Home

Date: 2019-10-21

# Epistemology 📹0:00

The philosophy of knowledge and science
The nature of knowledge, what is real knowledge, what is false, how do you achieve it, how do we know that we know, …

# Uncertenty 📹4:24

Something whose outcome is not well defined and clear
Example: Weather, quantum states, …

It is easy to give an answer in certainty, but it’s not so easy to do it in an uncertain situation.

# Alan Turing 📹14:15

## Contributions

• The Turing machine
• State machine, infinite memory, process data from the memory
• Turin test
• Determine the level of AI (Artificial Intelligence)
• If a human is uncertain about whether the user one is talking to is another human or an AI, then the AI is intelligent
• Definition of AI
• Decoded the enigma machine

# Decision making in 📹19:43

## Certainty

1. Given Preferences
2. We formulate Goals
4. Actions have consequences Consequences with some probability

## Uncertainty

1. Given Preferences
2. We formulate Aspiration levels
3. That will inform Information gathering
4. Apon which we plan Actions
5. That have Uncertain consequences
6. That might prompt to redefine Aspiration levels (→ back to 2.)

In computer science Example: feedback loops, back-propagation in machine learning, try-catch
Loops are not for free, they need money, power, energy, time. At some time you need to stop.

# How can you innovate in uncertainty 📹38:10

• Iterate
• Parallelize
• Run as many instances as possible
• Prototype to fail cheaper

Looping is not always the way.
Example: If you continue to loop to improve the candle, you will never discover lightbulb

# Beyond bounded rationality 📹45:59

## Auto-piloting

Autopilot is the action of doing an activity that requires the minimum amount of energy to be completed. It is often a “no”-answer because it is easier to deny something instead of thinking about the request and its feasibility. But doing so, you are limiting potentially good actions.
We take and uncertain request to a safe certain response.

Autopilot is riskless and stable but it will always search for the local optimum but you will probably not reach the global one.

### Overriding the autopilot

Sometimes the autopilot is good, one cannot put effort into every very decision because they will be overwhelmed. But sometimes we need to take control and overriding the autopilot.
When to override? Example: When there is a problem, emergency situation, training (forcing to push yourself forward), opportunity that do not lie in a usual scenario, something big to do, in case of procastination

# The unreal 📹58:50

Overruling manual-pilot
Situations in which you are supposed to be in manual-control but suddenly you switch to autopilot.

Example: When you are studying but you get distracted by the surroundings, when you are driving late at night and your body decides to sleep even if you don’t want to

Autopilot blocks you to do an action that will improve your situation

# Competing pilots buying an appliance 📹1:11:55

If you add parameters, the complexity grows exponentially and you may decide to fallback to an autopilot choice Example When buying stuff you fallback to a brand that you used in the past, or you may take the decision only based on the price removing other useful parameters, or just buy randomly, or even not buy anything

# Dating 📹1:17:18

Is dating a rational or emotional process?

## Dating parameters

Autopilot 🤖: you know by instinct
Manual pilot 🤔: you need to sit down and think
Direct 📌: It is directly noticeable
Proxy 📡: you have to guess it based on the context

Parameter Auto / Manual Direct / Proxy Labels
Age 🤖🤔 📌 Autopilot & Manual pilot | Direct
Appearence 🤖 📌 Autopilot | Direct
Interests 🤔 📡 Manual pilot | Proxy
Gender 🤖 📌 Autopilot | Direct
Wealth 🤖🤔 📡 Autopilot & Manual pilot | Proxy
Personality 🤖🤔 📌📡 Autopilot & Manual pilot | Direct & Proxy
Education 🤔 📡 Manual pilot | Proxy

But if you think about online dating, often the Direct become a Proxy Example: Age is Direct in real life but become a Proxy online
Usually, uncertainty decreases over time in the real world, instead, in an online space the uncertainty fluctuates and is really slow to change compared to physical space.
The context adds some autopilot: the way you interact, the fact that both of the people are actively searching for a partner in an online dating app, the context decides the type of information you can or cannot expose, ….

# The troubles with digital transformation 📹1:39:07

You apply bounded rationality in something that is intrinsically uncertain.
A huge disconnection between how you think things should work and how the people that designed the system are making them work.
Transferring human activities into business/economic activities

# Summary 📹1:40:11

Philosophical theories applied in other industries and contexts.
We are applying notions in other fields of life.
We are opening new spaces and creating problems for other spaces.

We need to take seriously the topics about the uncertainty that we discuss in order to position ourselves much better in the world.
When creating something you need to be aware of what you are including in the system to avoid to build a bomb while trying to innovate.