Friday, July 26, 2013

Don't think!

TLDR: Concept of living simpler life without much thinking that will go to a black hole of ideas.

Why do I have to think that much? Brain is the most energy-consuming organ of a body so in principle - using it must be justified. Working with complex IT systems requires a lot of attention so no wonder why these jobs are paid above the average.

Computers are getting cheaper, software intuitive, education available. There are some buzzwords out there:
- Machine Learning
- Artificial Intelligence
- Computer Vision
- Natural Language Processing
- Neural Networks
- Optical Character Recognition
- Expert Systems
- ...

Why not outsource thinking to the machines? I believe it's already done in the most lucrative areas such as High-Frequency Trading. Algorithms vs algorithms and people with arrays of panoramic displays connected to Bloomberg terminals. Why not democratize decision-support systems? I'm really satisfied iPhone5 but at all times I'm considering switching to Android realm because of Google Now. Wish there was a XenClient-like solution that would allow me to switch seamlessly on a top of the range device such as Ubuntu Edge.


This post is about not thinking. You see my rationale, now a few real-life examples... In the past when in doubt I sticked with cheapest option. My time wasn't worth that much money. Now I value more comfort more and quite recently I paid extra for automatic transmission or anytime ticket that didn't force me to return after 7pm. Yet there are still situations when there is no clear winner - no gut feeling, estimated value equal in both cases and convoluted dependencies such as getting ticket on Friday relying on Monday weather. Here comes the explanation - I share a ride with my neighbor depending whether it rains or not. With many possible ticket options and limited bus services on Sunday it actually makes sense to sleepover in London (among 7 million people you find some friends) and go directly to the office in Cambridge rather than sleep at home.

Task: Wife rental

Here is the problem of delivering wife to the airport:

- reserve car, pay when collecting: 29£
- reserve car, pay online: 24£
- probability of wife willing to travel on that plane, that depends on:
  - weather on that day
  - our arguments just before
  - her dentist appointment
  - ... (unknown, undefined factors)

I actually decided to reserve and pay on the day as there were too many factors to consider, gut feeling telling me that probability is less than 80%. After serious deliberation we decided that she'll go and it was right decision. Eventually I paid 29£ rather than 24£ and theoretically I've lost five quid but I in general I made the right decision with the information I had (profitable in the long run).

What I would like to achieve is the system that would ask me some assistive questions (retrieve missing information and discover missing factors) and tell me the answer: you should do this and that.

It's an interesting scenario, and I think of the risk assessment system (especially credit card/loan system) as similar use-case (correct me if I'm wrong). I currently work in the area doing analytic on the mid-range data set with machine learning algorithm and see the opportunity ahead using related AI and analytic modelling system.

There is an app called - setting goals. So let's assume I'll set a goal of going to Mars and the system should be asking assistive questions and processing graph of dependencies and finally arrive with conclusion - learn physics, become vegetarian, print yourself in 3D.

Back to example of delivering wife to the airport - I should establish priorities - my happiness, her happiness, our time, spendings. I believe that having clear goal is essential in the long run. With some nuisances mental effort is not worth the money. Wish money was never a concern... (and we are working on it)


Samadhi - - "subject becomes one with the experienced object, and in which the mind becomes still, one-pointed or concentrated while the person remains conscious"