Monthly Archives: September 2017

Humans Cells In Multicellular Future Minds?

In general, adaptive systems vary along an axis from general to specific. A more general system works better (either directly or after further adaptation) in a wider range of environments, and also with a wider range of other adapting systems. It does this in part via having more useful modularity and abstraction. In contrast, a more specific system adapts to a narrower range of specific environments and other subsystems.

Systems that we humans consciously design tend to be more general, i.e., less context dependent, relative to the “organic” systems that they often replace. For example, compare grid-like city street plans to locally evolved city streets, national retail outlets to locally arising stores and restaurants, traditional to permaculture farms, hotel rooms to private homes, big formal firms to small informal teams, uniforms to individually-chosen clothes, and refactored to un-refactored software. The first entity in each pair tends to more easily scale and to match more environments, while the second in each pair tends to be adapted in more detail to particular local conditions. Continue reading "Humans Cells In Multicellular Future Minds?" »

GD Star Rating
loading...
Tagged as: ,

Prediction Markets Update

Prediction markets continue to offer great potential to improve society at many levels. Their greatest promise lies in helping organizations to better aggregate info to enable better key decisions. However, while such markets have consistently performed well in terms of cost, accuracy, ease of use, and user satisfaction, they have also tended to be politically disruptive – they often say things that embarrass powerful people, who get them killed. It is like putting a smart autist in the C-suite, someone who has lots of valuable info but is oblivious to the firm’s political landscape. Such an executive just wouldn’t last long, no matter how much they knew.

Like most promising innovations, prediction markets can’t realize their potential until they have been honed and evaluated in a set of increasingly substantial and challenging trials. Abstract ideas must be married to the right sort of complementary details that allow them to function in specific contexts. For prediction markets, real organizations with concrete forecasting needs related to their key decisions need to experiment with different ways to field prediction markets, in search of arrangements that minimize political disruption. (If you know of an organization willing to put up with the disruption that such experimentation creates, I know of a patron willing to consider funding such experiments.)

Alas, few such experiments have been happening. So let me tell you what has been happening instead.

For example, some public markets, such as PredictIt continue to function, and others, such as InTrade, have gone away. While I wish such markets well, I’m not that optimistic about markets on broad public questions, sold to the public, relative to markets on specific organization questions, funded by those organizations. I’m skeptical that there is much public demand for betting on odd questions, or that such markets do much to promote the more promising organizational markets.

Academics and their patrons have showed continued interest, but mostly in the form of abstract efforts such as papers, theorems, and lab experiments. We continue to collect an overhang of abstractly promising mechanisms that haven’t been tried in real organizations. (Including my combinatorial prediction markets.)

A few books have reached wide popular audiences, but mainly by focusing on signs that readers can use to tell themselves that they are better forecasters than their rivals. For example, The Wisdom of Crowds lets ordinary people tell themselves that those damn self-appointed experts are over-rated, compared to we wise crowds. Superforecastors gives several other indicators that readers can collect to tell themselves they are better. While the authors of these books clearly favor creating more markets, they are aware that this isn’t the readers’ priority.

In business, the firms who a decade ago tried to sell straight prediction market software and services to firms for their key decisions have mostly either gone out of business or switched to safer related products. These related products tend to keep the appearance of markets but undercut their main incentive benefits, and they tend to stay away from topics on which key firm insiders might express opinions.

For example, “innovation markets” suggest new research or development projects for firms to pursue. But instead of betting on the consequences of starting such projects, which would take years to see, they just bet on which projects will be funded. And
prediction market research” replaces focus groups with an interface where users seem to bet on which products will be more popular. Except they are just taking a survey, not actually betting on real outcomes.

The last few years has seen great interest in “blockchain” technology and ventures. While these are often used for illegal purposes, authorities have not cracked down as much as they might, as blockchains are new and sexy and promise many non-illegal and useful applications. However, these other applications have been slow in coming. I expect that if blockchains do not soon deliver a majority of their activity as legal laudable applications, authorities will crack down. And while hardcore fans may do what it takes to continue to use them even in the face of such a crackdown, most users will cave and quit, resulting in far lower activity levels.

Many firms have recently issued “crypto coins” to support their blockchain efforts, and current market prices suggest that speculators see a substantial (even if low) chance of large future blockchain activity levels. Even if the underlying technology has promise, such market prices can still be in error in either direction, and it is my opinion that such prices are now too high.

Prediction markets are one of the most frequently mentioned blockchain applications, and I’ve advised several related ventures. As sports betting seems one of the most likely uses of such blockchain based prediction markets, it isn’t clear that activity in this area will count as the legal laudible applications that blockchains need to survive. Even so, many are pursuing this possibility.

Overall blockchain based ventures tend to be heavy on algorithms and software, and light on all the other inputs needed to make a successful business venture. Many of these ventures seek to create general platforms, and hope that other more specific ventures will fill in the more specific business details.

The first firm to issue a blockchain prediction market coin was Augur, two years ago. They still haven’t delivered their software product, but they say they are close, and hopefully extra quality and reliability will result from their extra effort.

Gnosis issued their coin back in April, and they also say they are near ready to deliver their software. They also plan to do a set of experiments to test decision market variations. Both Augur and Gnosis are focused on creating general platforms, expecting others to fill in the specific betting topics and to do the marketing to attract specific customers.

Stox issued its prediction market coin last month. They say they plan more to team with existing trading websites: ”Stox incentivizes other industry leaders with existing customer bases, like invest.com, to join the Stox network and drive traffic to the network.” But they have yet to announce specific teaming deals.

Enjin is issuing its coin now. Instead of supporting prediction markets, Enjin says its coin makes it easier to trade assets from games like Minecraft, even without the support of the makers of such games.

I wish all these ventures well, though I fear a blockchain price crash is coming soon, and I wish there was more of a focus on selling organization over amateur prediction markets, and on particular business applications, rather than general software platforms. But it isn’t yet too late for someone to start to focus there.

GD Star Rating
loading...
Tagged as:

Prepare for Nuclear Winter

If a 1km asteroid were to hit the Earth, the dust it kicked up would block most sunlight over most of the world for 3 to 10 years. There’s only a one in a million chance of that happening per year, however. Whew. However, there’s a ten times bigger chance that a super volcano, such as the one hiding under Yellowstone, might explode, for a similar result. And I’d put the chance of a full scale nuclear war at ten to one hundred times larger than that: one in ten thousand to one thousand per year. Over a century, that becomes a one to ten percent chance. Not whew; grimace instead.

There is a substantial chance that a full scale nuclear war would produce a nuclear winter, with a similar effect: sunlight is blocked for 3-10 years or more. Yes, there are good criticisms of the more extreme forecasts, but there’s still a big chance the sun gets blocked in a full scale nuclear war, and there’s even a substantial chance of the same result in a mere regional war, where only 100 nukes explode (the world now has 15,000 nukes).

I’ll summarize this as saying we face roughly a one in 10,000 chance per year of most all sunlight on Earth being blocked for 5 to 10 years. Which accumulates to become a 1% chance per century. This is about as big as your one in 9000 personal chance each year of dying in a car accident, or your one in 7500 chance of dying from poisoining. We treat both of these other risks as nontrivial, and put substantial efforts into reducing and mitigating such risks, as we also do for many much smaller risks, such as dying from guns, fire, drowning, or plane crashes. So this risk of losing sunlight for 5-10 years seems well worth reducing or mitigating, if possible.

Even in the best case, the world has only enough stored food to feed everyone for about a year. If the population then gradually declined due to cannibalism of the living, the population falls in half every month, and we’d all be dead in a few years. To save your family by storing ten years of food, you not only have to spend a huge sum now, you’d have to stay very well hidden or defended. Just not gonna happen.

Yeah, probably a few people live on, and so humanity doesn’t go extinct. But the only realistic chance most of us have of surviving in this scenario is to use our vast industrial and scientific abilities to make food. We actually know of many plausible ways to make more than enough food to feed everyone for ten years, even with no sunlight. And even if big chunks of the world economy are in shambles. But for that to work, we must preserve enough social order to make use of at least the core of key social institutions.

Many people presume that as soon as everyone hears about a big problem like this, all social institutions immediately collapse and everyone retreats to their compound to fight a war of all against all, perhaps organized via local Mad-Max-style warlords. But in places where this happens, everyone dies, or moves to places where something else happens.

Many take this as an opportunity to renew their favorite debate, on the right roles for government in society. But while there are clearly many strong roles for government to play in such a situation, it seems unlikely that government can smoothly step into all of the roles required here. Instead, we need an effective industry, to make food, collect its inputs, allocate its workers, and distribute its products. And we need to prepare enough to allow a smooth transition in a crisis; waiting until after the sunlights goes to try to plan this probably ends badly.

Thus while there are important technical aspects of this problem, the core of the problem is social: how to preserve functioning social institutions in a crisis. So I call to social scientist superheroes: we light the “bat signal”, and call on you to apply your superpowers. How can we keep enough peace to make enough food, so we don’t all starve, if Earth loses sunlight for a decade?

To learn more on making food without sunlight, see ALLFED.

GD Star Rating
loading...
Tagged as: ,

Too Much of a Good Thing

When people are especially eager to show allegiance to moral allies, they often let themselves be especially irrational. They try not to let this show, but most aren’t very good at hiding it. One cute way to watch this behavior is to ask people if it is possible to have too much of a good thing, or too little of a bad thing. The fully rational answer is of course yes, it is usually possible to go too far in most any direction. But many seem to fear seeming disloyal if they admit this.

For example, I recently gave this poll to my twitter followers:

One of my followers, who has many more followers than I, asked hers a related poll:

While my and Aella’s followers similarly say we do too little on global warming, hers are far more likely to say that isn’t possible to do too much. And my followers who think we do too much tend to be less reasonable in that more of them think it isn’t possible to do too little.

(Note that the third option in Aella’s poll is a logical contradiction: if people actually do too much, surely it must be possible to do too much. )

This seems ripe for a larger more representative poll. Which side is more reasonable in admitting that one could go too far in their direction? And which other kinds of people are more reasonable? How does this can’t-have-too-much effect vary with the topic?

Added 3:30p: If you can understand the first question, on if we do too much or little, you should be able to understand the second question, on if such things are possible. I don’t get how you can be confused about the meaning of the second question, yet can easily answer the first question.

GD Star Rating
loading...
Tagged as:

MRE Futures, To Not Starve

The Meal, Ready-to-Eat – commonly known as the MRE – is a self-contained, individual field ration in lightweight packaging bought by the United States military for its service members for use in combat or other field conditions where organized food facilities are not available. While MREs should be kept cool, they do not need to be refrigerated. .. MREs have also been distributed to civilians during natural disasters. .. Each meal provides about 1200 Calories. They .. have a minimum shelf life of three years. .. MREs must be able to withstand parachute drops from 380 metres, and non-parachute drops of 30 metres. (more)

Someday, a global crisis, or perhaps a severe regional one, may block 10-100% of the normal food supply for up to several years. This last week I attended a workshop set up by ALLFED, a group exploring new food sources for such situations. It seems that few people need to starve, even if we lose 100% of food for five years! And feeding everyone could go a long way toward keeping such a crisis from escalating into a worse catastrophic or existential risk. But for this to work, the right people, with the means and will to act, need to be aware of the right options at the right time. And early preparation, before a crisis, may go a long way toward making this feasible. How can we make this happen?

In this post I will outline a plan I worked out at this workshop, a plan intended to simultaneously achieve several related goals:

  1. Support deals for food insurance expressed in terms that ordinary people might understand and trust.
  2. Create incentives for food producers, before and during a crisis, to find good local ways to make and deliver food.
  3. Create incentives for researchers to find new food sources, develop working processes, and demonstrate their feasibility.
  4. Share information about the likelihood and severity of food crises in particular times, places, and conditions.

My idea starts with a new kind of MRE, one inspired by but not the same as the familiar military MRE. This new MRE would also be ready to eat without cooking, and also have minimum requirements for calories (after digesting), nutrients, lack of toxins, shelf life, and robustness to shocks. But, and this is key, suppliers would be free to meet these requirements using a wide range of exotic food options, including bacteria, bugs, and rats. (Or more conventional food made in unusual ways, like sugar from corn stalks or cows eating tree leaves.) It is this wide flexibility that could actually make it feasible to feed most everyone in a crisis. MREs might be graded for taste quality, perhaps assigned to three different taste quality levels by credentialed food tasters.

As an individual, you might want access to a source of MREs in a crisis. So you, or your family, firm, club, city, or nation, may want to buy or arrange for insurance which guarantees access to MREs in a crisis. A plausible insurance deal might promise access to so many MREs of a certain quality level per per time period, delivered at standard periodic times to a standard location “near” you. That is, rather than deliver MREs to your door on demand, you might have to show up at a certain more central location once a week or month to pick up your next batch of MREs.

The availability of these MREs might be triggered by a publicly observable event, like a statistical average of ordinary food prices over some area exceeding a threshold. Or, more flexibly, standard MRE insurance might always give one the right to buy, at a pre-declared high price and at standard places and times, a certain number of MREs per time period.  Those who fear not having enough cash to pay this pre-declared MRE price in a crisis might separately arrange for straight financial insurance, which pays cash tied either to a publicly triggered event, or to a market MRE price. Or the two approaches could be combined, so that MRE are available at a standard price during certain public events.

The organizations that offer insurance need ways to ensure customers that they can actually deliver on their promises to offer MREs at the stated times, places, and prices, given relevant public events. In addition, they want to minimize the prices they pay for these supplies of MREs, and encourage suppliers to search for low cost ways to make MREs.

This is where futures markets could help. In a futures market for wheat, people promise to deliver, or to take delivery, of certain quantities of certain types of wheat at particular standard times and places. Those who want to ensure a future supply of wheat against risks of changing prices can buy these futures, and those who grow wheat can ensure a future revenue for their wheat by selling futures. Most traders in futures markets are just speculating, and so arrange to leave the market before they’d have to make or take delivery. But the threat of making or taking delivery disciplines the prices that they pay. Those who fail to make or take delivery as promised face large financial and other penalties.

Analogously, those who offer MRE insurance could use MRE futures markets to ensure an MRE supply, and convince clients that they have ensured a supply. Yes, compared to the terms of the insurance offered by insurance organizations, the futures markets may offer fewer standard times, places, quality levels, and triggering public events. (Though the lab but not field tested tech of combinatorial markets make feasible far more combinations.) Even so, customers might find it easy to believe that, if necessary, an organization that has bought futures for a few standard times and places could actually take delivery of these futures contracts, store the MREs for short periods, and deliver them to the more numerous times and places specified in their insurance deals.

MRE futures markets could also ensure firms who explore innovative ways to make MREs of a demand for their product. By selling futures to deliver MREs at the standard times and places, they might fund their research, development, and production. When it came time to actually deliver MREs, they might make side deals with local insurance organizations to avoid any extra storage and transport costs of actually transferring MREs according to the futures contract details.

To encourage innovation, and to convince everyone that the system actually works, some patron, perhaps a foundation or government, could make a habit of periodically but randomly announcing large buy orders for MRE futures at certain times and places in the near future. They actually take delivery of the MREs, and then auction them off to whomever shows up there then to taste the MREs at a big social event. In this way ordinary people can sometimes hold and taste the MREs, and we can all see that there is a system capable of producing and delivering at least modest quantities on short notice. The firms who supply these MREs will of course have to set up real processes to actually deliver them, and be paid big premiums for their efforts.

These new MREs may not meet current regulatory requirements for food, and it may not be easy to adapt them to meet such requirements. Such requirements should be relaxed in a crisis, via a new crisis regulatory regime. It would be better to set that regime up ahead of time, instead of trying to negotiate it during a crisis. Such a new regulatory regime could be tested during these periodic random big MRE orders. Regulators could test the delivered MREs and only let people eat the ones that pasts their tests. Firms that had passed tests at previous events might be pre-approved for delivering MREs to future events, at least if they didn’t change their product too much. And during a real crisis, such firms could be pre-approved to rapidly increase production and delivery of their product. This offers an added incentive for firms to participate in these tests.

MRE futures markets might also help the world to coordinate expectations about which kinds of food crises might appear when under what circumstances. Special conditional futures contracts could be created, where one only promises to deliver MREs given certain world events or policies. If the event doesn’t happen, you don’t have to deliver. The relative prices of future contracts for different events and policies would reveal speculator expectations about how the chance and severity of food crises depend on such events and policies.

And that’s my big idea. Yes it will cost real resources, and I of course hope we never have to use it in a real crisis. But it seems to me far preferable to most of us starving to death. Far preferable.

GD Star Rating
loading...
Tagged as: , ,

Tegmark’s Book of Foom

Max Tegmark says his new book, Life 3.0, is about what happens when life can design not just its software, as humans have done in Life 2.0, but also its hardware:

Life 1.0 (biological stage) evolves its hardware and software
Life 2.0 (cultural stage) evolves its hardware, designs much of its software
Life 3.0 (technological stage): designs its hardware and software ..
Many AI researchers think that Life 3.0 may arrive during the coming century, perhaps even during our lifetime, spawned by progress in AI. What will happen, and what will this mean for us? That’s the topic of this book. (29-30)

Actually, its not. The book says little about redesigning hardware. While it says interesting things on many topics, its core is on a future “singularity” where AI systems quickly redesign their own software. (A scenario sometimes called “foom”.)

The book starts out with a 19 page fictional “scenario where humans use superintelligence to take over the world.” A small team, apparently seen as unthreatening by the world, somehow knows how to “launch” a “recursive self-improvement” in a system focused on “one particular task: programming AI Systems.” While initially “subhuman”, within five hours it redesigns its software four times and becomes superhuman at its core task, and so “could also teach itself all other humans skills.”

After five more hours and redesigns it can make money by doing half of the tasks at Amazon Mechanical Turk acceptably well. And it does this without having access to vast amounts of hardware or to large datasets of previous performance on such tasks. Within three days it can read and write like humans, and create world class animated movies to make more money. Over the next few months it goes on to take over the news media, education, world opinion, and then the world. It could have taken over much faster, except that its human controllers were careful to maintain control. During this time, no other team on Earth is remotely close to being able to do this.

Tegmark later explains: Continue reading "Tegmark’s Book of Foom" »

GD Star Rating
loading...
Tagged as: , ,