Career & Work
Undergraduation
paulgraham.com
Career & Work
16 min read → 60 sec
Advice for computer science undergraduates on becoming strong programmers, choosing worthwhile subjects, preparing for jobs, and deciding whether to attend grad school.
College works best when you follow genuine interest into hard problems.
That path builds programming skill, stronger hacking, and better odds in grad school.
02 / 20
003
Become a good hacker in two steps.
Get very good at programming. Learn specific cool problems.
03 / 20
004
Programming skill grows from hard work on hard problems.
Projects outside class usually teach it best.
04 / 20
Personal projects remove three common constraints.
INDEX
001
Novelty does not matter.
002
Permission does not matter.
003
Profitability does not matter.
05 / 20
006
Technical difficulty does not track prestige.
Games can be sophisticated. Databases can be trivial.
06 / 20
007
Graphics projects were especially promising.
Network games and 3D animation tools fit that bet.
07 / 20
008
Most hackers need only high school math and basic computation theory.
Math still pays off through metaphors.
08 / 20
Study fields where people work honestly on hard problems.
INDEX
001
Math.
002
Hard sciences.
003
Engineering.
004
History.
005
Architecture and classics.
09 / 20
010
Learn from smart programming tribes and professors.
Avoid research that is novel only because nobody cares.
10 / 20
011
The social sciences and much of philosophy teach less.
Fashions and weak methods drag them down.
11 / 20
012
A dropout graph can rank departments.
People switch from math to computer science, but not the reverse.
12 / 20
013
Language courses work better as extracurriculars.
Pair them with living where the language is spoken.
13 / 20
014
Real-world programming means larger codebases and existing code.
It also means defining the problem before solving it.
14 / 20
015
Open-source work teaches existing codebases.
Starting your own projects impresses employers even more.
15 / 20
016
Big companies reward Blub on Windows.
Cool small companies reward Ruby on Linux.
16 / 20
017
If you want to start a company, use the most powerful tools you can find.
Tools are your horse.
17 / 20
018
Grad school suits hackers more than it seems.
Much computer science research is hacking shaped into papers.
18 / 20
019
Top PhD admissions depend mostly on recommendations from known professors.
Tests and grades matter less.
19 / 20
20 / 20
Carouse
Turn any article, video, or podcast into a carousel you can browse in 30 seconds.