Paul Graham’s talk at Startup School 2005. Some notes I jotted down..

General Perception -> Idea for startup very hard..Million dollar idea.
Not lying around for anyone to discover. Most startups end up nothing
like the initial idea. In the process of building up you come up with
the real idea.

Present question instead of idea. Instead of “how about a
collaborative web based spreadsheet” vs. “is it possible to make a
collaborative web based spreadsheet”. Incrementally implementing your
way towards the final idea.

Be familiar with promising new technologies. Conversation with
friends…right kind of people. Ideas get developed while explaining
them to the right kind of people.

Let your mind wander. Harder fields are a better area of work for
generating ideas…eg. math. Related fields..Comp Sc and Electrical
Engg…importing ideas from related fields. You have to be working on
something..you cant let your mind wander at random. you have to be
working on something and then let your mind wander a “little bit into
randomness”.

Problem –>Denial of problems. People believed that spam cannot be
avoided..except for rule based methods. Even when presented with
solutions ppl believe they wouldnt work. To accept the problem as
intolerable and feeling that it must be possible to solve them.

Make something people want. Think in terms of users..how would it
benefit you..your friends..your cousins etc. Build stuff for yourself.
Look at stuff people use now which are broken. Take a luxury and make
it into a commodity. If ppl are willing to pay a lot for it..it must
be useful to them. Henry Ford. This might involve redefining the
problem. Model T didnot have all the features of luxury cars…but it
solved the problem ppl cared about…getting from place to place. Make
things easier to use. Simplicity takes effort.

Do one thing and do it well.

Some kind of exit strategy. Either get bought or go public. Only then
will you be able to attract talent (with stock options). Look at big
companies and think what they ought to be doing. Make something
multiple acquirers would want. Dont make Microsoft Windows fix. Make
something where there is competition.