davinci inventions.jpg

Last March, Paul Graham wrote an essay entitled “Frighteningly Ambitious Startup Ideas,” in which he spelled out seven massive problem areas/voids and ideas that could possibly fix/fill them. The adverb may be hypercorrection, but the essay isn’t hyperbolic, and it inspired throngs of intrepid entrepreneurs[1] to roll up their gingham and get crackin’.

Now that nearly a year has gone by, I thought it’d be fun to take stock of what’s been built. Of course, not all of the following applications and products are responses to Graham, but each of them are, I think, manifestations of the ideas. If you disagree, or have more to suggest, let me know in the comments!

The Idea: a New Search Engine

The way to win here is to build the search engine all the hackers use.

The responders:

  • Ark: People search, based on Facebook data. Now that Graph Search is here, I’m guessing Ark will need to harvest other data sources if it wants to stay in the game.
  • Izik: Izik was built from the ground up for mobile devices. CEO Richard Skrenta says mobile searchers are “more likely to be relaxed and in the mood to flip through many search results,” so Izik has swipeable magazine-style results and related topics.
  • Scholr.ly: Scholr.ly is a search engine for academic research and its authors. Results pages are split into two columns, one for publications and one for authors. Currently, it’s limited to computer science, but more branches are a'coming.
  • What Would Paul Graham Do: To answer queries, WWPGD plumbs the archives of Graham’s essays and HN comments. It’s maybe the simplest way to build a search engine all the hackers use. Suggestion: offer a trickier flavor: What Might Paul Graham Do, which would use machine learning to predict answers based on Graham’s corpus. 

The Idea: Replace Email

 I suspect that tweaking the inbox is not enough, and that email has to be replaced with a new protocol…the new protocol should give more power to the recipient than email does.

The responders:

  • Mailbox: Even if you’re not among the million+ who signed up for Mailbox, you’ve probably heard of the triage-focused iPhone app. Mailbox lets you swipe to delete, defer to specific date/time, or mark as done.
  • Mail Pilot: Desktop client Mail Pilot also approaches email from the todo list angle, but it categorizes all incoming messages as Incomplete. Users then mark them as done, defer them to a specific date, or move them to folders.
  • Skimbox: <shameless plug> Skimbox is a iOS app that sorts incoming mail into Inbox (people you have a history of responding with) and Skimbox (robots, strangers, Aunt Kathy’s FWD:FWD:FWD joke emails). Gander also uses swipe-based triage, allowing users to swipe messages to delete, mark as done, and move to skim.</shameless plug>

The Idea: Replace Universities

There will be many different ways to learn different things, and some may look quite different from universities.

The responders:

  • Thiel Fellowships: So, this one isn’t technically a response, as it started in 2010, but there is no higher profile example of the skip-college movement than Paypal founder (and Stanford alumn) Peter Thiel’s fellowships, which award high-achieving high school grads $100k, provided they don’t go to college.
  • Udacity: Udacity offers free college-level courses composed of short videos and interactive project. When it launched late last February, it had two courses, both in computer science. Today it also offers courses in mathematics, business, and physics. 
  • The Minerva Project: Minerva is billed as the place to get an online education that surpasses that offered by the country's top universities, for under half the tuition price. How: through combination of ace professors and constant individual assessment, plus a non-profit institute , led by former senator and New School president Bob Kerrey, that will "create new programs to finance students’ education and recruit top-level teaching talent."
  • Apprenticeship programs: Apprentices get paid—and paid well—to learn the skills required for a specific trade. Presently, only .3% of the American workforce are apprentices, but American University economist Rob Lerman told NPR he’s optimistic that number will grow. One example: Siemens expanded its four-year apprenticeship program, which ends with a guaranteed job, to the states last year. 

The Idea: Internet as an entertainment delivery vehicle

There are two ways delivery and payment could play out. Either some company like Netflix or Apple will be the app store for entertainment, and you'll reach audiences through them. Or the would-be app stores will be too overreaching, or too technically inflexible, and companies will arise to supply payment and streaming a la carte to the producers of drama.

The responders: 

  • Aereo: Aereo assigns each user an antenna that is connected to the internet. Users can then stream cable tv over their desktop computers, iPads, iPhones, Apple TVs (using Airplay), or Rokus. 
  • Netflix: Netflix’s first stab at original programming, the gangster-in-hiding series Lilyhammer, came out just before Thiel’s essay. House of Cards, its second, debuted last month. In both cases, Netflix released the entire season at once. Lilyhammer was successful enough to order a second season; House of Cards is considered by many critics to be an Emmy contender.

The Idea: Take over Steve Jobs’ vacant hardware visionary seat

If a new company led boldly into the future of hardware, users would follow.


  •  Bre Pettis, Makerbot: Is the future of hardware hardware that allows users to print more hardware (or guns, or food, or clothing)? Maybe. And while Makerbot isn’t the only 3D printing company around, it’s the first out of the consumer gate. Makerbot’s first retail store opened in September, and it started shipping its 3D home printers at the end of January. 
  • Sergey Brin Google[x]: Do I want to wear a location-sensitive, personalized computer on my eyes? No, I do not. Do many people? Yes. And do many developers see the Glass api’s potential? Yes. Along with Glass, Google X is also responsible for those famous driverless cars. 
  • Nadeem Kassam, Basis: Basis makes an activity tracker, wrapped in a watch. Its sensors capture heart rate patterns, motion, perspiration and skin temperature all day long, and display the results in an online dashboard.

The Idea: Software that makes Moore’s Law true again

It would be great if a startup could give us something of the old Moore's Law back, by writing software that could make a large number of CPUs look to the developer like one very fast CPU.

The responders: 

  • Meteor: Meteor is a pure javascript app development platform where code runs entirely on the client, with the client and server sharing the same API. Developers only need to write a particular line of code once, in one language, and it will be immediately pushed to the server. Meteor supports hot code pushes and latency compensation, so users always see the newest version of the app without having their work be disrupted.
  • Themis: Themis reduces MapReduce bottlenecks by minimizing the number of I/O operations. Themis’ implementation “reads and writes data records to disk exactly twice, which is the minimum amount possible for data sets that cannot fit in memory.” Themis is not commercially available right now, but I’m guessing someone with large pockets will jump on it, if they haven’t already.

The idea: Automatic, ongoing medical diagnosis

It will seem preposterous to future generations that we wait till patients have physical symptoms to be diagnosed with cancer. Cancer will show up on some sort of radar screen immediately.

The responders: 

  • Watson: The ‘bot that beat Ken Jennings and is now documenting neologisms at RIT is also being used as a virtual assistant for cancer doctors, researchers, medical centers, and insurance carriers. Because Watson can rapidly scan millions of cases, it can narrow-down or pinpoint diagnoses with greater efficiency and accuracy than a doctor can. It can also flag suspicious payment patterns.
  • Sequential decision-making models: Two researchers at the University of Indiana developed a machine learning-powered simulation model that understands and predicts the outcomes of treatment. When the researchers pitted their model against the actual doctor performance and patient outcomes for 500 patients, the model was over 50% cheaper ($189-per-patient vs $497-per-patient) and resulted in a 30-35% increase in patient outcomes. 

[1] Yep, including yrs truly