After reading Paul Graham's essay on Frighteningly Ambitious Startup ideas, I asked the same question on Quora and here are some of the answers. Go read them, I'll wait.... As the person who asked the question, I cannot answer directly on Quora, so let me put a few thoughts: First, much like the person who received the most credits on her answer, we have to be focused on what this means:
**'Frighteningly ambitious' means different things to different people. Paul Graham refers to 'frighteningly ambitious' ideas as ideas that are so big that they are terrifying. Not just because they seem like a lot of work, but because they threaten your identity. Ideas that say "even if you got me, you wouldn't know what to do with me." Ideas that are just on the right-side of impossible. This list reflects ideas with solutions to big, hard problems. Ideas either disruptive in nature, have the potential to take out huge bully companies that dominate entire markets, or solutions to problems that just seem impossible to solve
For those of you reading this, most of these ideas listed here are rooted in some of the discussions mentioned on this blog. So here we go:
* Better Story Telling: Ever since the dawn of time, we are story tellers. Inside that category I list three different ideas:
- A better search engine, or better yet an insight engine. I agree with Paul's first idea, improvements above what Google currently provides is needed. Nuit blanche is one of the few blogs to consistently write or report about certain aspect of a very technical subject, yet, the search for elements in the blog is not even crude A search engine should provide context.
- There is a need for tools aimed at the discriminate blogger or story teller.
- An AI-like engine having a conversation with you about where you are with regards to your projects. The idea is to decrease the amount of time it takes to reemerge back into earlier projects. The switching time between projects is the real killer.
Industries/Products being disrupted: Search Engines, Email, ToDo Lists, Media Companies...
* Communicating with Animals : One word. Reddit. It's pretty much on the right side of impossible yet any dog/cat owner can tell what their pets are saying. It seems obvious to me that some machine learning using some supervised learning could do much better than what we have right now, which is ... nothing.
* Cheap Recovery of Failed Hard Drives: How come there isn't a somewhat cheap imaging technique that allows one to copy hard drives. How come this task requires more than $1000 ? Most people who want access to old hard drives have family photos on them, They care only about getting something back, they don't mind if it takes a month or a year to do it. They also have copies of some of these photos and documents somewhere. Industries/Products being disrupted: HD recovery companies that charge upwards $1000 for HD recovery
* Robots for sorting landfills or dumpters : Robots have done well with regards to mining technology, how come there is no solution at the end of the spectrum for robots to mine landfills ?
* Real Telepresence: The business travel industry can be seriously disrupted with the right approach.
* New Types of Sensors: Before building an ongoing diagnosis -Paul's 7th idea-, you need new types of sensors first (compressive sensing devices of all kinds)
- Such sensor needs to be non conspicuous yet records information all the time (including sleep) and allow easy annotation.
- Instant information decoding/diagnosis is not necessary.
- A sensor that seamlessly obtain information about food intake and dejections
- A sensor that would remove the need for building a camera infrastructure inside a building/house
- Some sensors might use the concept of Imaging With Nature.
Other ideas include Stopping Tornadoes, Nation Building..
What are your ideas ?
Liked this entry ? subscribe to Nuit Blanche's feed, there's more where that came from. You can also subscribe to Nuit Blanche by Email, explore the Big Picture in Compressive Sensing or the Matrix Factorization Jungle and join the conversations on compressive sensing, advanced matrix factorization and calibration issues on Linkedin.