There was a lot good in Thiel's lecture.
One point he made at the end was good: Look for things
that are new. Indeed, venture capital very much needs
what is quite exceptional at least in exit value but
also in whatever the heck it is, likely new, that leads to that value.
But it seems to me that in the quest for work that
is now, powerful, and valuable, Thiel is overlooking something
both big and important and, actually, close to where he
lives in SV: He is overlooking some of the
crucial aspects of how SV became important, i.e., got there.
And what was that? Sure, US national security via
decades of high spending on electronics for aerospace,
yes, in SV.
I do agree with Thiel that it is better if a founder
has a really good idea for a startup and has this
without "get out of the building" feedback from
customers responding to agile developments.
Okay, here's something big, solidly in the background
of SV, that Thiel didn't emphasize: Suppose a founder
thinks of a big problem, that is, a big need.
Next, he wants a good solution: So, he wants, say,
the first good or a much better solution to the
big problem. And with that solution he wants a
monopoly. Good.
Now, how to get that solution? Well, SV long
showed an important way -- research. And the
researchers do not necessarily have to settle
for Thiel's Y = 0%.
So, for US national security we have a long list
of fantastic solutions, from research, heavily
done in SV, for
really challenging problems. Point: Such
work can be done. That is, when want a
really good solution to a big problem,
in the case of US national security,
research in SV has a terrific track record.
Yes, some of that research was expensive,
but not all research is! A lot of the best
research is one guy with paper and pencil.
E.g., did Thiel mention Einstein?
But, wait, there's more! Now SV likes
computing, that is, exploiting Moore's law.
And Thiel said he liked software. Okay,
it's about information. And for techniques
to take in available data, manipulate it, and
put out valuable information? Sure, and may
I have the envelope, please (drum roll):
Right, and the winner is applied mathematics,
commonly done with paper and pencil, by one guy.
Then convert to software and proceed! For
US national security, SV did a lot of that.
Well, then, since such work can be done,
do such work on commercial problems.
That way, get such good technology and have a monopoly,
the first good or a much better solution, much
better technology, etc. I didn't hear Thiel
explain the promise of this approach and thought
that he should have.