Here is the answer.

1. Divide.

The idea that there is such thing as a problem that is infeasible due to technical complexity is an illusion. Everything that appears complex is recursively made up of simpler components, and once you know those components — the fundamental concepts that anchor the problem (and potential solution) in the realm of possibility — the illusion vanishes.

The trick is knowing what these components are or knowing someone who does – the more knowledge you have here, the further you can progress before encountering an indivisible sub-problem.

2. Conquer.

Once the problem is broken down into its fundamental requirements, it’s much easier to understand which parts are easy or already solved, which may require some research, and which are new.

A problem is almost never unique although it may appear unique on the surface. By pattern matching on historical work, much of the solution complexity can be delegated.

So, now what’s left is a set problems that you’ve never seen before.

3. Play.

Playing with new problems is often the quickest way to become acquainted with them. People tend to get caught up in analysis paralysis or worrying about the “correct” way to do something. This is all overhead to finding a solution.

Get dirty, make mistakes — the path to a novel solution is not paved.

http://www.apple.com/stevejobs/

If you would like to share your thoughts, memories, and condolences, please email rememberingsteve@apple.com

I’ve just finished reading The Big Short on recommendation from my friend Marco. It’s an easily digestible introduction to the causes of the recent financial crisis. I highly recommend this book if you’re at all curious about the crash’s back story.

In this book: CDO, AIG, credit default swap, Goldman Sachs, Morgan Stanley, JP Morgan, Salomon Brothers, Lehman Brothers, LehmanLive, asset backed security, mortgage bond, ARM mortgage, negative-amortizing adjustible-rate subprime mortgage, S&P, Moody’s, tranche, AAA, BBB.

Hilarious.

In your api controller:
render :json => @items.to_json, :callback => params[:callback]

Result:
awesome(["item1", item2"]);

This gives you cross-domain JSONP. The callback variable in the URL is optional.

There’s no need to roll your own render_json.

It’s kind of pollution-y here. I still need to figure out if I’m headed straight for Microsoft Research or heading to my accommodations first.