Notes on talk by Jeevan Kalanithi, founder of sifteo

Stanford ugrad, then MIT media lab. After grad school, went to California to do design consulting type company (“just hung out and did cool projects”).

Some main points:

  • Hardware-software startup: definitely possible today
  • Be able to explain via video, no audio, in 30 seconds.
  • consulting vs starting a business: completely different


  • I can’t believe we thought twice about this: opportunity to do a talk at TED conference.
  • Regret minimization
  • CEO should care about three things (and maybe 8 more): money, people, vision.
  • when there is a hard problem, before asking for help make a proposal, doesn’t have to be completely right but helps things get done.
  • starting a company: have a cofounder
  • design inspiration:
    • Like to make decisions based on “culture” as opposed to technical architecture.
    • Apple, nintendo
  • it’s a mistake to think that you’re coming up with something new from scratch, it’s more like drawing from many sources of inspiration

About a hardware company

  • make hardware, business model is simple: make something, and sell it for more than what it cost
  • More complicated and more expensive
  • Software iterates on two week basis, hardware iterates more on like quarterly basis
  • Long term view is very important
  • minimum viable idea can be distracting
  • When a team is large (>= 5 people) it really needs to define when something is successful/done and when it isn’t
  • dragon innovation

Hardware creative process

  • start playing with a mockup, i.e. don’t immediately start laying out a circuit board
  • make a mock up, play with it, will give you an idea of the produce that you want to make
  • while a prototype can be expensive, but cost really matters in the final product
  • don’t jump into building it too quickly, and iterate because it’s hard to get it right the first time

Hardware interface

  • prototypes done in freemont
  • facilities and know how is abroad, it’s not in the US
  • mistakes by manufacturers aren’t necessarily lack of diligence, just not enough bandwidth for a conversation
  • would love to talk to someone who can do high volume in the US, but just doesn’t think it’s possible

Q: IP differences between hardware and software

  • IP how important is it to your company? biotech 10, webapp, 2, hardware 5-7.
  • BUT patent will not make or break sifteo

Q: distribution

  • getting them to stores, it’s important: yes people are shopping online, but it’s becoming true more slowly than mostly everyone thinks

Q: having an SDK

  • a guy found a leaked SDK and made an application to make sentences
  • then litle girl learned how to make sentences, rewarding.

Q: Competition

  • red ocean: bloody ocean, people fight on features. blue ocean: virgin territory, not fighting competition on the features.