Predictions about technology
are usually right
about direction
and badly wrong
about timing.
That is not a small detail.
That is the entire difference
between a good decision
and an expensive one.
The idea arrives.
It is obviously the future,
and it obviously is.
But obviously takes seven years,
and you have to survive
all seven of them
inside a company
that is measuring quarters.
The team that rebuilt
on the obvious future
in year one
spent two years
writing the tooling
that did not exist yet,
filing bugs against
a project with four maintainers,
and explaining to their manager
why the old way
was still faster.
They were not wrong.
They were early,
and from the outside,
in the moment,
those two things
are indistinguishable.
The lesson is not
to ignore what is coming.
It is to separate
learning from committing.
Learning is cheap.
Read it, build a toy,
run it in something
that does not matter,
keep a rough sense
of what it is good at.
Committing is expensive,
and the honest question
for committing
is never "is this the future".
It is "who has already run this
in production for a year,
and can I hire someone
who has done it before".
If nobody has,
you are not adopting.
You are doing unpaid research
on behalf of your users.
Keep your bets reversible.
A new thing at the edge
of the system
is an experiment.
The same new thing
underneath everything
is a marriage.
And when it does arrive,
which it will,
the person who quietly learned it
two years early
and did not bet the company on it
is in the best position
anyone can be in.
They know it,
and they did not pay
for knowing it early.
There is one exception
worth naming.
Sometimes being early
is the whole business,
and the risk is the point.
Fine.
Just say that out loud,
in the room,
so that everybody knows
which bet they are on,
instead of discovering it
two years later
when the tooling
is still not ready.
Be interested early.
Be committed late.
– Serguey Asael Shinder
Top comments (0)