Ken’s Blog

We all know that when you’re having fun, time seems to go faster. And conversely, when you are bored or having a really bad time, time can seem to drag on and on.

You could, if you wished, achieve the experience of living a longer life by always being bored. In practice, this isn’t a very good strategy, because what’s the point of life if you can’t enjoy it?

But maybe there is the germ of a good idea there. What if your level of happiness and your level of subjective time passing don’t change in the same way? In particular, maybe we can maximize the following product, added up over all the moments of your life:

Subjective Enjoyment (SE) =
      (subjective duration of each moment) ×
      (level of happiness at that moment)

 
For instance, let’s say that at some level of happiness A, time seems to go at a rapid rate T_A, and at some lower level of happiness B, time seems to go at a slower rate T_B.

Consider the level of happiness half way between A and B. What if time at that level of happiness seems to goes by more slowly than the average of T_A and T_B?

In that case, you can achieve a greater total lifetime SE score by hovering near this average state.

In fact, if we could measure both subjective happiness and subjective rate of time passing on a linear scale, we might be able to compute an optimal state of happiness, to best make you feel as though you’ve lived a long and satisfying life.

But don’t get too excited about this. If you get too happy, you might end up with a shorter subjective lifetime. 😉

When I was an undergrad at Harvard there was a one semester course that gave a broad survey of physics to non-majors. The instructors knew their students generally didn’t have a high level of mathematical preparation, and they also knew they needed to make things fun, because students were generally taking this as a requirement, not out of love for the subject.

The course was nicknamed, somewhat derisively, as “Physics for Poets”, since it was essentially trying to teach a highly mathematical field without actually using any math. So the question arises, are all such courses doomed to superficiality and irrelevance, or is there something good there?

Today some of us faculty at NYU were discussing something vaguely similar: Is there a good way to introduce principles of computer science to non-majors who have no prior background in CS?

We came to the conclusion that there isn’t one way, but there might be many ways. Computer science contains quite a few key concepts. To name just a few: looping, conditionals, variables, procedures, inheritance, computational complexity, recursion. The list goes on.

And college students have many and varied interests. To name just a few: Art, photography, music, sports, cinema, politics, literature, dance, poetry, economics, theater, journalism. The list goes on.

For any given interest a student might have, there is a way to teach a corresponding concept in computer science. Consider, for example, looping. Post-processing of photographs requires looping through pixels, music, poetry and dance require looping through rhythmic patterns, and so on. If you understand a field well enough, you can generally find a motivation for the use of any computer science topic in that field.

Of course to turn this insight into proper course design for a given student interest is far from easy. It requires real work and preparation on the part of an educator who loves both that subject and computer science.

But hey, isn’t that why we are here?

This last summer I saw a wonderful project by some graduate students in the Interactive Digital Media program at Trinity college. Noting that this year is the 150th anniversary of Alice in Wonderland, and the 100th anniversary of Einstein’s theory of general relativity, they created an original work of interactive art.

Much of the power and delight of Lewis Carroll’s classic comes from the way it warps time and space. Everything is relative, and notions of reality as a static frame of reference go out the window.

As the students observed, this is one of the fundamental predictions of Einstein’s theory. It was general relativity that definitively moved our view of reality itself away from the rigid framework of Newtonian mechanics, with its fixed notions of the nature of time and space.

The students created a clever assortment of interactive techno/art experiences that riffed on the connections twixt red shifts and Red Queens, manifolds and Mad Hatters, gravity and Gryphons. Much of it was quite delightful.

But why just those two points in time? What about 50 years ago? I got to wondering what might have happened in 1965 to shake up our fixed notions of time and space.

So I went on the Wikipedia and started snooping around. Quite a few notable things happened that year, in music, politics, science, literature and many other domains of human interest.

But one event in particular jumped out at me: On April 19, 1965, Gordon Moore published a paper laying out the principle that came to be known as “Moore’s Law” — that computation would become exponentially less expensive with each passing year.

In hindsight, that paper was the shot across the bow. It effectively predicted that our nation’s economic engine was going to shift radically from industrial to informational, from an ecology of fixed resources to one of exponentially increasing resources.

The vision of the future that Moore predicted 50 years ago — the world that we live in today — is indeed in the spirit of Carroll and Einstein. For it is a world in which the meaning of time and space, how we move through them, how we use them to communicate with one another, seems to change with every successive doubling of computational power.

I was telling a friend over dinner this evening about Joe Harris, who came up with the immortal line “Silly rabbit, Trix are for kids!” — as well as the artwork and entire idea for the spot.

In an interview years later, Harris lamented how General Mills had misunderstood his beloved character. The entire spot is built on the rabbit not getting the breakfast cereal. The way he tells it, one year GM decided to let the rabbit get some Trix. The result? Sales of Trix went down — because now there was no drama.

I wonder how many other examples there are like this in popular culture: Producers give the audience what they think it wants, and the audience rebels. Because sometimes, not getting what it wants is exactly what an audience really wants.

I have rather enjoyed the version of the memory palace portrayed in the current BBC series Sherlock. The idea of memory palaces goes way back — at least to Simonides of Ceos.

Specific memories are associated with an imagined physical architecture. In the mind of the mnemonist, each fact is placed in a particular location. One then simply needs to tour the imaginary building in one’s mind to retrieve each memory from its allotted place.

There is an argument to be made that the sort of future reality I have been describing in these posts — in which you can walk around freely, using your own physical body to visit imaginary places — would be the ideal interface in which to store a memory palace. Your own muscle memory, proprioception, body sense and geographic intuition could be integrated into the process of storing and retrieving memories.

Given what we now know about place cells and grid cells and how they operate, this seems like a very fruitful avenue of research to explore. These cells, buried deep in the hippocampus, turn out to be incredibly important.

In 2014 John O’Keefe and Edvard and May-Britt Moser won the Nobel prize for the discovery of such neurons, and for thereby showing that our ability to navigate the physical space around us is a fundamental part of how our memory works.

So perhaps the secret to future computer interfaces lies deep in our hippocampus, which is, by the way, a wonderful word. If translated literally, it means “horse fish”.

Sometimes in research we come up with a cool new way of doing things, but we aren’t always fully aware of all the purposes that new technique is good for. There are well known examples, like silly putty or the Smart Board, of inventions that were created for one purpose or market sector, but found their true utility elsewhere.

I wonder whether it would make sense to create a sort of on-line utility market, in which anybody can participate. The market would be a sort of “invention CraigsList”.

People who have invented (and presumably will patent protect) novel enablements would be able to describe those new techniques in an on-line forum. Other participants to the forum can then suggest interesting uses for those techniques, to which they can claim ownership.

In an economic sense, such a marketplace can benefit everyone involved, since each novel utility increases the potential value of the corresponding novel enablement — or combination of novel enablements.

It is obvious that there are similarities between actors from different eras of Hollywood. Audrey Hepburn and Amanda Seyfried share an elegantly elfin innocence with Leslie Caron, George Clooney and Cary Grant are the square jawed grown-up leading man with a sense of self-deprecating humor, Brad Pitt, Clark Gable and Hugh Grant the irresistable charmer with a devilishly boyish streak, Jerry Lewis and Jim Carrey the manically nutty comic with an undercurrent of bathos, aspiration toward romantic lead and a dash of occasional menace.

We recognize these similarities not merely as one individual to another, but in terms of various qualities that we associate with a Hollywood star. I suspect we could lay these qualities out in a multidimensional space, along such axes as childlike ↔ grown-up, masculine ↔ feminine, serious ↔ comic, upbeat ↔ tragic, knowing ↔ innocent, and so forth.

How many dimensions would we need to effectively classify all the major stars, and how accurately could we position each one, perhaps with the assistance of Amazon’s Mechanical Turk? With the right choice of dimensions, would they fall into constellations, with the constellation containing Clooney and Grant far away, in the celestial sphere, from the constellation that contains Pitt and Gable?

One part of Minority Report that many people remember very vividly is when Tom Cruise, wearing those cool black gloves, waves his hands around, arms outstretched in front of him, to move virtual objects on what appears to be a holographic screen. There is something iconic about this image, and I am sure it has influenced how many people think of the future of computer interfaces.

There has also been quite a bit of grumbling about this vision of interfacing with a computer. There is something awkward about needing to hold your arms out in front of you for extended periods of time. Even John Underkoffler, the real-life MIT researcher who designed this interface paradigm, has given up on it. Tom Cruise may look great doing all those arm movements, but they sure seem tiring.

But I think it all starts to make sense once we deconstruct Steven Spielberg’s probable intention in creating such an image. The fact that this is exactly the wrong vision for the future is precisely why it works so well.

In reality — as opposed to science fiction fantasies — the human body is supremely lazy, and that is a good thing. Our minds are incredibly good at moving our bodies, in performance of any given task, in a way that uses the smallest amount of energy.

You see this in just about every sort of human movement, from walking, to sitting down or standing up, to reaching for or throwing an object. No matter what the task, we have an uncanny ability to perform that task in an extremely energy conserving way.

This makes perfect sense from an evolutionary perspective. Food can be a scarce resource, and in day to day life there is no survival value in wasting energy on unnecessary movement.

The only situations in which such wasteful movements might be of use are where they carry a social message. Typically those situations come down to social dominance and sexual display. We deliberately move in an energy wasteful way to show that we can. By demonstrating our fitness, we prove a point to potential rivals or potential mates.

And this is precisely what Tom Cruise’s character is doing, on a level of storytelling. By having him perform these power gestures — gestures that nobody else in the film seems to be able to perform as well — Spielberg is telling us that John Anderton is the alpha male in our story. By doing that, he is usefully cluing in the audience to much that will happen later on in the film.

Ever the master visual storyteller, Steven Spielberg is not really interested in predicting what the technological future will look like. Rather, he is interested in guiding our emotions, via cinematic art, through a compelling character driven narrative.

I was just talking to a colleague here at NYU who, like me, is insanely busy and oversubscribed. We both have lots of projects going on in parallel, and sometimes just simply through the day seems to be one giant juggling act.

I told him my personal theory about these things. Some people — and I suspect he and I are both in this club — are plate jugglers. We work best when we are juggling lots of plates in the air at the same time.

Of course every once in a while you just can’t help yourself. You succumb to the temptation to look up and see how many plates there are in the air. And that’s when you notice that a large number of plates are hurtling down at you from above.

I told him that my usual strategy at such moments (and I suspect it is his strategy as well) is to look down, grab the nearest plate I see, and toss it in the air.

There are various kinds of time travel story. Some, like Rian Johnson’s Looper posit that time has many potential branches. When you travel back in time, you can change the course of history and thereby move your reality to another branch.

Others, like Robert Heinlein’s By his Bootstraps, posit that there is only a single time line. Anything you do by jumping back in time inevitably leads to the exact conditions that caused you to jump back in the first place.

I’m fascinated by the constraints imposed by this second kind of time travel story. Inevitably such stories convey a sense that free will is an illusion, since no matter what you do, you always end up in the same place.

But there is another interesting aspect to the linear time travel story: Higher bandwidth interventions lead to ever stranger realities. For example, if you could only send a single bit of information — true or false — from the future to, say, a year into the past, then it is reasonably plausible that whatever happens over the course of that year, the same bit value will always end up being sent back in time.

Even if a character in the story is actively trying to flip that bit, there are many reasonable storylines that could result in that character’s intent always being foiled. This is particularly true if we know that this is a linear time travel story, and that therefore paradoxes are not part of its fictional universe.

But every time we add more bits to the connection, things get a little nuttier. At the opposite extreme, a live video feed always shows reality as seen from one minute into the future. Whatever we try to do, our near future ends up being whatever is in that video. Clearly this scenario is vastly outside the bounds of any psychologically plausible narrative.

So the question I’m wondering is this: What information bandwidth that would be large enough to “break” a linear time travel story?