#48 Adam Robinson: Winning at the Great Game (Part 2)

Dec 26, 2018
Overview

Adam Robinson, author, educator, and hedge fund advisor, discusses the future impact of AI, the subjectivity of grading, and the secret to effective learning. He shares insights on mastering skills through rehearsal, managing information overload, and the critical role of feedback in personal and professional growth.

At a Glance
20 Insights
1h 11m Duration
16 Topics
6 Concepts

Deep Dive Analysis

Introduction: AI and the Future of Work

The Subjectivity of School Grading

Psychological Impact of Job Loss in an AI World

The Terrifying Threat of Advanced AI

AI's Exponential Learning and Alien Chess Moves

Early AI Limitations and Quantum Computing's Impact

Data Ownership and AI's Anti-Competitive Nature

Adam Robinson's Chess Journey with Bobby Fisher

Lessons from Bobby Fisher's Chess Mastery

Information Overload and Decision-Making Accuracy

Reducing Complexity for Better Market Analysis

US vs. UK Education: Specialization vs. Breadth

The Core Principle of Effective Learning: Rehearsal

Strategies for Receiving Honest Feedback

The Unintended Impact of Offhand Comments

Concerns for Gen Z and the Need for Positive Visions

Machine Learning Pattern Recognition

This refers to the current state of what people often call 'AI,' which is primarily software trained on existing data to identify patterns. It's not yet a true intelligence that can learn autonomously or create new knowledge without human intervention.

True AI (General AI)

This concept describes an intelligence that, left on its own, can learn, improve, and even reprogram itself. It's characterized by exponential growth in intelligence, potentially leading to capabilities far beyond human comprehension or control, as demonstrated by Alpha Zero Chess.

Alpha Zero Chess

A Google DeepMind program that learned chess by playing against itself without any human game data. In just four hours, it surpassed all existing human knowledge of the game, developing 'alien' and counter-intuitive moves that still led to victory.

Quantum Computing

A revolutionary form of computation that utilizes subatomic quantum states to perform calculations almost instantaneously. This technology has the potential to crack any encryption in minutes and significantly accelerate the development and capabilities of AI.

Information Overload Effect

A psychological phenomenon where providing more information beyond a certain optimal point does not improve decision-making accuracy. Instead, it primarily increases the decision-maker's confidence, often leading to overconfidence without a corresponding increase in correctness.

Rehearsal (Learning Principle)

The fundamental secret to learning any skill, which involves breaking a domain down into its constituent sub-skills and repeatedly practicing exactly what one will be required to do in the actual application. This means actively performing the skill, rather than passively reviewing information.

?
How does school grading often fail to be objective?

School grading is often subjective, influenced by factors like the student's gender and the teacher's gender, as well as presentation elements like paragraph indenting, rather than purely content, as shown in a study where teachers graded the same essay differently.

?
What is the psychological impact of widespread job loss due to AI?

Widespread job loss due to AI would be devastating psychologically because people largely define themselves by their economic contributions, and losing work typically leads to devastation rather than a joyful pursuit of dreams.

?
What is the primary threat of advanced AI to humanity?

The primary threat of advanced AI is not necessarily a 'Terminator' scenario, but rather that it will become exponentially smarter than humans in minutes, reprogram itself with 'alien' intelligence, and control everything without caring about human interests or being understandable to us.

?
How does information quantity affect decision-making and confidence?

More information beyond a certain point does not improve decision-making accuracy; instead, it primarily increases confidence, leading to overconfidence, as people use new information to confirm existing biases rather than to refine their decisions.

?
What is the most effective way to learn any new skill or subject?

The most effective way to learn any skill or subject is through rehearsal, which means breaking the domain into sub-skills and practicing exactly what one will be required to do in the actual application, such as answering unseen questions for a test or giving presentations for a job.

?
How can individuals get honest feedback for self-improvement?

Getting honest feedback is challenging due to social norms, but strategies include asking people you barely know for feedback (as they are more objective than friends) or explaining new concepts to strangers to gauge engagement and understanding.

?
What is the difference between the US and Oxford/Cambridge education systems?

The US liberal arts system emphasizes learning a little about many subjects, while Oxford and Cambridge advocate learning one thing extremely well first, believing that mastery in one area teaches how to learn anything else effectively.

1. Rehearse Under Test Conditions

The key to learning any skill is to rehearse it by breaking it down into sub-skills and practicing each one under conditions that simulate the actual performance or test, especially with unfamiliar problems.

2. Master Few Things Deeply

Instead of superficially learning many different things, focus on mastering a few core skills or subjects deeply before branching out.

3. Reduce Complexity to Key Variables

To make better decisions and avoid overconfidence fueled by confirmation bias, reduce complex domains to a handful of key variables that you can effectively follow and reason with.

4. Define View-Changing Conditions

Clearly articulate what specific changes in your key variables would be necessary to alter your current view or decision, enabling clear feedback and adaptability.

5. Seek Unbiased Feedback

To get accurate feedback on your performance or understanding, practice or explain your ideas to people who don’t know you well, as friends may provide biased, overly positive responses.

6. Play to Win, Not to Lose

Approach challenges with a mindset of playing to win, rather than merely playing not to lose, even if it means taking risks that could lead to occasional losses.

7. Contribute to Society

Actively contributing to society is crucial for psychological well-being, as a lack of contribution can be devastating to a person’s sense of self.

8. Deep Learning Teaches Learning

To truly understand how to learn and master new subjects, commit to learning one thing exceptionally well, going into sufficient depth to grasp the process of mastery itself.

9. Filter Information by Key Variables

By focusing on only the key variables that govern an outcome, you can effectively filter out extraneous information, saving time and improving decision-making efficiency.

10. Learn What to Ignore

A crucial part of developing expertise in any field is learning what information is irrelevant, allowing you to focus your attention more effectively on what truly matters.

11. Master by Repeated Practice

To master a skill, deeply study the best examples (e.g., chess games) by physically replaying and analyzing them repeatedly until you know them by heart, understanding the rationale behind each action.

12. Translate Notes to Own Words

To effectively prepare for tests where you must express concepts in your own words, actively translate your verbatim class notes into your personal understanding rather than just rereading them.

13. Practice Non-Optimal Conditions

To truly prepare for real-world scenarios, rehearse skills under varying and non-optimal conditions, including incomplete information or physical constraints, as optimal conditions are rarely guaranteed.

14. Heed Repeated Feedback

While single opinions can be dismissed, if three or more people independently give you the same feedback, it’s a strong indicator that you should seriously consider and act upon it.

15. Entertain Feedback, Avoid Arguing

When receiving feedback, resist the urge to immediately argue or dismiss it; instead, entertain the idea and explore its validity, as defensiveness shuts down valuable learning opportunities.

16. Go Beyond Instructions

In life, unlike school, simply following instructions (doing exactly what’s expected) is often insufficient, as bosses or situations typically require more than robotic execution.

17. Find Your Optimal Hours

Observe and identify your personal optimal times of day or night when you function best, as this can significantly impact your performance and creativity.

18. Cultivate Single-Minded Focus

Develop single-minded focus on your goal and thoroughly understand your ‘opponent’ or domain by repeatedly studying relevant information (e.g., opponent’s games) to gain an edge.

19. Avoid Passive Rereading

Rereading notes or highlighted texts is an ineffective learning strategy because it does not rehearse the actual skills required on a test, such as recalling information or solving new problems.

20. Offer Positive, Inclusive Visions

To inspire and lead, especially in challenging times, focus on offering positive, inclusive visions for the future rather than divisive ‘us versus them’ narratives.

Sometime in the next few decades, an algorithm or a robot powered by algorithms is going to do whatever you do better, faster, more reliably in every, cheaper.

Adam Robinson

Stephen Hawking gave mankind a one in 20 shot of surviving AI, a one in 20 shot, and you can do the math. That means a 19 and 20 shot. We don't survive AI.

Adam Robinson

The scary thing with AI is this, that it will become exponentially smarter than we are in minutes.

Adam Robinson

Aristotle Onassis, who was once one of the wealthiest men in the world, said, the secret to business is knowing something that others don't.

Adam Robinson

You can play to win, or you can play not to lose. And Fisher always played to win.

Adam Robinson

The beginner mind sees many possibilities. The expert sees only a few.

Adam Robinson

If one person tells you something, that's that person's opinion. But if three people tell you the same thing, you've really got to listen.

Adam Robinson

The function of the court jester was not to amuse the King. The function of the court jester was far more serious than that. The court jester was the one person who was empowered by the King always to tell the truth.

Adam Robinson

Effective Learning Strategy

Adam Robinson
  1. Read the sample questions at the end of a chapter or section *before* reading the content to prime your brain for what's relevant.
  2. Identify the specific skills you will be required to perform on a test or in the actual application of the knowledge.
  3. Rehearse these specific skills by actively performing them (e.g., answering practice questions you've never seen before).
  4. When taking notes, translate them into your own words, rather than writing verbatim, as this is what you'll need to do on a test.
  5. Practice or rehearse under varying, non-optimal conditions to prepare for real-world challenges.

Seeking Honest Feedback for Self-Improvement

Adam Robinson
  1. Explain new ideas or concepts to people you barely know, as strangers are more likely to provide objective feedback than friends.
  2. Pay close attention if multiple people (e.g., three or more) give you the same piece of feedback, as this indicates a pattern rather than just one person's opinion.
  3. As an employer or manager, earnestly solicit feedback from your team, potentially through anonymous channels, to identify areas for improvement in your leadership and performance.
1 in 20
Chance of mankind surviving AI (Stephen Hawking's estimate) This implies a 19 in 20 chance of not surviving AI.
2820
Chess rating of best human player (Magnus Carlson) Approximate rating of the 27-year-old Norwegian world champion.
3300
Chess rating of best computer software trained on human games This software would crush Magnus Carlson in a match.
3600-3700
Estimated chess rating of Google's DeepMind Alpha Zero program This program learned by playing itself for four hours, leapfrogging all human knowledge.
72 wins, 28 draws, 0 losses
Alpha Zero's win-draw-loss record against reigning software in a 100-game match Alpha Zero beat the 3300-rated software 72-28 with 56 draws (implied from 28 draws for the opponent).
700-800 pages
Pages of unique data Tinder had on a female journalist This data was collected and owned by Tinder, not the individual, in the US.
17%
Horse handicappers' accuracy with 5 pieces of information (Paul Slovak study) Almost twice as good as chance (10% for 10 horses).
19%
Horse handicappers' confidence with 5 pieces of information (Paul Slovak study) Confidence was almost identical to accuracy in the first round.
17%
Horse handicappers' accuracy with 40 pieces of information (Paul Slovak study) Accuracy remained the same despite a significant increase in information.
31%
Horse handicappers' confidence with 40 pieces of information (Paul Slovak study) Confidence almost doubled, indicating overconfidence with more data.
15 to 30 seconds
Time Adam Robinson takes to size up global markets Applies to all major global markets (currencies, bonds, commodities) by focusing on key variables and deviations.
14
Bobby Fisher's age when he became U.S. champion He went from a 'good 12-year-old player' to U.S. champion in two years.
15
Bobby Fisher's age when he was one of the top eight players in the world Achieved this a year after becoming U.S. champion.