A heuristic is a "rule-of-thumb", advice that helps an AI program or human think and act more efficiently by directing
thinking in an useful direction.
Some of these heuristics are age-old wisdom, bordering on cliche, but
most are actually helpful.
If you want something done, do it yourself
Comment: Obviously true, and doing it is usually very good for your self
esteem. A surprising amount of work can be done this way, and experts are not
always necessary. However, there is a risk of becoming overworked if you try to
do everything yourself - we all need other people after all.
Never procrastinate anything you can do right now
Comment: Very powerful. There are many things that can be fixed or
solved with a minimum of effort, but are often pushed aside as unimportant.
Unfortunately they won't go away, and in time the feelings of guilt for
not having done them will make you even less likely of fixing the problems.
When you have several things you could be doing
and don't know which to do: Just do any one of them!
If you cannot decide between two or more possibilities, then there is a good
chance that the differences don't matter. However, most people begin to hesitate in this kind of situation (Fredkin's paradox). If you are conscious of this, you
can just choose one choice randomly or according to some standard method.
Always assume that you will succeed
Comments: If you don't expect to succeed in an endeavor, then you will
not do your best and will not notice possible solutions, while if you feel
that you will eventually succeed you will concentrate all your power at the
problem. Of course, there is no
point in attempting what you cannot do, a certain amount of self-knowledge is
If you can't find a solution, change the rules.
Comment: Remember that there are no no-win scenarios.
If you cannot do anything about something, there is no point in worrying about it.
Comment: Worrying is stressful, and in most situations doesn't accomplish anything - it just wastes energy. Instead of worrying about things, either
do something about them or find ways around the problem. One useful idea is to write down your worries on slips of paper, and then put them away in a box. Regularly, once a week or so, you open the box and see what you can do about the worries that are still relevant.
Do not rely on conscious decisions for speed - Just Do It
The conscious mind is surprisingly slow, conscious choices and actions
are delayed for a significant time (a reflex acts within some tens of milliseconds, an unconscious reaction to external stimuli circa 100 milliseconds and
a conscious choice several seconds). The duty of the conscious mind is usually
to inhibit rather than start action, and if you become too conscious of what
you are doing in a tense situation you will hesitate or slow down.
It is a good idea to learn to rely on your non-conscious mind, since our
conscious mind is slow and has very low bandwidth while the other systems
in our brains have a tremendous capacity and actually do most of the real
Don't try to explain away your actions for yourself
Comment: While we often do things we do not want to explain our real motivations for before other people (out of fear of embarrassment, anger or loss
of image), it is a bad idea to try to convince oneself that the motivation was anything different from what it was. It will only reduce your self-knowledge
with deliberate misinformation, and it is often valuable to understand what
motivations you have (even if you dislike them or would never admit them in public).
Listen to your intuition, but do not believe it unconditionally
Comments: Intuitive or emotional thinking, analogies, "gut feelings" or "flashes of inspiration"
can sometimes give fantastic new insights or show problems from a new direction. Unfortunately such thinking isn't always reliable, and quite often completely
wrong! Such insights should never be accepted because you admire their beauty or they are intuitive, only because they fit with reality.
Up to the Mental Enhancement Page
Up to the Transhuman Page
Anders Main Page
Anders Sandberg / email@example.com