Difference between revisions of "Technical Aspects of Future Predictions"

From Lunarpedia
Jump to: navigation, search
m
Line 1: Line 1:
Whenever we talk about the Moon, we are most often trying to predict the future.  There are a number of technical approaches we can take to do this job that are far more effective than just guessing or looking into crystal ball.  Here are the processes that were used to predict the future for the paper [Future Work for NASA].
+
Whenever we talk about the Moon, we are most often trying to predict the future.  There are a number of technical approaches we can take to do this job that are far more effective than just guessing or looking into crystal ball.  Here are the processes that were used to predict the future for the paper [[Future Work for NASA]].
  
 
Future prediction is at best guesswork.  Technical knowledge does, however, provide certain limited advantages in prediction.
 
Future prediction is at best guesswork.  Technical knowledge does, however, provide certain limited advantages in prediction.

Revision as of 11:23, 26 May 2008

Whenever we talk about the Moon, we are most often trying to predict the future. There are a number of technical approaches we can take to do this job that are far more effective than just guessing or looking into crystal ball. Here are the processes that were used to predict the future for the paper Future Work for NASA.

Future prediction is at best guesswork. Technical knowledge does, however, provide certain limited advantages in prediction.


Why predict the future?

Figure 1, Diagram for problem solving


Away from pain, toward pleasure

People can be motivated either to move away from pain or to move toward pleasure. The easiest thing to do is to first tell people a scary story and then tell them what they must do to get away from the pain it causes. This move-away-from-pain approach works well in the short term but wears people out. Fear does not keep people in action for the long haul.

The second approach is to tell people a positive story about the future and then generate buy-in in them for actions to get there. This approach works the best for the long haul, but only if the vision of the future has a solid foundation in reality. Although the vision need not be too detailed, when the going gets tough the vision must hold together against determined critics. The vision of the future must be supported by substantial evidence and the plan to get there must be well thought through. Both the vision and the plan must hold up to criticism over time.

Most of the problems of the 21st century have so far been approached with move-away-from-pain ideas. This approach can only go so far.

To use the more effective move-toward-pleasure approach we must develop sound visions of positive futures. These visions must first face the problems squarely and then help people get into actions that can reasonably generate a pleasant future even if in the end it is not exactly the one envisioned. The problems cannot be denied to exist, fortunately it is quite acceptable for the actions to be very challenging to the people.

The problem then is to take a realistic view of each problem, envision a believable future in which the problem is addressed, and then devise a series of actions to get there. You must keep a continuous watch on the problem, but you do so by looking over your shoulder. While you are in action, you must you keep your head looking forward, you keep your eyes on the prize.

The myriad actions developed for individual problems then must be coordinated into to an overall plan to cover all the problems.

The Side of the Angles, Not

An extreme, and powerful, version of move-away-from-pain exists, but is not suitable for use in developing a sustainable future. If one group of people defines another group as the most evil people who have ever walked the Earth and defines their own stand as the one true good, then the members of that group can draw truly enormous amounts of personal energy from being on the side of the angles. This is the effect that supported the Cold War and was effectively used by both sides for decades.

The side-of-the-angles effect requires two simplistic models of the future, one of the evil that will happen if the other side wins and another of the sunny future if we win. Only five to twenty percent of a population is likely to buy into this model but a much larger number may go along with it particularly as the hardcore people have such energy and power.

This effect simply cannot provide a positive future for all as there must be an evil group who looses and this group must be made up of people. It is therefore not suitable for developing a sustainable future.

Generating Buy-in

How did the great pyramids get built? How could anybody get a large number of people to take on such a daunting project? Buy-in is how. To understand which large projects will go forward and which ones will fall by the wayside, we must understand the very human process of buy-in.

All large projects done by humans came into existence through language. Somebody quite literally talked them into existence. Apollo to the Moon is a clear example of this effect, in that the person is known, President Kennedy, and the speech is famous “We shall go to the Moon”. Technical people refer to this process as generating buy-in and we now have a basic understanding of how this language process operates in the human brain.

Most technical people have a gut feel for the concept of buy-in. If you experience a buy-in, you hear an idea, you comprehend the idea, you envision yourself succeeding with the idea, you express your support for the idea, and you get in action on the idea. For most people the buy-in state of mind includes a clear daydream of themselves succeeding with the project.

The structures in the human brain responsible for buy-in are beginning to be understood, but we have not really nailed them yet. We have located the specific bit of brain that generates the eureka state of mind, which is similar to the buy-in trigger but is probably independent. Developments in understanding the modules in the human brain are moving forward very quickly. I am confident that we will soon nail buy-in.

Generating buy in is dependent on having a clear, top level design that will generate a vision of success in the listeners. Those grand ideas and projects that easily support the generation of buy-in in the general population are the ones most likely to be completed. These ideas make it easy for anyone to envision themselves and their nation succeeding. These are the ideas we need to give weight to in our predictions.

One purpose of this paper is to help define the top of NASA’s designs which are critical to generating buy-in for our projects. In 1989 we talked about returning to the Moon. Our design then failed at the top. It was too expensive; it did not suit Congress’s needs; and it did not generate buy-in in either Congress or public. We are now involved in another return to the Moon design. Public support for the idea is now disappointingly low. If our top level design does not meet true needs of today, it will not generate buy-in and it will fail just like the 1989 effort.

Also, strong buy-in is necessary but not sufficient for an idea to affect society and be given weight in this analysis. Ideas with weak buy-in rarely affect society, but ideas with strong buy-in but weak technical value also fail. Buy-in must be supported by value to be effective.

Talking into Existence

The buy-in idea has a somewhat unexpected consequence. All the great projects of the human race came into existence by being talked into existence. From the great pyramids, through the Gothic cathedrals, to Apollo to the Moon, the key to making big things happen is language. You can draw pictures to get the idea across but to get people into action and to keep them in action you must express your idea in language. It may be written or verbal, in person or recorded, but it must be in words.

Look at the example of President Kennedy. He needed a project that would focus the energies of the American people. He did not have one already chosen, so he talked over several ideas with his advisors. A number of projects were suggested and several quick studies were prepared to get estimates of cost and workable timelines. His team then chose a goal and set the date for completion at 1967, but he later rounded up to the end of the decade simply because the phrase, “by the end of this decade,” had a nice ring to it.

On a cold fall day, September 12, 1962, Kennedy delivered what was originally scheduled to be a short routine speech at Rice University in Houston Texas. He was standing at an outside podium in the football field; the sky was overcast and gray. He had no visual aides beyond his bare hands. There was TV coverage for the nightly news, but it was only routine coverage.

He spoke, “We choose to go to the Moon in this decade, and do the other things, not because they are easy, but because they are hard; …”

The speech lasted less than 18 minutes but a nation was listening. The complete buy-in process actually took a couple more weeks with several repetitions of the speech to growing TV coverage.

A nation bought into the idea. It mattered not how outlandish the idea actually was. He gave his nation a powerful vision of success that it wanted very badly. Several million people bought in and then got into action. Millions more watched and supported the effort.

We went to the moon. We now need to give extra weight to ideas that have proponents who can powerfully elucidate their virtues.

Accidents don't happen

The reverse of the talking into existence idea is important too. People fear that just talking about a bad possibility might accidentally make it happen so we should not even mention them. Such an accident is extremely unlikely. It takes a lot of work to setup and execute a buy-in. Just mentioning something is rarely enough to get the ball rolling. The environmental movement spends lot of time talking about the bad effects of pollution with little effect. It is only when actions that can lead to success are talked about to people get into action.

We need to talk through the bad things that could happen. Only then can we arrive at effective actions to address the problems.

Modeling Surprise

We are not really very interested in the future if that future is exactly what we expect it to be. It is only when something happens, good or bad, that is different from our expectations that our interest is peaked. What we need is a program that warns us of surprises in time to take action.

For a forecast to warn us of surprises, it must first develop a base model of what we are expecting to happen and then a way to be flagged when reality deviates from that expectation. The what-we-are-expecting task requires only a moderately good forecast while the spotting-surprises task requires much more detail predictions, especially in areas that have surprised us before, like the sudden appearance of the Web.

Top-Down versus Bottom-Up Design

In a real sense we are here designing the future of NASA. There is a school of technical design that says you must start a design at the bottom and work up. This school pays close attention to the properties of materials and new technologies. There is another school of design that says you must start at the top and work down. This school pays attention to the environment and customer needs.

Bottom-up design can often produce clever devices that are of little use to real people, like the Segway. Top-down design tends to produce pie-in-the-sky ideas that simply cannot be built, such as floating cities in the sky. In truth, the best approach is to start at both ends and work to the middle. It is only then that true needs mesh with new technologies, and as a result, society changes. For example, obscure improvements in digital communication opened up a possibility, the need of humans to communicate meshed with that opening and the Internet was born.

To do a successful top-level design, you must have a very good idea of the whole environment it must function in and of the needs it must fulfill. This paper may be fairly judged on how well it meets this requirement for defining the top of our NASA mission design.

How do you go about predicting the future?

Complex Models on Super Computers versus Desk Top Computers

The modern way to predict the future is to build a complex and sophisticated mathematical model, input massive amounts of data, and run it on a super computer. NASA is a leader in this field for environment and climate. Other companies and institutions run proprietary models in many fields such as economics and energy.

This is the state of the art in future prediction, but it is very expensive. Clear applications of great economic value must be clearly seen to justify the cost. Many more of these systems exist than are available to public discussion as most corporations keep theirs confidential. When the results of these large systems are available we need to seek them out and use them. When they are not we are forced to look elsewhere.

The alternative is to use much simpler models run on modest computer systems. These are comparatively simple and cheap so they tend to be public, but by their nature they are gross simplifications. Much work at this level must be done before the cost of the huge systems can be justified. The results of these small systems often come in the form of graphs, as seen in Section 3. Where the large models are not available we must fall back on the best of the small system efforts.

Figure 2, Feedback Curves

Simple Feedback System Responses

Most of the complex systems that make up our world exhibit feedback. A small portion of the output has an effect on the input. The feedback can be positive, which often causes oscillations, or it can be negative, which can help stabilize the system. A positive feedback example is: increased carbon dioxide in the atmosphere heats the oceans which then release more greenhouse gases. A negative feedback example is: an excess of predators results in reduced numbers of prey animals, which in turn, results in a reduced number of predators.

Complex feedback systems can often be modeled reasonably well with simple feedback systems on modest computers. Simple feedback systems often produce graphs with a common appearance that are easy to recognize. Complex systems often produce similar graphs, but with the complications of noise or multiple graphs stacked on top of each other. Recognizing these common graph forms helps us make practical predictions.

The two most common feedback responses are shown in Figure 2. In the first the system is fed a change in input to a specific level that continues indefinitely; this is called a unit step function. The system response is to rise, overshoot, and settle to a new value. This pattern can occur when a new sustainable resource becomes available, like arable land.

The second graph shows the system response to a single pulse of defined area. The system response first rises to a peak and then settles back to the original value. The area under the curve is proportional to the area in the original pulse. This pattern can occur when a resource fixed in total amount becomes available, like oil.

We will need to keep an eye out for these patterns; spotting them can be a great aid in prediction.

Figure 3, General Exponential Growth

Follow Exponential Growth

As shown in Figure 3 above, the concepts that define our technical society can grow exponentially, but only for a time. Most such parameters grow slowly at first, but a few take off and change our world. Later the idea reaches a technical limit and flattens out again to slow growth. Most effects never achieve exponential growth; those that do we need to recognize and then follow carefully. A number of examples of this effect are considered in Section 3.

Exponential growth has three repeating stages. In the first stage, the growth is very slow and appears linear. In the second stage, it takes off and rises in an exciting and sometimes frightening manner. It the third stage, it exhausts some key resource and returns to a slower, more linear growth. This effect is common enough to have a name; it is an “S” Curve and is often the first section of the Unit Step Response curve described above.

Often an important technology will run out of steam in its current version but will have new technologies standing by in the laboratory to take over. In this way its “S” Curve can be extended for decades. One good example is sound recording. It started slowly with wax cylinders. Soon it moved on to plastic disks, 78's. Then it moved on to vinyl stereo LP's. Then it moved on the CD's. And now the leader is direct information downloads. Along the way there were 45s, eight-tracks, and cassettes that never made it big. We need to spot this type of growth and learn to tell the winners, like CD’s, from the losers, like eight-tracks.

To predict our future work, we are primarily trying to predict the future growth of a number of technologies and from that develop a general picture of the future of society as a whole. Predicting where specific technologies are on the “S” Curve is then a critical exercise. The ones now in exponential growth, or likely to enter it soon, are the ones we must pay attention to. If a technology is in linear growth and shows no signs of a breakthrough, we must not assume it will breakthrough, no mater how much we need it to do so.

Our task then is to identify technologies currently in exponential growth, estimate how long this growth will be sustained, and look to the laboratories to see if new technologies are set to keep the trend going.

Why is the future predictable at all?

The Universe is Probabilistic, Not Deterministic

The future does not now exist. It is not determined in advance. Nothing is certain to happen although many things are likely to happen. The sun will almost certainly rise in the morning. The weather a month from now is much more open to variation and is not now set.

We have started understanding the probabilistic nature of the universe with Heisenberg's Principle early in the 20th century. Einstein spent the last thirty years of his life trying to prove the universe is deterministic and failed. His famous quote, “God does not play dice,” has proved to be wrong.

This probabilistic quality has major implications for predicting the future. We can make precise predictions for only a short period of time. We can then make more general predictions for longer periods. We can never make exact predictions except by pure chance.

Predicting climate is one thing NASA does and does well. As long as NASA is in the climate prediction business, we are in the future prediction business. We will be better at the future prediction job if, first, we admit that we are in this business, and second, we look closely at the underlying assumptions we are making in doing this task. The value of accurately predicting the future is so great that we have no choice but to use all the technical theory and information we can muster.

Complex systems at the edge of Chaos

Common simple feedback systems are often stable and require a significant disturbance to move them from one output state to another. We know and depend on this effect. Complex systems can behave surprisingly differently.

Many complex systems, particularly those involving life like the ones we are dealing with here, operate at the edge of chaos. If their internal feedback paths were just a little bit more positive, their output would be chaotic or would oscillate. Many living systems take advantage of being at the edge of chaos so that they can make maximum use of the free energy and random events in their environment.

When a system is near the edge of chaos it can change state as the result of a very small event in its ever changing internal states. A small event can then cascade into large events, and the output of the entire system can radically change. This is sometimes called the Butterfly Effect and is usually evoked to present the idea that a small apparently random event, say the flapping of a butterfly’s wing, could have an unexpectedly large effect like a wind storm a continent away. This particular vision is of no practical use because it is impossible to determine which small effect will grow among the great majority of effects that diminish.

For this analysis we will use the concept quite differently. We will take the Butterfly Effect to mean that a very large number of small positive actions taken by many people can result in the later solution of great problems. A few actions are unlikely to generate a winning hit. No single great action or leader can be relied on to be sufficient. It will take a very large number of positive actions both to hit the random ones with large positive effects and to overpower any negative effects caused by chance. This approach is consistent with the positive results generated by recent movements that changed society, like the American civil rights movement.

At this critical juncture that is the 21st century, our achievable actions now can have large positive effects on all future generations. Under this view, Dr. Martin Luther King, Jr. did not invent the civil rights movement; the civil rights movement invented Dr. King.

What does this all means to us?

None of the Above

Most future predictions are simply pie-in-the-sky daydreams that show nothing more than the author's hopes and personal buy-in. These rarely have practical value. For our predictions to serve a useful purpose we must be very harsh in eliminating this type of wishful thinking.

If an idea, even one we like, fails to fit into any of the selection categories above, we must demote it to be only a minor effect in our forecast, an eight-track player. It simply does not matter how much we need the idea or how much we might like for the idea to take off, we simply cannot use it. Flying cars have great buy-in but are worthless in all the ways that really count. We must crash and burn flying cars. This restriction is our primary defense against building a future on a foundation of wishful thinking.

The hypothesis of this paper

From all this we can develop a working hypothesis on future prediction. Future prediction is:

Possible, but only within limits

  • A heavy user of both manpower and computer power
  • Very difficult
  • Very controversial
  • And, absolutely necessary

(End of the appendixes)


Supportive readings:

James Gleick, Chaos: the Making of a New Science (Penguin, 1988)

M. Mitchell Waldrop, Complexity: The Emerging Science at the Edge of Order and Chaos (Simon, 1992)

Ray Kurzweil, The Singularity is Near (Penguin, 2006)

Lunarpedia, “Buy-In Explained”, [[1]]

Diego Gambetta, Steffen Hertog, “Engineers of Jihad” (University of Oxford, 2007) [[2]]