Thinking Fast and Slow

Linda Rising
[email protected]
www.lindarising.org
@RisingLinda
What is thinking?
Thinking going on here?
 Commonalities
 Staring, focused concentrated attention
 Pressure to produce, decide as quickly as
possible
 Don’t stop or take a break – that would
mean you’re not really working
 This is what passes for thinking

4
 Behavioral economics
Psychologist Daniel
Kahneman won the
Nobel Prize for
“Economic
Sciences” in 2002.

Kahneman &
Tversky changed
the way we think
about thinking.
5
 My goal: useful tips
Cognitive science is fast growing, uses controlled
experiments, on-going re-testing, explosion in
results
Our field is slow growing, few, if any controlled
experiments (mostly anecdotes, case studies)
We should pay attention. It might help us do our
work better !
 Which System?
System 1 System 2
Unconscious (runs 24/7) Conscious (on-line when awake)
Fast Slow
Intuitive Rational
Can multi-task Linear
Remembers everything Forgetful
~11 million bits/sec ~40 bits/sec
95% of cognitive function 5% of cognitive function
BUT inaccessible Chatters constantly while we are
awake

Our gut feeling, source of insight We identify with it, believe it’s in
charge
7
8
 More on System 1
 Heuristics, expertise, automatic mental
activities of perception and memory.
 Associative memory continually creates
a story interpreting what is going on at
any moment--your personal story where
you are the hero.

9
 System 1 Heuristics, biases, errors

Confirmation bias: tendency to search for, interpret,

favor, recall (it affects memory) information to confirm
our beliefs. We seek confirmation, not information.

Cognitive dissonance: struggle to hold two disconfirming

ideas at the same time.

Naïve realism: belief that we are rational and those who

disagree with us only need “facts” to see our side.

10
 Any help for cognitive bias?
 Hardwired
 Difficult to overcome on a personal level
 So easy to see in others

11
 The result: severe limitations on
thought
 Excessive confidence in what we believe we
know.
 Inability to acknowledge the extent of our
ignorance and the uncertainty in the world.
 Overestimate how much we understand about
the world and underestimate the role of
chance (randomness) in events. We like
“patterns” and explanations.

12
I'm considered one of the worst offenders on many of
these errors. I'm overconfident, when I preach against
that. I make extreme predictions, when I preach against
that. Some people read Thinking, Fast And Slow hoping it
will improve their minds. I wrote it and it didn't improve
my mind. Those things are deep and powerful and hard to
change.

Daniel Kahneman

13
 More on System 2
 Allocates attention to effortful mental
activities.
 Associated with action, choice,
concentration.
 When fully engaged can only focus for
50 minutes (max).

14
 How we learn
 New skills are acquired by System 2. With
practice/experience, we can move them to
System 1 as they become automatic and we
become an expert.
 Examples: walking, driving, playing a musical
instrument, engaging in a sport.
 When we “overthink” in a domain where we
have skill we can get in the way of our
expertise.
15
 Models are not reality
 Kahneman’s model is an abstraction
 All models are wrong but some are useful.
George Box, statistician
 Kahneman made no effort to map Systems 1 &
2 to brain regions.
 Except for well-known areas of the brain
involved in sensory and motor functions,
almost every part of the brain is involved in
almost everything the brain does.
16
Jonathan Haidt & Heath Brothers

17
 System 2 takes energy
 Conscious decision-making (System 2) takes time and
energy (about 25% of your intake).
 Difficult cognitive reasoning or self-control causes
drop in blood glucose.
 “Pay attention.” There is a cost to using System 2.
 It’s an illusion that System 2 = thinking.
 Don’t waste conscious effort

18
 Limitations of System 2
 We have a limited pool of “mental energy.”
 This is why conscious multi-tasking is
impossible for us
 This is why we make bad decisions when we
are tired or hungry
 This is why we need to take breaks

19
 How it works
 System 2 normally in a comfortable low-effort mode,
using a fraction of its capacity.
 System 1 continuously generates suggestions:
impressions, intuitions, impulses, intentions, feelings.
If endorsed by System 2, impressions and intuitions
become beliefs; impulses become voluntary actions.
 When all goes smoothly (most of the time) System 2
adopts System 1’s suggestions.
 When System 1 has difficulty, it calls on System 2 for
processing.
 Who’s in charge?

20
 System 1, of course!
 System 2 believes it runs the show, but
System 1 is in charge.
 It's good to have System 1 in charge of
stuff -- it would be bad (as David
McRaney says) to hand the keys
to the pancreas over to System 2 who
routinely forgets your phone in
restaurants.
21
Cake or Fruit Salad?

22
 Make life simpler
You'll see I wear only gray
or blue suits, [Obama] said.
I'm trying to pare down
decisions. I don't want to
make decisions about what
I’m eating or wearing.
Because I have too many
other decisions to make.

Fast Company 2/12/2014

23
 Rule #1- don’t do it
 If possible—don’t spend time on it!
 OR hand it off to System 1
 Professionals spend an average of 9
minutes choosing their office attire.
OfficeTeam survey of 2015 workers in North America.

 System 2 is a limited resource, do not

waste it.

24
 Stories of using System 1
 Solving math problems; debugging code
 “I back away from conscious thought and turn the
problem over to my unconscious mind. It will scan a
broader array of patterns and find some new close fits
from other information stored in my brain.”
Arthur Fry, co-creator of the Post-it Note
 The question I never asked myself: How long do I have
to work on something before getting insight?

25
 My problem-solving approach
1. Define problem. Say it out loud, write it down.
2. Enough data?
3. Invest minimal amount of time solving it (possibly
adjust answers to 1 & 2) – no more than 10 min.
4. If solution is not forthcoming, leave it (one or more of
the following):
- work on another task
- stand up/sit down, stretch
- bio break
- longer break, e.g. exercise, meal, errand, sleep
5. If no insight, repeat steps.
6. System 1 is not always “right.” System 2 makes final
decision.
26
You may have more knowledge, data,
information than you realize. It’s better to
go ahead with what you have than to wait
for a complete understanding, which may
never come and might not be necessary.
Remember that System 1 is already
working on it.

27
 System 1 is creative
 The problems you solve using System 1
are not binary solutions (should I do this
or not) but can be complex, innovative
approaches.
 Songs and novels have been written this
way.
 Story of finding music.

28
 Multi-task? Sorry !
System 2 cannot multi-task but must context
switch as does any linear processor
Heavy multi-taskers have been shown to have no
beneficial abilities and to be suckers for
irrelevancy
You can’t compute 17 x 24 while making a left
turn in heavy traffic. You can do several
things at once, but only if they are easy and
undemanding, so you can turn most of them
over to System 1.
29
 This includes habits
 Defining a process in advance not only
hands the work over to System 1, it saves
wear and tear on System 2.
 Travel, hotel rooms, purse/briefcase
 Sometimes this means changing habits, a
difficult endeavor.

30
The best book on changing
habits

31
 When to use System 2
 Complex tasks, e.g. income tax reports
 Reading. Taking notes.
 Practicing anything.
 Meetings, lectures, podcasts.
 Anytime you have to “pay attention.”

32
 Flashes of Insight
When System 2 is trying to solve a complex
problem, often you need to distract it. If System 2
stops working on the problem, then System 1
(which has been working in the background) can
potentially offer up results. Doing a different,
unrelated activity, e.g. taking a shower, going for
a walk, mind wandering, sleeping, allows System
1 to intrude. These ‘flashes of insight’ originate in
System 1.

33
 Do “Nothing”
The areas of the brain involved in (prospect)
thinking about the future – insight – occurs
during idle time – are the same areas of the
brain involved in thinking about what others
are thinking.
Right temporoparietal junction RTPJ, one of the
areas that’s most different in our brains.
It takes a long time to develop ~age 5.
We need to take time off, let the mind wander,
listen to others, take another point of view. 34
I insist on a lot of time being spent, almost
every day, to just sit and think. That is very
uncommon in American business. I read
and think. So I do more reading and
thinking, and make less impulsive decisions
than most people in business.
Warren Buffett

35
“Let’s go for a walk.”

Steve Jobs

36
 The Ultimatum Game
In pairs of subjects, Player A is given money (e.g.
$10) and makes an offer to Player B. If B
accepts offer, both keep the money. If B rejects
offer, both get nothing.
Typically, low offers (e.g. $2) are rejected.
Research shows that if players take a 10-min
break, low offers are accepted, so TAKE TEN!

37
Virginia Apgar, M.D. (1909-1974)

38
 Is an ulcer malignant?
 Simple algorithm – 7 equally weighted
variables predicted cancer well
 Algorithm compared with expert
diagnosis of x-rays of 96 ulcers
 Physicians’ diagnoses all over the place,
inconsistent on hidden repeats (errors
20% of the time), little agreement

39
 Algorithms win
The same inputs generate the same outputs every
time. They don't get distracted. They don't get
bored. They don't get mad. They don't get
annoyed. They don't have off days. They don't
fall prey to the litany of biases humans do.

The algorithm doesn't have to be complex. As

demonstrated with the radiology example, simple
rules work well.

40
A study of 101 independent auditors asked
to evaluate the reliability of internal
corporate audits showed similar
inconsistencies.

A review of 41 studies of the reliability of

judgments of auditors, pathologists,
psychologists, organizational managers,
and other professionals suggests that this
level of inconsistency is typical, even when
a case is reevaluated within a few minutes.
41
The surprising success of equal-weighting schemes has an
important practical implication: it is possible to develop
useful algorithms without prior statistical research.
Simple equally weighted formulas based on existing
statistics or common sense are often good predictors of
significant outcomes.

Daniel Kahneman

42
 Small Steps
We have a sense that large, complex problems require
large, complex solutions.
Organizations and software are examples of complex
adaptive systems where the impact of any change is
difficult to predict
The often-overlooked approach of Small Steps (it’s a
pattern!) is a better road to success where the result of
each small experiment can be seen before the next step
is taken

43
 Continuous experiments
Ask System 1 questions and wait for an answer,
e.g. “How can I find more time for practicing?”

The answer, when it comes, may not be definitive,

but can suggest a small experiment that likely
leads to more questions and more experiments.
This process never ends.

Thanks for listening!

44
 References
 Brain Rules, John Medina, brainrules.net
 The Art of Changing the Brain, James Zull
 Any book by Dan Ariely
 http://youarenotsosmart.com/podcast/
 Thinking Fast and Slow, Daniel Kahneman

45