Sunday, October 25, 2015

Good Communication, Bad Communication, and Head-Desk-Related Injury

A few days ago I had the joy of listening to a designer and an engineer discuss a minor change to a web page. It went something like this:

Designer: "Ok, I want it just like it was before, but put this part at the top."

Engineer: "Like this?"

Designer: "No, I don't want everything else moved down. Just keep everything else where it was, and put this at the top."

Engineer: "But putting that at the top pushes everything else down."

Designer: "It doesn't need to. Look, just..."

... this went on for about 30 minutes, with steadily increasing frustration on both sides, and steadily increasing thumping noises from my head hitting the desk.

Thump. Thump. Thump.

It turned out that the designer's tools built everything from the bottom of the page up, while the engineer's tools built everything from top down. So from the designer's perspective, "put this at the top" did not require moving anything else. But from the engineer's perspective, "put this at the top" meant everything else had to get pushed down. This revelation did not reduce the pain from thumping my head on the desk.

Whenever people communicate, a certain amount of translation has to happen. We don't all think in exactly the same way, so somebody has to translate from what makes most sense to me into what makes most sense to you, and vice versa. A "good communicator" can handle all of the translation single-handed. They can word things to make sense to anyone, and they can tease out whatever anyone tries to tell them. The best communicators take it a step further and also tease out the things their partners are trying to NOT tell them (often much more fun and interesting than what people are actually saying). A good communicator is the universal remote of human language.

A bad communicator, on the other hand, does not translate anything at all. They can't understand what other people want to say (even though they might THINK they understand), and other people can't always understand them (though again, they might THINK people understand). This underlying problem can produce symptoms which we often think of as communication problems in their own right. The most common is lots of talking and very little listening. A person with poor communication skills will frequently not understand what others try to say, so they avoid the problem by talking themselves. As long as the poor communicator is around better communicators, their partners will shoulder the effort of translating, and some understanding will be achieved. But put two poor communicators together, and it takes 30 minutes to figure out that one is building from the top and the other from the bottom.

Nothing is as frustrating as not understanding. Stick two poor communicators together, and frustration will inevitably result. "Why can't you just put this at the top and keep everything else where it is? Are you being deliberately obtuse? Just stop arguing and do it!!!"... "Why can't you see that that's not how it works? I've explained it five times! Are you just not paying attention?!?!"... Thump. Thump. Thump.

So how do we prevent head-desk-related injuries? There's a lot of answers. At an organizational level, good management and processes can handle this problem. A good manager is always a good communicator. All they need to do is stand behind the designer and the engineer and translate. In more formal hierarchies, the designer and engineer are required to communicate through the manager. If one or both of the designer/engineer are good communicators (or worse, the manager is a bad communicator), then communicating through the manager is useless. But if the designer and engineer are both poor communicators, and the manager is a good communicator, then the problem is solved.

Even absent good management, process can substitute for management. In this particular case, the engineer suggested that all design changes, no matter how minor, had to come in visual form. This creates quite a bit of extra work for the designer, but it means not burning 30 minutes failing to communicate.

Of course, both managerial and process solutions are highly inefficient. They require an extra person, extra work, or both. They don't always generalize well. Ideally, we want people to communicate directly. Most people, most of the time, can communicate reasonably well. They're not the best communicators, but not bad either. Most of us aren't universal good communicators, but we learn to communicate with those around us.

Feel free to leave good communication advice in the comments. Better yet, leave cryptic advice in the comments and let everyone else try to figure out what you meant.

Thursday, October 1, 2015

The Breakthrough How-To, Part 3

What sort of environment lends itself to breakthroughs? This is a pretty broad question. Environment covers a lot of variables, including many which affect cognition in general. I'll focus mainly on the social environment of breakthrough.

In the previous post, we compared designing an oil rig (a hard but non-breakthrough type of problem) to P-NP (a very hard breakthrough type of problem). Designing an oil rig is hard because it has lots of pieces, but each piece is straightforward. To design an oil rig, we need a large team of engineers. Each engineer needs to work on some straightforward part of the problem. The engineers will need to work together to make sure that their parts are all compatible, e.g. the structure is strong enough to support the pipe and drill motor. Management will be needed to make sure each part gets done and everyone works together smoothly.

Now imagine a similar team working on P-NP. We gather a large team of computer scientists, and management tells them to... um... do something. Yeah, go solve that problem! And the computer scientists sit there staring at each other.

There's an old parable about the wisdom of crowds. According to the parable, a certain emperor was never seen in public. A student wanted to find the size of the emperor's nose, but since the emperor was never seen in public, the student could not measure it directly. So, the student resorted to the wisdom of crowds: the student ran a huge survey of a major city, asked every resident to estimate the size of the emperor's nose, and averaged all the estimates. Of course, none of the people questioned had ever seen the emperor's nose either. Low and behold, the student wound up with a number which had nothing whatsoever to do with the actual size of the emperor's nose.

If nobody's ever seen the emperor's nose, then no matter how many people you survey, you won't get any closer to an accurate estimate its size.

There are a number of morals to that story, but the moral for us is that putting together a lot of people with no information does not create information. If no one has any idea at all how to approach P-NP, then putting a thousand such people in a room will not get you any closer to solving P-NP (no matter how much experience management has).

That does not mean that teamwork is useless for breakthroughs. History does show an abundance of insight by individuals (Isaac Newton formulated most of his ideas during a one-year stint in the countryside). But history also shows that certain kinds of groups make breakthroughs as well. Breakthrough just doesn't come from the kind of teams that produce oil rigs.

Let's consider a simplified, abstract model of a breakthrough. Let's say that our breakthrough involves getting from point A to point D. Anyone with a working knowledge of algebraic topology can get from A to B, a smart geneticist can get from B to C, and some basic but non-obvious high school algebra can get from C to D. One way to make this breakthrough is for a single generalist with a knowledge of both algebraic topology and genetics to sit down and play with the problem for a while. But how might a team make the jump?

Let's say our team is a mathematician and a biologist, each with the requisite skills for the problem. The main difficulty for the team is that the intermediate points B and C don't seem useful by themselves. The mathematician can see the connection A -> B, but B doesn't seem useful to the mathematician. B does seem useful to the biologist, because the biologist can see the B -> C connection easily. But C doesn't seem useful to either of them, at least until they play around with it a bit and realize that it's equivalent to D. In order to make the jumps, the mathematician has to show the A -> B connection even though it seems silly, and the biologist has to show the B -> C connection even though it seems useless, and they both have to play around a bit with C even though it seems irrelevant. After all, if the intermediate steps were obviously useful, someone would have made the connections immediately and the problem would not require any breakthrough at all.

This whole process is socially awkward. We're socially trained not to present ideas that seem useless, and in group discussions such ideas tend to get shot down. This problem was a major focus of Isaac Asimov's 1959 essay on how people get new ideas. Asimov concluded that in order for a group to make breakthroughs, their meetings had to be somewhat silly. By the very nature of breakthrough-type problems, people need to be willing to throw out ideas which may or may not be relevant, which seem silly or unrelated. People need to be willing to sound foolish, and anyone who is "unsympathetic" to the foolishness will quickly kill the mood and destroy any hope of making the requisite connections.

Asimov recommended small groups, no more than five, so that people would not feel the pressure of waiting to speak. He also pointed to a relaxed, pressure-free atmosphere as a key component. He pointed out that being paid for the meeting generally created more pressure, and might be undesirable. Similarly, if one person had a much higher reputation, it could chill discussion. Finally, Asimov believed that such sessions needed to alternate with people going off to think on the problem alone and process whatever came up in discussions. As he put it: "The creative person is, in any case, continually working at it. His mind is shuffling his information at all times, even when he is not conscious of it."

Asimov's ideas about the stifling effects of pressure on group behavior bear a remarkable resemblance to psychological research on the candle problem. The candle problem looks simple: put a few people in a room with a box of matches, some thumbtacks, and a candle. Their challenge is to mount the candle on the wall and light it. Of course, it's not as easy as it seems. Simply sticking the candle to the wall with the thumbtacks invariably fails. Solving the problem requires a tiny breakthrough, a novel use of the materials.

Setup for the candle problem.

Just as Asimov said, adding pressure to the group makes the candle problem harder. A group under time pressure with monetary rewards is less likely to find the solution at all, and takes longer to find it when they do. Similarly, larger rewards result in slower progress, and make people more likely to debate bad solutions rather than find the good solution.

However, there is a way to reliably improve performance in the candle problem. If participants are instructed to first discuss the problem as much as possible WITHOUT actually solving it, then they are far more likely to find the solution. In other words, to encourage breakthroughs, explicitly tell people that they should just try to explore the problem rather than solve it. Then people are much more willing to suggest ideas which don't seem immediately useful. After all, that's what the instructions say to do.

The Breakthrough How-To, Part 2

The previous post gave some background on Kuhn's theories about scientific development and breakthrough, and extended those ideas to industry. In this post we'll talk about how to find breakthroughs.

1. What sort of problems lend themselves to breakthroughs?
Let's start with the problem. There's a reason this blog is called "Seeking Questions". Breakthroughs tend to start with an open, unsolved, hard problem. But not just any sort of hard problem; certain kinds of problems lend themselves to breakthrough solutions.

Problem 1: Build a deep-sea oil rig. Breakthrough? Optional.

Problem 2: P vs NP. Breakthrough? Required.
Consider building a deep-sea oil rig. This is a very hard problem. Deep-sea oil rigs are very complicated, with dozens of subsystems and hundreds of thousands of parts. On the other hand, each component of a deep-sea oil rig is straightforward. The drill, the casing, the motor, the circulation, the stabilization... each of these is a previously solved problem. To build a deep-sea oil rig, each subsystem is assigned to a small group of engineers, and each group can design their part. The end result is complicated, but it is complicated only because of the number of parts, not because of the complexity of any single part. Deep-sea oil rigs do not require fundamentally new insights; they do not require any major breakthrough.

For a breakthrough, we don't want a problem which is hard only because it has a large number of straightforward pieces. So what kind of "hard" do we want?

Let's consider a particular open problem: P-NP. P-NP is a problem in computer science which asks whether or not two particular large classes of problems (P problems and NP problems) are equivalent. At this point, a solution or even any significant progress on P-NP would certainly be a breakthrough. Alas, we have little idea of how to approach the problem. It's hard to even find a starting point (a useful starting point, anyway). P-NP is not a problem with lots of straightforward pieces; it is a problem with a single large, cloudy piece which we don't understand. It requires fundamentally new insight.

So we have two examples of hard problems: a problem which is hard because it has lots of pieces, and a problem which is hard because it requires new insight. The former requires hard work and a large team to solve. The latter requires insight, and lends itself to breakthrough. I believe that most if not all hard problems fall into one (or sometimes both) of these categories. For breakthroughs, we want to look for the latter type of problem: problems which require new insight.

There's still variety within problems which require insight. There are insights which take five minutes and there are insights which take five years. Presumably, most of the time, more difficult insights are needed for harder problems and correspond to bigger breakthroughs. I might have a small breakthrough in a project at work in five minutes; I might have a large breakthrough on a major open problem in five years.

2. What sort of skillsets lend themselves to breakthroughs?
A big takeaway from the last section is that breakthrough-type problems involve new insight. At first, that makes it hard to anticipate what sort of skillset will be useful. If nobody's had the key insight yet, then presumably the key insight will not be included in any extant skillset. But that's not quite true...

For starters, the insight need only be new to the people working on the problem, not necessarily original. For example, both physicists and chemists have a long history of regularly invading biology (sort of like China's relation with the steppes). These invasions tend to produce major breakthroughs in biology, including bacterial locomotion and the birth of molecular biology. Taking knowledge from one field and applying it in another is about as close as we can get to a reliable recipe for producing breakthroughs. It's certainly not the only way, but it's probably the most reliable.

This suggests a generalist skillset as ideal for finding breakthroughs. Contrast to a specialist skillset: a specialist gains lots of practice within a particular paradigm, able to quickly and efficiently apply the tools of that paradigm. But when the specialist's tools fail, they have nothing to fall back on. When a different paradigm is needed, the specialist can make no headway. A generalist, on the other hand, has many more tricks to try when one paradigm fails. The more fields the generalist knows, the better. Of course, generalization has its tradeoffs: a generalist will usually be slower and more error-prone with any particular tool than a specialist. For problems which the specialist can solve, the specialist will produce a better solution faster. But the generalist shines when the specialist's tools fail altogether.

There's more to the story. A general skillset is preferable for breakthrough-type problems, but not all fields are equal. The tools of some fields are far more general than the tools of other fields. Mathematics, in particular, offers the most flexible and powerful tools for technical problems. Within mathematics, the tools of applied math tend to prove useful in new problems. After mathematics, computer science is a close second, especially in today's environment. That said, the generalist rule still applies: the more different tools you have, the more likely one of them will have the right insight for a new problem.

Next post we'll look at what sort of environment lends itself to breakthroughs.

The Breakthrough How-To, Part 1

"Thomas Kuhn" by Source. Licensed under Fair use via Wikipedia - https://en.wikipedia.org/wiki/File:Thomas_Kuhn.jpg#/media/File:Thomas_Kuhn.jpg
Thomas Kuhn

Probably the best-known name in the field of science history is Thomas Kuhn. This post is mostly going to be background on Kuhn's paradigm of paradigms, and some extension of his ideas into industry. If you're familiar with Kuhn and you can see how startups fit into his ideas, feel free to skip to the next post.

Kuhn's book "The Structure of Scientific Revolutions" argues that "science" is really an amalgamation of two very different processes. The first part of science is "normal" science, the everyday work of most people in scientific research. Normal science incrementally develops existing theories. Things like measuring the gravitational constant, simulating the folding of specific proteins, isolating molecular species, deducing evolutionary trees, or smashing together high-energy particles all fall under normal science.

The second part of science consists of breakthroughs. In contrast to normal science, a breakthrough is not incremental. It is a significant change to the existing model with the potential to explain a wide variety of previous-poorly-understood phenomena. Discoveries like the heliocentric model of planetary motion, Newton's theory of gravitation, Einstein's big four papers, evolution, DNA, polymers, the periodic table, and high-temperature superconductors all fall under breakthrough science.

Abstractly, Kuhn characterizes these two types of scientific work in terms of paradigms. A paradigm is a model or framework for understanding, like the DNA -> RNA -> protein framework in biology or Newtonian mechanics in physics. I mostly use very big paradigms as examples, because most people have heard of them. However, paradigms can be much smaller as well, like the Cooper pair theory of superconduction. Normal science operates within a paradigm, applying and extending existing ideas. Breakthrough science creates a new paradigm.

For example, when superconductivity was first discovered, the existing paradigm of electrical resistance could not explain it. Researchers began to explore the phenomena, experimentally measuring superconductivity in many different materials and developing many different models which could explain certain aspects of superconductivity. Eventually, Cooper realized that quantum behavior of electron pairs at very low temperature could neatly explain the accumulated experimental results, and this became the central paradigm of superconductivity (the old electrical resistance paradigm was not abandoned, we just needed a new paradigm for these special materials at low temperatures). Cooper's model drove decades of research in superconduction, allowing researchers to predict which materials would superconduct at which temperatures and to develop new superconductors with useful properties. Later, high-temperature superconductors were discovered, and Cooper's model could not explain it. The cycle began again, and today researchers are still experimenting with different materials and models in search of a good model of high-temperature superconductivity.

This is the usual progression of science. Most scientists spend most of their time experimentally measuring things, developing partial models, and applying current knowledge create useful new things. This is normal science. Every now and then, something big shakes up the paradigm: the discovery of superconductors or high temperature superconductors, or the discovery of a new theory like Cooper's theory of superconduction. This is breakthrough science.

Kuhn talked quite a bit about the social aspects of the two types of science. People doing normal science aren't always happy when someone comes along and upsets their paradigm. Kuhn's book is great if you want to hear more about that. Meanwhile, I'm going to take it in a different direction.

Although Kuhn mostly stuck to academia, the patterns of normal vs breakthrough science apply outside of the sciences. Industries have their own paradigms, and these are regularly upset, often by technical breakthroughs. The lightbulb, the transistor, the assembly line, containerized shipping, stock options, personal computers and the internet, radio broadcasting, the iPhone, Facebook... each of these was a breakthrough which shook things up and created a new paradigm in business. Throughout the twentieth century, the pace of breakthrough has accelerated, and large businesses today find themselves under pressure to produce breakthroughs just to keep up. In recent decades, we've even seen the emergence of a new type of business which is defined by explicitly seeking a breakthrough: the startup.

On other end of the spectrum, non-breakthrough work is on the decline. Things within the current paradigm are things we understand well, and things we understand well are precisely the things which can be automated or outsourced to the lowest bidder. Traditional management is quite good at taking simple, well-understood tasks and getting people to do them quickly and at low cost.

I would argue that every field out there has its paradigms and its interruptions. Some are far more stable than others, but the pace of breakthrough has only accelerated for more than a century. If the past two centuries are any indicator, the future will see more people spending more time explicitly working toward breakthroughs, and normal within-paradigm work will become increasingly automated.

There's a problem, though. Historically, most people have spent most of their time on normal work rather than breakthrough work. Consequently, our education system, our management structures, and our work culture are all optimized for non-breakthrough work. The breakthrough process is largely mysterious; we still don't understand what sort of background or environment will make them happen. To that end, the next post will talk about what sort of knowledge, environment, challenges and thought patterns lend themselves to breakthroughs.