What Can AI (and AlphaGo) Teach us About Being Human?

Dear friend,

I’ve been philosophizing a lot about AI and studying AI (especially AlphaGo– just watched the Netflix documentary and was super inspired). Anyways, the question I have is this:

What can AI teach us about being human, or how can we better leverage AI to help us become smarter, more intelligent, or “better”?

AI isn’t the end-game

How Alphago sees

The end-game isn’t for us to build an AI like terminator or sky net to destroy humans. The point is to use AI as a tool to empower us.

Or we can even use AI as a training tool to help us think better, to challenge ourselves, and to help us come up with more novel or innovative ideas/solutions.


What I learned from watching AlphaGO

There’s lots of stuff I learned from the AlphaGO documentary on Netflix. Some ideas:

1. What looks like a mistake isn’t really a “mistake”

This is a huge thing that is fascinating– when AlphaGo was playing LEE SEDOL, the human commentators saw AlphaGO play a move they called a “mistake”. But in-fact, it wasn’t a mistake. It was a move that AlphaGo used to leverage in order to win (AlphaGo was playing the long-game in order to win).

Philosophically this is fascinating– because of this:

What is a “mistake”?

The problem in the game of GO (Baduk in Korean) is that there are too many conventions. We become trapped in this mental prison of certain moves being “good moves” or certain moves being “bad” moves, or certain moves being “ugly moves” or certain moves being “elegant moves”.

However — once again, this is all based on convention. The point of the game GO isn’t to make good moves or to look elegant; the point is to win!

2. You just need to win by 1 point

In the game GO, this is the problem with human players:

We try to win by a large margin by securing lots of territory (which ends up being superfluous in the long-term).

AlphaGo seeks to win, even if only by 1 point!

Some practical human ramifications of this:

  1. In life, we don’t need that much money. You only need enough money for tools and services that empower you. You don’t need lots of money by a large margin, in order to compare yourselves to others, or to show others that you are “successful”.
  2. If you’re playing a game, or life, figure out what your end-game is. Then realize in order to win, you don’t need to win by a large margin. Even if you’re Usain Bolt and you’re a racer, you only need to win by 1 milli-second to win!

3. Challenge what is ‘good’ or ‘bad’

AlphaGo made a move against LEE SEDOL which commentators called “bad”. But what is “bad”?

Later on, when AlphaGo ended up beating LEE SEDOL in the match, THEN the commentators called the move as “original”.

Thus this was my takeaway:

A lot of truly original moves, ideas, or actions are often seen as “ugly” or “bad” — because people don’t understand it (yet).

For example in the high-jump, the first guy to have done it backwards was seen as being stupid, and that it looked “ugly”. Yet he was able to win, because he jumped higher!

Shaq had an “ugly” form with his free-throw, but actually improved his free-shot percentage with this “ugly” form.

Therefore this is my personal takeaway:

Don’t seek elegance or “looking good” — just do new things which are better, which help you win!

4. Confusion

A lot of commentators for the AlphaGo match were confused. Why? Because AlphaGO didn’t play like a “traditional” human and made a lot of moves which the commentators believed to be bad moves, but actually ended up being good moves in the end.

Therefore the moral of the story is this:

Ignore commentators.

5. Don’t change your style

The most upsetting part of LEE SEDOL vs AlphaGo is that LEE SEDOL ended up “losing his cool”, by changing up his style (in a negative way, which ended up hurting him).

The fascinating thing about playing against AI (according to one of the people in the film) is that you’re playing against yourself– it is like looking at yourself in the mirror, naked.

Which made me think:

Perhaps the benefit of AI is that it helps us look at ourselves in the mirror, to really help us better understand ourselves, our motivations, our actions, and our own human biases.


Conclusion

AI is beneficial for humans, because it helps us question ourselves, our motives, and it challenges traditional ways of thinking.

AI will help benefit us as humans to “think outside the box”, and challenge what is ‘good’ or ‘bad’, ‘beneficial’ or ‘harmful’.

But once again, let us think critically:

How can I best leverage AI to best help me, and humanity in general? (not the other way around).

ERIC


Machine Learning

Brave new world of photography: