Four basic factors … are present in every story, with varying degrees of emphasis. … milieu, idea, character, and event.

• The milieu is the world surrounding the characters, the landscape, the interior spaces, the surrounding cultures the characters emerge from and react to; everything from weather to traffic laws.
• The idea is the information that the reader is meant to discover or learn during the process of the story.
• Character is the nature of one or more of the people in the story – what they do and why they do it. It usually leads to or arises from a conclusion about human nature in general.
• The events of the story are everything that happens and why. …

Each factor is present in all stories, to one degree or another. Every factor has an implicit structure; if that factor dominates a story, its structure determines the overall shape of the story. …

All [these] factors are present in The Lord Of The Rings, but it is the milieu structure that predominates, as it should. It would be absurd to criticize The Lord Of The Rings for not having plot unity and integrity, because it is not an event story. Likewise, it would be absurd to criticize the book for its stereotyped one-to-a-race characters or for the many characters about who we learn little more than what they do in the story and why they do it, because this is not a character story. …

Character stories really came into their own at the beginning of the twentieth century, and both the novelty and the extraordinary brilliance of some of the writers who worked with this story structure have lead many critics and teachers to believe that only this kind of story can be “good.” … [But] other kinds of stories have long traditions, with many examples of brilliance along the way. ….

It is a mistake to think that deep, detailed characterization is an absolute virtue in storytelling. .. If you choose not to devote much time to characterization in a particular stormy, this won’t necessarily mean you “failed” or “wrote badly.” It may mean that you understand yourself and your story. (more; pp.62,63,74,75; see also)

Card suggests that the current high status of character stories is a temporary historical accident, which suggests that it will eventually decline. Someone will write such a damn impressive milieu, idea, or event story that others seeking to look impressive will try their hand, making that structure the ideal of a “good” story.

I’d guess that rising incomes contributed to the rising status of the character story. Rich self-indulgent folks are more likely to be obsessed with their own internal feelings, and our wealth has allowed us the slack to often have dramatically dysfunctional character features. Also, our psychological aversion to seeing ourselves clearly has made those who can overcome such aversions more clearly impressive. However, if our descendants are less rich, less free to change their social roles, or if they can more easily see themselves clearly, character stories may seem less compelling. My weak bet is on the eventual rise in status of the milieu story, as I’ve recently come to see how very hard it can be to describe a coherent yet different world.

Added 8:30p: I went searching for criticism of Card’s framework here, and couldn’t find any. Odd.

Added 7a: On reflection, it is also pretty plausible that increasing density, size, and specialization has only recently created a niche for cognitive elites to write for other cognitive elites, which let writers focus on impressing such elites. Impressively realistic character stories are mostly impressive to other cognitive elites, and much less so to ordinary readers.

Exploring, not the time-bound consequences of human actions, but the timeless implications of known physical law. …. Call it “exploratory engineering”; as applied by Tsiolkovsky a century ago, this method of study showed that rocket technology could open a world beyond the bounds of the Earth. Applied today, this method shows that atomically precise technologies can open a world beyond the bounds of the Industrial Revolution.

Drexler’s most famous book was his ’86 Engines of Creation, but his best was his ’92 Nanosystems, which explored nanotech engineering. The book shows impressive courage, venturing far beyond familiar intellectual shores, impressive breadth, requiring mastery of a wide range of science and engineering, and impressive accomplishment, as little in there is likely to be very wrong. This makes Drexler one of my heroes, and an inspiration in my current efforts to think through the social implications of ems.

Alas, Drexler also deserves some criticism. His latest book, Radical Abundance, like several prior books, goes well beyond physical science and engineering to discuss social implications at length. Alas, though his impressive breadth doesn’t extend much into social science, like most “hard” sci/tech folks Drexler seems mostly unaware of this. He seems to toss together his own seat-of-the-pants social reasoning as he can, and then figure that anything he can’t work out must be unknown to all. Sometimes this goes badly.

Drexler is absolutely right that physical law allows a full nanotech manufacturing capacity that can make physical devices with vastly improved abilities, efficiencies, and costs. Even limiting ourselves to planet Earth, the ultimate limits to our productive capacities exist but are many orders of magnitude beyond today’s capacities. Drexler is also probably right that this manufacturing tech would eventually:

1. reduce relative demand for rare elements,
2. make fractionally less waste in production,
3. shorten supply chains and thus cut long distance physical trade,
4. have factories not much bigger than the products they make,
5. usually have marginal costs near that of required energy, cooling, and raw materials, implying far less variation in such costs per-pound,
6. allow rates of production and retooling far faster than today, and
7. have lower fixed costs of manufacturing capital (implied by 5&6).

These conclusions are non-trivial and useful for estimating social implications. Bravo.

Note that Drexler’s basic tool of exploring the engineering consequences of basic physical laws says little about how fast we will acquire these capacities. He describes a wide landscape of intermediate capabilities, across which we have now traveled a limited distance, so the question is how fast we will venture across the rest of that lanscape.

For example, Drexler’s basic conclusions are consistent with current rates of tech growth continuing steadily on for another few centuries, before finally running up against nanotech limits (at least on Earth). They are also consistent with rates of tech growth falling by a factor of ten and then continuing steadily on for millennia, with per capita income falling to and then staying at subsistence levels.

I’d guess that in the 27 years since Drexler’s first book we’ve come ~10-30% of the way toward full nanotech, with no noticeable acceleration, suggesting we have one to three centuries still to go. I’d love to hear more expert estimates on this.

Contrary to steady growth scenarios, much of Drexler’s further analysis seems to assume that full nanotech will appear rather suddenly, perhaps soon:

Progress has gone surprisingly fast and far, with no barriers in sight. … A Version 2.0 of world civilization, a change as profound as the Industrial Revolution, but unfolding at Internet speeds. … Change in a [nanotech] era could be swift indeed— not stretched out over millennia, like the spread of agriculture, nor over centuries, like the rise of industry, nor even over decades, like the spread of the Internet’s physical infrastructure. … [Nanotech] holds the potential for a physical revolution that, if unconstrained, could unfold at the speed of new digital media. … We can expect to see an accelerating upward spiral of capabilities, with [nanotech]-level technologies as a natural destination.

[Drexler prefers the awkward phrase "atomically precise manufacturing"; I've substituted [nanotech] in the quotes.]

Such suddenness is akin to landscape hikers being picked up by a helicopter and transported to a distant valley. It has Drexler worry about sudden big military power imbalances:

The history of technological competition, whether in nuclear weapons, satellites, computer chips, or stealth aircraft, shows that competitors typically reach technology thresholds at times that differ by years or more. What is different in the realm of [nanotech]-based technology is the potential speed of advanced-stage development and scale up, which could dramatically heighten the resulting disparities in capabilities.

And terrorists:

Unconstrained access to an unconstrained range of [nanotech]-level technologies would place unpredictable capabilities in the hands of hostile non-state actors, leading to unacceptable and unpredictable risks. Preventing unconstrained applications … is thus a vital national interest … and a potential driver for increased worldwide cooperation.

Though Drexler takes great pains to carefully argue that advanced nanotech is possible, he offers no argument for why we should expect sudden nanotech, beyond the fact that with full nanotech one can do lots of things quickly. But the fact that industrial jet planes can circle the world fast today hardly implied that the first nation with any industry could take over the world. As the saying goes, one shouldn’t confuse a clear vision with a short distance.

A sudden appearance of advanced nanotech would imply a vast increase in per person income, at least until population levels could catch up. Drexler thinks most would feel this new income to be “enough,” and so care little about income differences, and have little reason to conflict:

The global prospect would be, not scarcity, but unprecedented abundance— radical, transformative, and sustainable abundance. … First and foremost, broad international access to appropriate [nanotech]-level production capabilities would decrease pressures to compete for access to markets and natural resources simply because there can be no vital interest in resources that are no longer scarce or important, nor a vital interest in export markets once imports and trade balances are no longer essential to material well-being. …

[Nanotech]-based production, however, can enable the material plenty needed for human development while breaking the link between human development and material economic progress. … Today, lack of material development and human development can reinforce one another in a circle that perpetuates poverty. … [Nanotech]-based production will encapsulate complexity in a form that users have no need to examine. The complexity of a cell phone’s computer chip is enormous, yet self-contained; its complexity is beyond the full comprehension of any person alive, yet to use it requires little skill or investment in learning.

As I posted before on Krugman and Keynes, most economists have long since given up on the idea that humans will be satiated when they finally get “enough.” Post-sudden-nanotech folks being lots richer gives us little reason to expect that they won’t work hard to get even more. People and nations with less access to energy, raw materials, and better designs, would still be poor by comparison, and struggle to escape that status.

While Drexler is admirably careful and expert in his exploratory engineering of the capacities of advanced nanotech, when he moves outside his tech expertise he succumbs to seriously wishful thinking in expecting nanotech to come soon and suddenly, and to think that people would feel they had “enough.”

One last nitpick. Drexler says:

For [nanotech]-based production, like information technologies, the crucial labor consists of creative work.

Info tech work is complex and requires high levels of knowledge, experience, and intelligence. That hardly means most of it is “creative”, however. Schools that train people for such careers tend to try to drill the creativity out of them for good reason – people are otherwise inclined to be more creative than is helpful.

The essence of science is inquiry; the essence of engineering is design. Scientific inquiry expands the scope of human perception and understanding; engineering design expands the scope of human plans and results. …

•     Scientists seek unique, correct theories, and if several theories seem plausible, all but one must be wrong, while engineers seek options for working designs, and if several options will work, success is assured.
•     Scientists seek theories that apply across the widest possible range (the Standard Model applies to everything), while engineers seek concepts well-suited to particular domains (liquid-cooled nozzles for engines in liquid-fueled rockets).
•     Scientists seek theories that make precise, hence brittle predictions (like Newton’s), while engineers seek designs that provide a robust margin of safety.
•     In science a single failed prediction can disprove a theory, no matter how many previous tests it has passed, while in engineering one successful design can validate a concept, no matter how many previous versions have failed. ..

Simple systems can behave in ways beyond the reach of predictive calculation. This is true even in classical physics. …. Engineers, however, can constrain and master this sort of unpredictability. A pipe carrying turbulent water is unpredictable inside (despite being like a shielded box), yet can deliver water reliably through a faucet downstream. The details of this turbulent flow are beyond prediction, yet everything about the flow is bounded in magnitude, and in a robust engineering design the unpredictable details won’t matter.  …

The reason that aircraft seldom fall from the sky with a broken wing isn’t that anyone has perfect knowledge of dislocation dynamics and high-cycle fatigue in dispersion-hardened aluminum, nor because of perfect design calculations, nor because of perfection of any other kind. Instead, the reason that wings remain intact is that engineers apply conservative design, specifying structures that will survive even unlikely events, taking account of expected flaws in high-quality components, crack growth in aluminum under high-cycle fatigue, and known inaccuracies in the design calculations themselves. This design discipline provides safety margins, and safety margins explain why disasters are rare. …

The key to designing and managing complexity is to work with design components of a particular kind— components that are complex, yet can be understood and described in a simple way from the outside. … Exotic effects that are hard to discover or measure will almost certainly be easy to avoid or ignore. … Exotic effects that can be discovered and measured can sometimes be exploited for practical purposes. …

When faced with imprecise knowledge, a scientist will be inclined to improve it, yet an engineer will routinely accept it. Might predictions be wrong by as much as 10 percent, and for poorly understood reasons? The reasons may pose a difficult scientific puzzle, yet an engineer might see no problem at all. Add a 50 percent margin of safety, and move on.

The general point is that we quite often collect and act on rather weak info clues. This could make good info sense. We might be slowly collecting small clues that eventually add up to big clues. Or if we know well which parameters matter the most, it can make sense to act on weak clues; over a lifetime this can add up to net good decisions. When this is what is going on, then people will tend to know many things they cannot explicitly justify. They might have seen a long history of related situations, and have slowly accumulated enough relevant clues to form useful judgments, but not be able to explicitly point to most of those weak clues which were the basis of that judgement.

Another thing I noticed on vacation is that a large fraction of my relatives age 50 or older think that they know that their lives were personally saved by medicine. They can tell of one or more specific episodes where a good doctor did the right thing, and they’d otherwise be dead. But people just can’t on average have this much evidence, since we usually find it hard to see effects of medicine on health even when we have datasets with thousands of people. (I didn’t point this out to them – these beliefs seemed among the ones they held most deeply and passionately.) So clearly this intuitive collection of weak clues stuff goes very wrong sometimes, even on topics where people suffer large personal consequences. It is not just that random errors can show up; there are topics on which our minds are systematically structured, probably on purpose, to greatly mislead us in big ways.

One of the biggest questions we face is thus: when are judgements trustworthy? When can we guess that the intuitive slow accumulation of weak clues by others or ourselves embodies sufficient evidence to be a useful guide to action? At one extreme, one could try to never act on anything less than explicitly well-founded reasoning, but this would usually go very badly; we mostly have no choice but to rely heavily on such intuition. At the other extreme, many people go their whole lives relying almost entirely on their intuitions, and they seem to mostly do okay.

In between, people often act like they rely on intuition except when good solid evidence is presented to the contrary, but they usually rely on their intuition to judge when to look for explicit evidence, and when that is solid enough. So when those intuitions fail the whole process fails.

Prediction markets seem a robust way to cut through this fog; just believe the market prices when available. But people are usually resistant to creating such market prices, probably exactly because doing so might force them to drop treasured intuitive judgements.

On this blog I often present weak clues, relevant to important topics, but by themselves not sufficient to draw strong conclusions. Usually commenters are eager to indignantly point out this fact. Each and every time. But on many topics we have little other choice; until many weak clues are systematically collected into strong clues, weak clues are what we have. And the topics of where our intuitive conclusions are most likely to be systematically biased tend to be those sort of topics. So I’ll continue to struggle to collect whatever clues I can find there.

On the surface the processes that produce faster hardware and faster algorithms seem quite different. Hardware is made by huge companies that achieve massive scale economies via high levels of coordination, relying largely on internal R&D. Algorithms instead seem to be made more by many small artisans who watch and copy each other, and mostly focus on their special problem area. How is it that these two very different processes, with very different outputs, both grow at roughly the same remarkably fast rate? The obvious hypothesis is that they share some important common cause. But what? Some possibilities:

• Digital – Both computer hardware and algorithms are digital technologies, which allow for an unusually high degree of formal reasoning to aid their development. So maybe digital techs just intrinsically grow faster. But aren’t there lots of digital techs that aren’t growing nearly as fast?
• Software – Maybe software development is really key to the rapid growth of both techs. After all, both hardware and algorithm experts use software to aid their work. But the usual descriptions of both fields don’t put a huge weight on gains from being able to use better productivity software.
• Algorithms - Maybe progress in hardware is really driven behind the scenes by progress in algorithms; new algorithms are what really enables each new generation of computer hardware. But that sure isn’t the story I’ve heard.
• Hardware - Maybe there are always lots of decent ideas for better algorithms, but most are hard to explore because of limited computer hardware. As hardware gets better, more new ideas can be explored, and some of them turn out to improve on the prior best algorithms. This story seems to at least roughly fit what I’ve heard about the process of algorithm design.

This last story of hardware as key has some testable predictions. It suggests that since gains in serial hardware have slowed down lately, while gains in parallel hardware have not, parallel algorithms will continue to improve as fast as before, but serial algorithm gains will slow down. It also suggests that when even parallel hardware gains slow substantially in the future, because reversible computing is required to limit power use, algorithm gains will also slow down a comparable amount.

If true, this hardware as key theory also has policy implications. It suggests that it is much better to subsidize hardware research, relative to algorithm research; even with less research funding algorithm gains will happen anyway, if just a bit later. This theory also suggests that there is less prospect for self-improving algorithms making huge gains.

So what other explanations can we come up with, and what predictions might they make?

Added 5June: There are actually several possible ways that software progress might be determined by hardware progress. In the post I mentioned better hardware letting one explore more possible ideas, but it could also be that people already knew of better algorithms that couldn’t work on smaller hardware. Algorithms vary both in their asymptotic efficiency and in their initial overhead, and we might slowly be switching to bigger overhead algorithms.

Those promised quotes:

The White House advisory report cited research, including a study of progress over a 15-year span on a benchmark production-planning task. Over that time, the speed of completing the calculations improved by a factor of 43 million. Of the total, a factor of roughly 1,000 was attributable to faster processor speeds. .. Yet a factor of 43,000 was due to improvements in the efficiency of software algorithms.

Solve some test LPs … hardware seems to be currently winning the speed race. Even so, studies suggest that over time, say 20 years, the speed-ups have been quite evenly divided between mathematics/algorithms and hardware improvements.

25 years progress solving sparse linear systems. Hardware alone: 4 orders of magnitude. Software alone: 6 orders of magnitude. The N-Body Problem: in 30 years 10^7 hardware, 10^10 software.

Added 5June: roystgnr points us to a similar result in magnetohydrodynamics.

These quotes are from a forthcoming MIRI tech report by Katja Grace currently titled “Assorted evidence on algorithmic progress.”

Chess programs … over the last four decades. … Estimates for the significance of hardware improvements are very noisy, but are consistent with hardware improvements being responsible for approximately half of progress. … Go programs … for the last three decades. Hardware doublings produce diminishing Elo gains, on a scale consistent with accounting for around half of progress. … Improvements in a variety of physics simulations (selected after the fact to exhibit performance increases due to software) appear to be roughly half due to hardware progress.

The puzzle is that in most of our social worlds we instead focus on rewarding the best, not punishing the worst. If you search for “punish reward” you will mostly find the issue raised about how to treat kids; we are mainly willing to use punishment flexibly on them. And this when young kids are the main exception – for them punishment works worse. For adults, we tend to limit punishment’s use to extreme behavior that we all strongly agree is bad, like crime. And when you ask adults, they much prefer to be part of a group that uses rewards, not punishment.

As a college teacher, I expect that I’d get more effort from most students by regularly pointing out the worst student in the class than the best. But I also expect students to hate it and give me low evaluations. Similarly, I expect that if I wrote the occasional post criticizing a bad blog commenter here, instead of praising a good one, I’d get more change in commenting behavior. But I also expect that person to complain long and loud about how I was biased and unfair, and others to come to their defense. I expect a lot less complaining about bias in picking the best.

In both the class and comment cases, I expect people to see me as mean and cruel for punishing the worst, but kind and generous for rewarding the best. This even though all of these effects are relative – punishment would raise the rest of the class, or the rest of the commenters, up above the worse.

Note that rewarding the best is in practice more elitist than punishing the worse; punishing creates an underclass, not an overclass. And in fact our hyper-egalitarian forager ancestors were quite reluctant to overtly reward or praise; they focused their social coordination on having the group punish norm violators. Our hyper sensitivity to being punished, and our elaborate instinctual strategies to give excuses and to coordinate to retaliate against any who might suggest we should be punished, are probably human adaptations to that forager history. And they make us especially unwilling to accept punishment by an authority, instead of by the informal consensus of the group.

This seems an interesting example of our seeking to avoid aspects of the forager way of life. Our forager evolved aversion to being singled out for social shame is so strong that we’d rather create elites instead. At least this applies when we are relatively rich and comfortable. If we really feared being destroyed for lack of sufficient efforts, as farmers often did, we’d probably be a lot more eager to raise overall efforts by punishing the worse. I suspect that foragers themselves didn’t punish much in good times; punishment was invoked more, and mattered more, in hard times. In good times foragers probably more tolerated praising some as better, and weak forms of bragging.

In a more competitive future, with organizations and individuals that compete harder to survive, I’d expect more use of punishment, in addition to reward.

Today if you have a group that really needs to succeed, and to induce strong efforts all around, consider paying the social disruptions costs of punishing the worst, instead of rewarding the best. You will probably get more effort that way, even if people end up hating you and calling you evil for it. And if your group doesn’t punish and fails, know that your reluctance to punish was probably a contributing factor.

Those promised quotes:

A study of 150 public-school teachers … split the teachers into two groups and told both that their bonuses would be linked to student test scores. Teachers in the first would receive a bonus at the end of the year if student test scores improved. Members of the second group received a check for $4,000 in September and agreed to return the money if test scores failed to rise by June. Loss aversion worked: Teachers who faced the threat of having to refund their bonuses produced student test scores that were about 7 percentage points higher. (more)

Performance improved substantially more when the feedback was positive in the case of the younger children, telling them they did well when they did, rather than negative, telling them that they did poorly when they did. Just the opposite proved true in the case of older children, who functioned just like young adults aged 18-25 who were also tested. That is, negative feedback improved performance more for these individuals than did positive feedback. (more)

In a laboratory experiment, we have implemented three main incentive mechanisms: reward-only, punishment-only, and reward and punishment. We have also varied the size of the tournament. Although the baseline model predicts that employee effort should be the same in all treatments, our empirical results have indicated that this is not the case. In general, no mechanism generates higher effort levels from the agents than the one which combines reward and punishment. We have also found that punishment produces similar results to the combined mechanism in tournaments of a relatively small size (three participants), while the reward-only and punishment-only mechanisms are equivalent in terms of effort in tournaments of a relatively large size (six participants). … Social dilemma … experiments have [also] found that punishment is generally more effective than rewards in terms of promoting cooperation between subjects in voluntary contribution games. (more)

It’s much easier to change an established behavior by offering rewards, rather than threatening with punishments. … But once you have established a norm, sustaining it by the threat of punishment is cheap. Only a few people will violate the norm, so you will rarely have to follow through with your threat. (more)

[We] extend the standard public goods game in a variety of ways, in particular by allowing for endogenous preference over institution, … [We] elicit preferences in an incentive-compatible manner over voluntary contribution mechanisms with and without reward and punishment options. Finally, [we] randomly assign subjects to one of the four institutions and observe repeated play. They [we] that payoffs are significantly greater when punishment is allowed but that only a small minority of participants prefers such an environment. (more)

I’ve gotten a lot of informal complaints that this approach is badly overconfident, unscientific, and just plain ignorant. Don’t I know that any particular forecasted combo is very unlikely to be realized? Well yes I do know this. But I don’t think critics realize how robust and widely used is this best combo approach.

For example, this is the main approach historians use studying ancient societies. A historian estimating Roman Empire copper trade will typically rely on the best estimates by other experts on Roman population, mine locations, trade routes, travel time, crime rates, lifespans, climate, wages, copper use in jewelry, etc. While such estimates are sometimes based on relatively direct clues about those parameters, historians usually rely more on consistency with other parameter estimates. While they usually acknowledge their uncertainty, and sometimes identify coherent sets of alternative values for small sets of variables, historians mostly build best estimates on the other historians’ best estimates.

As another example, the scheduling of very complex projects, as in construction, is usually done via reference to “baseline schedules,” which specify a best estimate start time, duration, and resource use for each part. While uncertainties are often given for each part, and sophisticated algorithms can take complex uncertainty dependencies into account in constructing this schedule (more here), most attention still focuses on that single best combination schedule.

As a third example, even when people go to all the trouble to set up a full formal joint probability distribution over a complex space, as in a complex Bayesian network, and so would seem to have the least need to crudely avoid complexity by focusing on just one joint state, they still quite commonly want to compute the “most probable explanation”, i.e., that single most likely joint state.

We also robustly use best tentative combinations when solving puzzles like Sudoku, crossword, or jigsaw. In fact, it is hard to think of realistic complex decision or inference problems full of interdependencies where we don’t rely heavily on a few current best guess baseline combinations. Since I’m not willing to believe that we are so badly mistaken in all these areas as to heavily rely on a terribly mistaken method, I have to believe it is a reasonable and robust method. I don’t see why I should hesitate to apply it to future forecasting.

Individuals with an accessible independent self-view (a characteristic of members of most Western cultures) place high values on self-reliance and autonomy. They strive toward being unique, different, and separate from others. Of key importance to the independents is the “inner core” of the self—internal attributes and traits that are enduring and invariant over time and context. In contrast, individuals with a more accessible interdependent self-view (a characteristic of members of many Eastern cultures) value relationships with others and interpersonal harmony. They view the self as part of a social group and strive toward blending and fitting in. …

There are reasons to believe that the two distinct self- views are associated with different levels of construal and psychological distances. First, interdependents are concerned about relationship harmony and are sensitive to the interconnectedness between people and events. From this perspective, it is both desirable and necessary that they pay close attention to the immediate environment to ensure that relationship harmony is attained and preserved. This attention to the “here” and “now” likely prompts a low-level construal and its corresponding proximal temporal perspective. Second, feelings of agency and control may also lead to higher construal levels among those with an independent self-view. (more)

This suggests that westerners tend to think more in a far view, which suggests that they are more idealistic, plan further into the future, are more socially inclusive, and think more via analogy.

Today, secrets are lumpy. You might know some info that would help you persuade someone of something, but reasonably fear that if you told them, they’d tell others, change their opinion on something else, or perhaps just get discouraged. Today, you can’t just tell them one implication of your secret. In the future, however, the ability to copy and erase minds (as in am em scenario) might make secrets much less lumpy – you could tell someone just one implication of a secret.

For example, what if you wanted to convince an associate that they should not go to a certain party. Your reason is that one of their exes will attend the party. But if you told them that directly, they would then know that this ex is in town, is friendly with the party host, etc. You might just tell them to trust you, but what if they don’t?

Imagine you could just say to your associate “I could tell you why you shouldn’t go to the party, but then I’d have to kill you,” and they could reply “Prove it.” Both of your minds would then be copied and placed together into an isolated “box,” perhaps with access to some public or other info sources. Inside the box the copy of you would explain your reasons to the copy of them. When the conversation was done, the entire box would be erased, and the original two of you would just hear a single bit answer, “yes” or “no,” chosen by the copy of your associate.

Now, as usual, there are some complications. For example, the fact that you suggested using the box, as opposed to just revealing your secrets, could be a useful clue to them, as could the fact that you were willing to spend resources to use the box. If you requested access to unusual sources while in the box, that might give further clues.

If you let the box return more detail about their degree of confidence in their conclusion, or about how long the conversation took, your associate might use some of those extra bits to encode more of your secrets. And if the info sources accessed by those in the box used simple cacheing, outsiders might see which sources were easier to access afterward, and use that to infer which sources had been accessed from in the box, which might encode more relevant info. So you’d probably want to be careful to run the box for a standard time period, with unobservable access to standard wide sources, and to return only a one bit conclusion.

Inside the box, you might just reveal that you had committed in some way to hurt your associate if they didn’t return the answer you wanted. To avoid this problem, it might be usual practice to have an independent (and hard to hurt) judge also join you in the box, with the power to make the box return “void” if they suspected such threats were being made. To reduce the cost of using these boxes, you might have prediction markets on what such boxes would return if made, but only actually make them a small percentage of the time.

There may be further complications I haven’t thought of, but at the moment I’m more interested in how this ability might be used. In the world around you, who would be tempted to prove what this way?

For example, would you prove to work associates that your proposed compromise is politically sound, without revealing your private political info about who would support or oppose it? Prove to investigators that you do not hold stolen items by letting them look through your private stores? Prove to a date you’ve been thinking lots about them, by letting them watch a video of your recent activities? Prove to a jury of voters that you really just want to help them, by letting them watch all of your activities for the last few months? What else?

In general, this would seem to better enable self-deception. You could actually not know things anywhere in your head, but still act on them when they mattered a lot.