated, optimistic individual who believes that through ingenuity, new technology, and hard work, humanity will easily overcome the ecological and economic crises we face. Person two might be considered a ‘catastrophizer,’ overly focused on worst-case scenarios, and generally only imagines a future of extinction for homo sapiens. It’s my hope that through following the wide-boundary, emerging story presented on The Great Simplification, these two individuals with drastically different worldviews might increasingly converge toward a wider, more balanced distribution that captures not just what’s possible, but what’s likely. Because if their distributions overlap somewhere, a conversation can occur.

This is going to be especially difficult as the end of growth will likely lead to deep and existential challenges for democracy, and even truth itself. Science, the best tool we have for understanding the world, struggles in an information ecosystem without a shared language, where conversation often breaks down before it begins.

If we consider the blue line on the above graph to represent an aggregation of all human and ecological events in the future, it may be more dystopian than modern culture and politicians (e.g. the ‘green curve’) are able to understand or willing to admit.

But there are many futures still under that curve. An understanding of the wide array of possibilities might shift your individual focus of interventions, preparations, and pro-social projects beyond what you otherwise would have considered.

For example, if you are quite confident there’s going to be a near-term human extinction (e.g. the ‘red curve’), you’re unlikely to work on projects and policies focused on long-term well-being. In contrast, if you think humans will soon be colonizing Mars, you probably aren’t going to be planting a garden or talking to your neighbors about fostering community resilience.

So if your own personal distribution of possible futures is like the red curve or the green curve above – how certain are you? And how much of what you’re doing today assumes that your expected future is guaranteed to arrive?

My underlying hope in all of this is that by understanding the relationships between these distributions and differing visions of the future, we might be able to break a deeper tension that we discussed earlier in this post: the inverse relationship between what’s good for humans and what’s good for the planet.

Could we imagine a point in the future where some things that are good for the environment are also good for the human system – and vice versa? Ultimately, the purpose of The Great Simplification is to shift the underlying reality toward better odds and outcomes for both of these categories, shown above as moving from curve A to curve B.

I hope some insights made it through this post. If not, you have a better understanding of what goes through my mind as I visualize the future and what’s going on in other people’s minds.

And that – plus $10 – will get you a coffee in New York City.

Thanks for reading, until next time.

---

P.S. The logo for The Great Simplification is partially inspired by the probability distributions discussed in this essay.

---

Want to Learn More?

This essay is adapted from a previous Frankly video titled “Reality, Probability, and Perception.” In the future, we’ll be adapting more Frankly videos to written versions and continuing to post them on Substack, so stay tuned for more.

If you would like to see a broad overview of The Great Simplification in 30 minutes, watch our Animated Movie.

---

If you want to support The Great Simplification podcast…

The Great Simplification podcast is produced by The Institute for the Study of Energy and Our Future (ISEOF), a 501(c)(3) organization. We want to keep all content completely free to view globally and without ads. If you’d like to support ISEOF and its content via a tax-deductible donation, please use the link below.

---

Want to connect with other TGS listeners?

---