Why Teachers Need to Be Human

Today a friend suggested that I type “why” into Matlab. You should too. Not only had I never tried it, but it never crossed my mind that it was something that one should try.  I’ve never wondered “what happens when I type ‘why’ in Matlab?”

That’s the value of person-to-person teaching that’s irreplaceable in textbooks and MOOCs – the human ability to recognize a problem someone doesn’t even know they have. Great teachers and entrepreneurs both possess this magically human ability. Once I know I have a problem (and how to phrase it correctly) the Internet is amazingly useful, but Dr. Google is pretty useless if you just type ‘help!’

Continue reading



Yesterday’s post got me thinking more about useful questions. Finding a useful solution is always predicated on someone asking a useful question that puts you down the path to that solution. Going forward, the ability to ask useful questions is going to be more and more key to gainful employment, as computers and robots take over the task of generating answers. Even answers we thought only humans could find.  However, the day when IBM’s Watson will come up with innovative questions is still far beyond the horizon.

How do we learn to ask good questions? As I mentioned yesterday, it has a lot to do with experience. Either you generate your own hard-earned experience or draw on that of a mentor. In this regard, I think apprenticeships, trade schools, and PhD programs have a vast advantage over most traditional undergraduate education.

In question training, Cornell engineering seems to create a division: the undergraduate education teaches you to create answers, and a graduate education prepares you to create questions. The problem with this structure is that people are less and less needed to create answers, which I’d guess be an unaddressed factor in the declining value of an undergraduate degree in many fields. Also, the longer someone is taught only to find answers, the harder it is for them to start asking good questions.

Idea–>Implementation Tools

Some insightful comments from my advisor have got me thinking about innovation tools. His thought was that many more kids would want to go into STEM (Science, Technology, Engineering and Math) fields if we separated out the creation parts from the grungy math and/or fabrication parts. I’d bet that any engineer (or really anybody who creates things) shares in the experience of having to slog through grungy, parts of the process driven only by the vision of the final product.

This ties right back to thinking about the question ‘what are people good at? vs. what are computers good at?’ Two relevant answers (among many): ‘people are really good at making (sometimes bizarrely) abstract connections and new ideas’ and ‘computers are really good at grungy, known processes.’

Basically, the ideal goal is that the shorter the effort-distance between brain-flash and real-world implementation, the more brain-flashes can be tried. This in turn increases the rate of innovation iteration, accelerating new technology.

Technically, anything you didn’t have to create from raw materials technically fulfills this role. The computer of course, is the shining star in this area and the 3-D printer holds promise for applying similar iterative power to the physical world. Both still require a lot of grunge work in order to translate an idea into reality – programming languages, CAD programs along with considerations of physical and processor restrictions all constitute the bog of grunge that must be waded through.

Of course, there are tradeoffs associated with tools that shorten the idea-implementation distance. Anything that automatically abstracts the problem makes it that much harder to fix or tweak if it’s not to your liking. Here’s our modern technology black-boxification problem again – the grunge reducing tool does exactly what you want 99% of the time, and is nigh impossible to manipulate the other 1%.

The example that comes to mind is Microsoft Word vs. LaTEX – I won’t wade into the treacherous waters and voice an opinion, but I think everybody can agree that for someone who, say, only knows how to use a typewriter, the barrier to entry for MS Word is far lower. As long as you want to do what it was designed to let you do, it’s great – until you hit a corner case, at which point the ability of LaTEX to dive into the guts of formatting shines.

Some things I’m thinking about:

How small do the corner cases need to be for a tool to be more useful than annoying?   The problem is that innovation is all about exploring corner cases, so what corner cases are even acceptable in innovation tools? Could these tools perhaps be made in multiple layers?

An open question

An open question that I’ve been struggling with since the conference:

What deep-seated human –need- can space exploration fulfill for a massive number of people?

This question is based purely out of selfishness: space exploration fulfills my basic human need for greater meaning and excitement. But it doesn’t do that for the vast majority of people: it’s pretty clear that if millions of people shared my feelings about space exploration (and maybe even science in general) the world would be a different place.

I would argue that anything with sustained success addresses at least a part of some fundamental human need, whether directly or indirectly. Sometimes I have trouble justifying space exploration along these axes – yes, the awesome factor is a player, but things done simply because of their awesome factor have no sticking power. They’re like empty calories. Space exploration needs to find its veggies and meat.

This is why I want to find a way for space exploration (not just launching satellites) to become profitable. That goal still goes back to my opening question: what human needs can space explorers meet?