Category Archives: Learning

BIll Gates’ Advice to His 19-Year Old Self

Source: CNBC, Feb 2017
<reddit source>

If the 61-year-old Microsoft co-founder and billionaire Bill Gates could whisper in the ear of his 19-year-old-self, he would have a few things to say about what he’s learned.

Intelligence is multifaceted

“I would explain that smartness is not single-dimensional and not quite as important as I thought it was back then,” Gates says.
While he now realizes that IQ may be overrated, Gates has often emphasized the importance of being curious. An interest in the world can and should be fostered, he says.
“I think having parents and teachers reinforce your curiosity and explain what they are fascinated with makes a big difference,” Gates says.
He encourages people to remain committed to learning throughout their lives. “A lot of people lose their curiosity as they get older, which is a shame,” he says. “One thing that helps nowadays is that if you get confused about something it is easier than ever to find an article or video to make things clear.”


Free (Online) Deep Learning Textbook

Source: DeepLearning book website, 2016

Lectures and Slides Available 

5 Increasingly Dominant Tech Companies

Source: Zero Hedge, Apr 2017

  1. The most valuable companies in the U.S. are increasingly tech companies.
  2. The concentration of value (as denoted by market cap) being created in tech, is increasingly being created by the five largest companies, Facebook, Apple, Microsoft, Google and Amazon (aka FAMGA)
  3. The concentration of market cap in just five hands will have an increasingly profound impact on innovation and wealth concentration in the U.S..

Solomonoff Induction

Source: Less Wrong, Jul 2012


1. Algorithms — We’re looking for an algorithm to determine truth.

2. Induction — By “determine truth”, we mean induction.

3. Occam’s Razor — How we judge between many inductive hypotheses.

4. Probability — Probability is what we usually use in induction.

5. The Problem of Priors — Probabilities change with evidence, but where do they start?

The Solution

6. Binary Sequences — Everything can be encoded as binary.

7. All Algorithms — Hypotheses are algorithms. Turing machines describe these.

8. Solomonoff’s Lightsaber — Putting it all together.

9. Formalized Science — From intuition to precision.

10. Approximations — Ongoing work towards practicality.

11. Unresolved Details — Problems, philosophical and mathematical.

Be More Convincing

Source: Business Insider, Dec 2016

if you want to convince someone that your explanation for something is the best way to explain it, you might want to tack on some useless (though accurate) information from a tangentially related scientific field.

It turns out that when you tack on additional information from a respected field of study, people think that makes an explanation more credible.

one of several cognitive biases we have in favor of certain types of explanations. We think longer explanations are better than short ones and we prefer explanations that point to a goal or a reason for things happening, even if these things don’t actually help us understand a phenomenon.

As the authors behind this most recent paper note, previous research has also shown that we prefer explanations of psychology when they contain “logically irrelevant neuroscience information,” something known as the “seductive lure effect.”


  • Good explanations matter, and were rated better than bad explanations (even if the bad explanations had reductive information).
  • Adding useless reductive information made the biggest difference when researchers added neuroscience to an explanation of psychological science.
  • Participants trusted psychology the least and — in the one exception to the general rule — didn’t think adding psychological explanations to social science made those explanations more credible (though these particular findings weren’t statistically significant).
  • Study participants actually considered neuroscience more rigorous and prestigious than the sciences considered more fundamental by researchers (biology, chemistry, and physics). This could explain the big effect that neuroscience explanation has when added to explanations of psychological science.
  • Mechanical Turk respondents thought the explanations with reductive information were better than undergraduates thought they were. That information made a big significant difference for them, but it was less of a big deal for undergraduates. Different groups of people are going to evaluate information in different ways, and neither of these groups of people can accurately represent the way the entire population evaluates information.
  • People who were better at logical reasoning were better at evaluating explanation accurately (they gave less credence to reductive information). The researchers think this could mean that philosophers who have studied logic are less susceptible to this cognitive bias.
  • People who knew more about science were also better at telling good explanations from bad explanations.


Usefulness of Useless Knowledge

Source: IAS, 2017

In his classic essay The Usefulness of Useless Knowledge, Abraham Flexner, the founding Director of the Institute for Advanced Study in Princeton and the man who helped bring Albert Einstein to the United States, describes a great paradox of scientific research. The search for answers to deep questions, motivated solely by curiosity and without concern for applications, often leads not only to the greatest scientific discoveries but also to the most revolutionary technological breakthroughs. In short, no quantum mechanics, no computer chips.

This brief book includes Flexner’s timeless 1939 essay alongside a new companion essay by Robbert Dijkgraaf, the Institute’s current Director, in which he shows that Flexner’s defense of the value of “the unobstructed pursuit of useless knowledge” may be even more relevant today than it was in the early twentieth century. Dijkgraaf describes how basic research has led to major transformations in the past century and explains why it is an essential precondition of innovation and the first step in social and cultural change. He makes the case that society can achieve deeper understanding and practical progress today and tomorrow only by truly valuing and substantially funding the curiosity-driven “pursuit of useless knowledge” in both the sciences and the humanities.

Einstein, Godel and Von Neumann were at IAS

Improving your E-Mails

Source: Fast Company, Feb 2017

Before you dash off a hasty email and risk offending or annoying the receiver, check out these common but unpopular lines and opt for an alternative instead.


say: “Thank you for being patient with me.”

2. Instead of “WHATEVER YOU THINK”

say: “I’m open to your ideas and am happy to do some more brainstorming.”

3. Instead of “PLEASE ADVISE”

say: “Let me know if you have any thoughts on how to proceed with this.”


say “I’m interested in your feedback on this update.”