TITLE: I Can't Imagine...
AUTHOR: Eugene Wallingford
DATE: November 23, 2022 1:27 PM
DESC:
-----
BODY:
I've been catching up on some items in my newsreader that went
unread last summer while I rode my bike outdoors rather than
inside. This passage from
a blog post by Fred Wilson at AVC
touched on a personal habit I've been working on:
I can't imagine an effective exec team that isn't in person
together at least once a month.
I sometimes fall into a habit of saying or thinking "I can't
imagine...". I'm trying to break that habit.
I don't mean to pick on Wilson, whose short posts I enjoy for
insight into the world of venture capital. "I can't imagine" is
a common trope in both spoken and written English. Some writers
use it as a rhetorical device, not as a literal expression. Maybe
he meant it that way, too.
For a while now, though, I've been trying to catch myself whenever
I say or think "I can't imagine...". Usually my mind is simply
being lazy, or too quick to judge how other people think or act.
It turns out that I usually can imagine, if I try. Trying
to imagine how that thinking or behavior makes sense helps me see
what other people might be thinking, what their assumptions or
first principles are. Even when I end up remaining firm in my own
way of thinking, trying to imagine usually puts me in a better
position to work with the other person, or explain my own reasoning
to them more effectively.
Trying to imagine can also give me insight into the limits of my
own thinking. What assumptions am I making that lead me to have
confidence in my position? Are those assumptions true? If yes,
when might they not to be true? If no, how do I need to
update my thinking to align with reality?
When I hear someone say, "I can't imagine..." I often think of
Russell and Norvig's textbook Artificial Intelligence: A
Modern Approach, which I used for many years in class [1].
At the end of one of the early chapters, I think, they mention
critics of artificial intelligence who can't imagine the field of
AI ever accomplishing a particular goal. They respond cheekily to
the effect, This says less about AI than it says about the critics'
lack of imagination. I don't think I'd ever seen a textbook dunk
on anyone before, and as a young prof and open-minded AI researcher,
I very much enjoyed that line [2].
Instead of saying "I can't imagine...", I am trying to imagine.
I'm usually better off for the effort.
~~~~
[1] The Russell and Norvig text first came out in 1995. I wonder
if the subtitle "A Modern Approach" is still accurate... Maybe
theirs is now a classical approach!
[2] I'll have to track that passage down when I am back in my
regular office and have access to my books. (We are in temporary
digs this fall due to construction.) I wonder if AI has
accomplished the criticized goal in the time since Russell and
Norvig published their book. AI has reached heights in recent
years that many critics in 1995 could not imagine. I certainly
didn't imagine a computer program defeating a human expert at Go
in my lifetime, let alone learning to do so almost from scratch!
(I wrote about AlphaGo and its intersection with my ideas about
AI a few times over the years: [
01/2016
|
03/2016
|
05/2017
|
05/2018
].)
-----