I am extremely happy about the new Marvel series Loki, having just seen the first episode. It feels like the fulfillment of the promise of WandaVision.

Rather than the same old boring superhero narrative, we are being treated to something truly original and imaginative. Unlike nearly everything we have seen in the Marvel universe, Loki is actually very good science fiction.

And on top of that, it’s funny as hell. Can’t wait to see how this plays out.

Programmum non nocere

I am in the middle of a rapid software development project. There is a first demo in a few weeks, and I need to add a lot more capability between now and then.

Every day I add new features, or replace old ones. Which means that every day there are opportunities for massive failure.

I keep backups at each step, so failure is never catastrophic. Still, I don’t really want to lose the bulk of a day’s work by introducing some bug that will force me to go back to a previous version.

So I’m adopting a kind of variation of the physician’s motto “First do no harm.” Whenever I add a new feature that will replace an older one, I keep the older feature in for a while.

The idea is that everything should continue to operate exactly the way it did before I added the new code. Except now there are new things you can also do if you know how.

I keep it that way until I am confident that I haven’t broken anything. Only then do I get rid of the old version of the feature.

I realize that the stakes here are a lot lower than what doctors deal with every day. Still, it’s good to have a way to work that doesn’t keep breaking anything.

Architect birthdays

I find it fascinating that both Frank Lloyd-Wright and Tim Berners-Lee were born on this day of the year. The former was born in 1867, the latter in 1955.

Both are our most famous architects in two very different domains: The former is the most famous of all architects in the traditional sense. The latter is the most famous of all architects of the World Wide Web.

Right now we think of those two domains as disparate. But that distinction might one day come to seem less significant, depending on how things go.

I wonder whether there will come a time when we will just think of both as famous architects.

Virtual tall ceilings

I love tall ceilings. Just give me a good ten or twelve foot tall ceiling in a room and I am happy.

But as we know, You can’t always get what You want. There are practical reasons that ceilings need to be lower than we would desire.

I am wondering whether, when we all have those future mixed reality glasses, we will be able to have virtual tall ceilings wherever we want them. It might be expensive, but I imagine it would be a lot cheaper than actual twelve foot high ceilings for every room in a house or apartment in a building.


Once upon a time, I dreamt I was an artificial intelligence, calculating here and there, to all intents and purposes a machine. I was conscious only of my happiness as a robot, unaware that I was myself. Soon I awoke, and there I was, veritably myself again. Now I do not know whether I was then a man dreaming I was a computer, or whether I am now an A.I., dreaming I am a man.

With apologies to Zhuangzi


Our lab is still virtual. But now we are fairly certain that it will once again be a physical place when the semester starts in early September.

And I can see how that knowledge is changing everyone’s mindset, day by day. As we near that critical date, the idea of going from virtual to physically co-present stops being an abstraction, and starts becoming a practical reality.

People are rethinking where they will be geographically. It is now important once again to think about how we will move our bodies through space and time every day.

Even after our lab and in-person classes reopen in September, I am going to try to hold on to some aspects of the virtual. It is not clear to me that meeting in person every day is categorically better, or that we should hold our classes in a physical room every time, even if we can.

Maybe some new mix of real and virtual, enabled by ever better software, is closer to the ideal. At any rate, it’s certainly worth exploring.


These last two days I have been going back through my computer files for the last decade or so. This is because we are presenting a course at the Siggraph computer graphics conference, which will include a retrospective of our lab’s research.

It’s quite a remarkable experience going back through the last decade or so of your life. The memory gets jogged in all sorts of interesting ways.

I am rediscovering memories of people and places I have not thought about for quite a while. And I am getting a sense of what my mindset was back then — which in many ways was very different from the way I think now.

If you are the sort of person who keeps around files from years ago, I highly recommend this sort of retrospective exercise. Go back and look at all your files, in chronological order.

The experience of visiting with your past self can be fascinating, if possibly unsettling. You may just learn a thing or two.

The company of strangers

What is it about a bustling restaurant that is so appealing? Objectively speaking, it would seem that strangers making noise all around you should be bothersome.

And yet most of us enjoy the feeling of dining in a crowded place, surrounded by people we don’t know. What exactly is going on here?

I wonder whether this feeling hearkens back to an earlier time in human pre-history, when the presence of the tribe denoted safety. If you are sitting around the tribal campfire, you are much less likely to be attacked by predators.

Could it really be that simple? Or is there some other principle at work here?

Designing for future hardware

We just got back some reviews for a research paper we submitted. The general topic was computer/human interfaces.

In particular, we were describing a good way to interact using devices that don’t quite yet exist. Examples of such devices include really good augmented reality glasses, and really good camera-based tracking of hand and finger movements.

In both cases, crude devices do already exist, but they are not yet really good. And for some things, you need something really good.

For example, you can’t really play the piano in virtual reality based on the kinds of camera-based hand tracking you can get today. You can sort of play the piano, but the results will be terrible.

In five years or so that will no longer be true. A combination of parallel advances in hardware and software will see to that.

In the case of our paper submission, some of the reviewers seemed confused by what we were telling them. This might be because the entire notion of “designing for hardware that does not yet exist” can be tricky.

When you talk about things like that, you need to explain very carefully what you are trying to do. Otherwise people have a tendency to just look at you in either confusion or disbelief.

Still, it’s important for us to keep trying to explain what we are doing and why. Designing for the future is hard, but it’s worth it.