How to Invent the Future I: class 9

CliffsNotes + commentary on Startup School

Class with Alan Kay.

I encourage everyone to watch the full lecture (58 mins, link at end of this post), but if you are just curious or want a 5 minute refresher, here we go:


If you want to make money, don’t bother with a startup, create an industry because then you get trillions instead of billions — not going incrementally from the present but carving out an entirely new area.

You want to invent, not innovate. To do that you need to escape the tyranny of the present. If you work off of the present, you work will be incremental.

“School is the best place ever invented to keep you from thinking about something important for more than a few minutes.”
— Marvin Minsky

Picasso said “Learn the rules like a pro, so you can break them like an artist.”

You have to learn everything and find a way to forget it so you can have your own ideas.

Xerox PARC 1973, we created the Xerox Alto, an early computer. We made 2,000 of these. It led to the development of the Mac. We had 25 researchers in 5 years develop the first modern personal computer, graphical user interface, desktop publishing, WYSIWYG (what you see is what you get), object oriented programing, laser printer, fonts & postscript, Ethernet, peer-to-peer and client server, and the Internet.

Cost: $10–12m / year.
Return: $40 trillion total.

“Real” Xerox PARC ended in 1983 when they fired Bob Taylor. If large companies were rational, then startups shouldn’t exist.

Since then, it’s been extraction but almost complete lack of curiosity and interest in the processes that produced them.

The world is different now. Read “The Dream Machine”. There had been a long continuity that led to PARC. The Manhattan Project, MIT Radar Project, MIT Whirlwind Project brought scientists and engineers to work together.

Ivan Sutherland created graphical sketches on computers in one year because he didn’t think it was hard; he was just working on the problem.

J.C.R. Licklider “Lick” was the leader of the ARPA and he developed 16

1. Picking an idea worth dedicating your life if necessary.
2. Forget about goals; focus on visions.
3. Fund people, not projects.
4. Only fund the very best people.
5. It’s a research community, not a research project.
6. Fund problem finding, not just problem solving. Figure out what the actual problem is.
7. Milestones not deadlines.
8. Baseball not golf. If you strike out, you don’t cry. Batting .300 is great.
9. You can’t think inside the Beltway (Washington D.C.)! There’s too much noise.
10. It’s about shaping computer stuff to human ends per the vision.
11. If you can make your own tools, HW and SW, then you must! Otherwise you are working on some other vendor’s product.
12. Software drives the hardware, rather than vice versa.
13. Make enough of the inventions so that more than the inventors can use them. PARC, whatever you did, you have to make over 100 of them.
14. Argue for clarity, not to win.
15. Develop people for the long term. Grad schools admitted students who were interesting, but didn’t make judgment on them until a few years down the road.
16. The reaching is the reward.
“You miss 100% of the shots you don’t take.”
“A good hockey player goes to where the puck is; a great one goes to where the puck will be!”
— Wayne Gretzky

Escaping the Present

  1. “Cosmic goodness intuition” — start with an intuition about what products might exist and would be useful.
  2. “Identify favorable exponentials” — see if there are fundamental shifts in technology that would enable your intuition.
  3. Take the cosmic intuition out 30 years. Extrapolate what 30 years of the shift might be. For example, 30 years of Moore’s Law would create 32,000 times more computing power (if doubling every 2 years).
  4. Can we say “it would be ridiculous if we didn’t have this” after 30 years.
  5. Now, bring that 30 year future idea back to 10–15 years out. Can something be done now to reach that 10–15 year version?
  6. At this point you can just pay money. If you pay a lot of money now, you can get the commodity computer of 10, 15 years in the future.
  7. You make a bunch of them and intertwine them, i.e. run the software.
  8. Experiment with software applications to design products.

Watch the full video: