[ExI] mit's answer to the stanford ai class

[Adrian Tymes]

On Thu, Dec 22, 2011 at 2:31 AM, Kelly Anderson <kellycoinguy at gmail.com> wrote:
> On Wed, Dec 21, 2011 at 3:51 PM, spike <spike66 at att.net> wrote:
>> How well I know it.  In 2002, I had a crazy-smart physicist working with me.
>> I showed him how to use excel.  In 2002!  He thought it was the greatest
>> invention since sex.  That was the last we heard of him for months, as he
>> did what he was famous for doing: super obsessive laser beam study of a
>> topic.  He wrote some interesting software in the macro language.
>
> I had a professor Evan Ivie, brilliant guy, and a great computer
> scientist, worked at Bell Labs with Kernihan and Richie and the gang.
> He spent quite a while figuring out what the limits of programming in
> the shell were... similar kind of story, I'd imagine. When all you
> have is a hammer, suddenly, the whole world is a nail!!! LOL

This is a problem in itself: knowing what tools are out there, that you've never
heard of, that can adequately complete entire sections of your new project -
even professionally, as in "handle details you haven't thought of yet because
you just now discovered the need for this, but you'll have to take care of to do
the project".

This is a classic problem with government contracts, for example.  Myriad
reporting requirements that are unfamiliar to newbies to the government
space, and often, you stumble your way through them only to not actually
win the bid - but you know what to do next time, you think.  And yet, each
of the requirements made sense from the point of view of someone who was
dealing with a problem that cropped up in the past, and there's been over
200 years of institutional learning on this (in the US - more in many other
countries, and the US sometimes copies that as needed).

Or, take an example from just last night.  A friend of mine was converting
logs from a chat tool for posting on the Web.  A long, tedious task of
cleanup...until I told her that I'd made a tool to do exactly that, and would
she like to borrow it?  What once took an hour now took a few minutes,
and all that changed was being informed of the relevant tool.

And in most Web programming jobs, when running into any novel task,
the first thing to do is to search the Web for anyone else who's solved
that task and posted public domain source code you're free to copy and
use.  For example, I would venture that most people on this list do not
know by heart the Luhn algorithm, which is used to determine whether a
credit card number could be valid or is just a typo.  (This algorithm is
100% protection against mistyping a single CC digit with another digit,
and good protection against other CC number mistyping errors.)  But
yet, just knowing that it exists and what the name of it is, any of you
can easily google for "luhn algorithm [LANGUAGE]" for any modern
programming language and find examples of it.

This requires a bit of a change in thinking, from "how do I solve this
problem" to "how do I find a solution to this problem".  Of course, not
all such challenges have solutions out there to find, so you have to
figure out when to stop looking and start solving it yourself.