Ken's Blog

Every prototyping project has its own rhythm, but there are patterns within these rhythms that one can see emerging over time — eternal constants in the process of creation. This week, as I iterate yet again on a prototype, I am noticing a pattern that I’ve seen before.

In particular, I notice that my process of prototyping is divided into three successive phases. Each phase, or “version” if you will, serves a distinctly different purpose.

The first version of the prototype says “This is just a rough sketch of what the result will look like”. In that first stage, the important thing is to quickly and efficiently throw something together which conveys an idea, with no attempt at robustness or reusability.

The second phase says “Watch this demo — it has many cool features, but it will break if anybody uses it except me”. This phase is a bit of a magic act. The idea is to show the full potential of the approach, but without the robustness of a real product. As long as I’m the only one using it, it looks wonderful. But if anybody else were to try to use it, they would quickly put their foot through a hole in the floor, and things would all fall apart pretty rapidly.

The third phase says “Go ahead, you use this.” This is a much harder thing to create, and it requires everything that was learned during the first two phases. It is the stage where you build something that can withstand being used by random people — people who aren’t trying to avoid the weak spots.

If you get that far, you may well be on your way to a usable product.

This week a representative from the White House Office of Science and Technology Policy (OSTP) came to our university to talk about Grand Challenges. These are big research questions that the U.S. government considers to be difficult or long term, that usually require university level research in addition to corporate research, and which can have a big impact on society or lots of great spin-off benefits.

In a PowerPoint presentation we learned of such Grand Challenges are the self-driving car (which started at universities, but which Google has now taken over — the hope is to save many lives each year), sequencing the human genome (which was largely successful), and an inexpensive implementation of the “Tricorder” — the hand-held device on Star Trek that can diagnose diseases. Believe it or not, people are actively working on a real version of that.

A canonical example of a Grand Challenge was the MoonShot — John F. Kennedy’s challenge in May 1961 to land a human being on the moon before the decade was out. Many useful spinoff technologies were developed during the eight years it took to realize that goal.

A number of such challenges were listed, and one in particular caught my eye: The “SunShot”. Eventually we learned that this was a program to develop truly practical next generation photovoltaic solar power.

Unfortunately, I saw the word “SunShot” on the PowerPoint slide well before I heard this explanation. So for several minutes I found it very difficult to concentrate on anything else.

This is because I was trying to work out in my head what seemed like an insoluble problem: How were they going to land a person on the Sun?

A friend told me today about the novel Dust, a 1999 sci-fi thriller by Charles Pellegrino, in which humanity is attacked and threatened with extinction by massive killer hordes of voracious mites.

My friend said that one thing he found notable about the premise is the contrast in behavior between individual mites and the giant swarms they form: Whereas each individual mite is incredibly stupid and dimwitted, when they band together in massive numbers the collective group entity displays a surprising degree of intelligence and inventiveness.

This struck me as very funny, because with people it works in exactly the opposite way.

I told a friend today that something I’d really love to do, if I had the resources, would be to throw parties that let people break free from their every day life.

More specifically, I would want to throw themed parties — ancient Romans, depression-era gangsters, the court of King Arthur, cowboys of the old West, roaring twenties flappers — anything that could invite a group of people into a shared spirit of romance and delight.

And here’s the best part: At each party there would be a room, near the entrance, devoted entirely to providing costumes for the guests. By the time you entered the party proper, you would be fully decked out in your personal vision of that night’s theme. So no pressure to find the perfect costume.

It’s such a lovely thought, isn’t it? Then again, if I really had enough money to do something like that, I’d probably end up using it to help the world by battling world hunger and promoting universal education.

Hmm. Maybe I could do both. 🙂

Sometimes I wish I had a perfect memory. This happens during those many moments when I find myself frustrated that a name, or a date, or some other factoid I’m sure I used to know, is stubbornly eluding my mental grasp.

Usually, if I allow the question to linger in the back of my mind for a few minutes, the answer will float magically up to the surface, like one of those mysterious predictions in the window of a Magic 8 Ball. But alas, not always.

Oh, for perfect powers of recollection!

Yet such powers might turn out to be not a dream, but a nightmare. One of the wonderful things about being human is our ability to weave our memories into a coherent narrative, one that allows us to make sense of things. Such weaving requires not only the ability to remember, but also the ability to forget.

We humans are pattern makers, who create meaning for ourselves not merely from the raw material of lived experience, but from the lights and shadows we are able to perceive within that experience. Yes, some of those shadows contain dark memories, doorways into experiences of hurt and pain.

From time to time we need to stare into those shadowy doorways, drawing what we need from the darkness within. But afterward, we must remember to shut the door.

I finally got around to seeing “Brick”, Rian Johnson’s awesome post-modern film noir from 2005. The execution of the film is brilliant, but the premise is very simple: “Suppose a 1940’s hardboiled detective story were set in a modern suburban high school?”

This film is able to work so well because it does not condescend. It treats the premise and the characters with utter seriousness, carefully transposing all of the noir character archetypes, relationships and dramatic conventions to this new setting. The clever and inventive camera work and editing are also entirely consistent with the tropes of noir. I didn’t detect a single moment when the film winked at the audience.

I wonder whether this is the key to updating a traditional genre: Rather than merely using it as source material, go deep and understand what really makes the genre tick, and respect its conventions.

After all, there’s a reason that classic genres are classic.

When you are constructing a prototype, sometimes you first need to put it together with string and chewing gum, just so you can quickly get a sense of things. The result doesn’t need to be robust or even completely functional — it just needs to tell you whether you are going in the right general direction.

But then at some point, after you’ve figured out what direction you are heading, you need to replace the string and chewing gum with bricks and mortar — in my case, solid techniques of software construction that will be sturdy and reliable and built to last.

I often find the transition from the former to the latter a bit stressful. Here I’ve just had a nice little lean-to that worked, albeit badly, and now I am going to tear it all down and replace it with a real building. That is going to take a while, and during the transition I won’t really have anything to show for my efforts. But of course it’s worth it.

Yesterday I noticed something: I wouldn’t let myself start construction on the more robust version of my prototype until I got the shoddy temporary version basically working. I found this puzzling. Here was something I knew I was going to toss out, something that I knew wasn’t even worth very much, but I simply had to finish it before I would let myself start on the next phase of the project.

At first I was perplexed, but eventually I realized that the benefits of this strategy were very real, although entirely psychological. Knowing that I had finished something, even something not made to last, would give me the confidence I needed to tackle the much more challenging problem ahead.

Robert M. Persig put it very well: “Peace of mind produces right values, right values produce right thoughts. Right thoughts produce right actions and right actions produce work.”

The other day I touched on the difference between Google Glass and the (highly conjectural) concept of “Matrix”-style direct to brain knowledge downloading. The ensuing discussion has been very interesting.

One thing that has come out of that discussion is the question of how best to use Glass-style augmentation in the context of an ongoing conversation. It might very well turn out that as augmentation technology matures, the user experience will evolve into a kind of continuous background process, never quite requiring our attention, but rather feeding unobtrusively into all our conversations.

In this kind of scenario, the true power-up technology will not be the display itself, but the coordinated pipeline of automatic speech recognition, contextual look-up, and semantic inference that will listen to our conversations and provide immediate information in response.

In essence, our augmentation devices will “whisper in our ear”, feeding us new information that is informed by what was just discussed, and that is optimized for suggesting new and fruitful conversational directions. If such technology is working properly, discussants need never focus on the existence of this back-channel.

Such a real-time knowledge supplement stream would be the conversational equivalent of wearing contact lenses or a hearing aid, or having an air conditioner. For if such augmentation does its job properly, it will become as clear as glass — and we will forget it is there. We will simply become better and more informed conversationalists, while forgetting that there is any technology involved.

This evening I attended a small social gathering at my sister’s house. I was surprised to find that the five year old daughter of one of one of my sister’s friends knew that a Googol is 10¹⁰⁰, and that she was wondering how long it would take to count up to a Googol. I told her it would take way too long, so she suggested maybe counting by hundreds or thousands to speed things up.

I then suggested it might work better if she counted by double each time — one, then two, then four, and so on. She quickly saw the logic of this and started working it out. “Let’s see,” she said, “one, then two, then four, then eight, then sixteen, then …” she thought a moment, “thirty two!”

Soon the topic shifted to parallel universes. I mentioned having read recently that scientists think there might be an infinite number of parallel universes, so it’s possible that we exist in many different worlds, only we could be different versions of ourselves in each one.

After thinking about this for a bit, she pointed out that there might be a universe where there are still dinosaurs. “Yes,” I said sadly, “but we could never see them, because you can’t get from one universe to another.”

“But maybe,” she said, “we could invent robots that could travel between worlds. We could send them to take pictures of the dinosaurs for us.” I was doubtful, but I had to admit it was an intriguing thought.

Our conversation then broadened. Since parallel universes can contain anything, we conjectured, there could be universes where dragons are real. Then we realized that there might be a universe where dragons are having this very conversation. “Imagine,” we wondered, “one dragon talking to another dragon about some other universe where people are real, and not just imaginary creatures.” We both thought this was very funny.

It was one of the more interesting conversations I’ve had recently.

My post yesterday asked what it might be like if we actually had the instant knowledge upload capability from “The Matrix”, In response, J. Peterson commented:

“Your Google Glass frames could handle this. Start face recognition on the person your talking to. Listen for voice and context cues (”…I write for the New York times”, GPS coordinates in the city). Then it’s just a matter of managing the awkward stare as you try to skim the articles floating just above her head.”

It’s an insightful comment, but one could argue that what Google Glass is, in a sense, the antithesis of Trinity’s skill upload. The former provides a tantalizing glimpse into a world of knowledge, whereas the latter actually delivers that world.

Imagine, for the sake of argument, a future society where everyone is “wearing”. In such a world, my conversant would expect me to know, mere seconds into our conversation, the titles of her articles in the NY Times — they would be showing on my heads-up display. But I would still know none of the substance.

In reality, I spent more time reading those articles after the party than we had spent conversing. Google Glass and similar devices, such as the Vuzix M100, cannot replicate that experience. Whereas an actual brain upload (which is most likely decades away), might indeed provide true in-depth knowledge in real time.

In a way it’s like the difference between a spyglass and a jetpack. The spyglass will let you see the contours of a distant mountain that remains tantalizingly out of reach. The jetpack will take you there.