a16z Podcast: Automation + Work, Human + Machine

Measurement will be one of the core habits of a machine learning/AI-centric company. What will be the other core habits?

  1. Build the tools & hide the AI
    • Building the scaffolding of tools to be able to test the quality and reliability of the data. Computer development has robust tools for testing, data now needs the same.
    • Hide the AI – e.g. spell check. 
  2. Find the right people
    • People with a deeper insight into Deep Learning and AI.
  3. Focus on the data
    • Start with measurement question and then maniacally focus on data and data quality.

Human Potential and Talent

skip to 34:40

Are we working too hard to measure and spot talent – branding and marking people too early?

Gladwell: Success is correlated with having a ‘broad, early base’. Specializing too soon can be problematic.

There are a body of skills you learn only if you have a broad, early base. For example, many professional athletes, in their childhood, excelled in sports other than their vocation. The other sports likely had a strong influence on their ‘natural talent’ in the vocational sport.

The best preparation for one domain is often a completely different domain.

We do a bad job of spotting early talent. Time and accuracy have an inverse relationship. Formative growth happens early. The earlier you try to spot talent, the lower your odds of success. Additionally, in the nurturing/development of that early talent, you may be damaging the overall prospect with specialization and burnout. People bloom at different ages. Runners who excel at an early age very often do not remain a top talent into their prime years.

Overly competitive environments, at an early age, are very problematic. Example: Prestigious schools. Putting a very smart child – one who is in the 99 percentile – alongside classmates in the 99.9 percentile often can discourage and demotivate the child.

Takeaway: Delay specialization as long as possible. Burnout is real and early prediction is inaccurate.

Read Classics

Read some classics. You probably read the equivalent of a few books a month in tweets, hot takes, and other low-calorie material. What if you shifted some of that consumption to not just what’s most recent, but the best humankind had to offer over the last few millennia? The influences compound and tickle your creativity in profound ways.David Heinemeier Hansson


Seth Godin on Education and Learning

Start at 9:09

Mostly, organized public education was designed and paid for to create generations of kids who can sit still, pay attention, do what they were told, and who would be ready for factory work. So the education system is all about processing human beings like a factory. If they’re defective, hold them back and process them again. Everything meeting spec. Standardized tests. Test and measure.

If you want to teach someone to be a baseball fan, you don’t give them a baseball textbook and tell them there’s going to be a test. And if they do well on that test, there’ll be another test. And if they do well on four years worth of tests, maybe they can go to a game. This is not how baseball fans are created. We don’t have to do anything to get baseball fans to measure batting statistics – or get involved with metrics. They do it because they want to.

And if we find someone who is deep into American History, they are not deep into American History because there is a test, they’re deep into it because they like the status that comes from mastery. They like satisfying their curiosity. They are on a quest. They’re exploring.

So our future. Our future where Google knows everything. Why in that future would we ever teach people to memorize stuff? What is the value of memorizing anything? If its worth memorizing, it’s worth looking up.

What we need to teach kids, from an early age, is resiliency. And curiosity, and energy, and positivity, and connection, and respect, and dignity.

Exactly the way we teach a kid how to walk. Every single kid who is physically able to walk, who is 10 years old, can walk. So what happened? Along the way, at some point they didn’t know how to walk and then they did. Not from a text book, not from a video, but from standing up and falling down. And standing up and falling down. That approach was worth it because they wanted to be able to walk.

WOD 7/31/18

A: Shoulder Press (5-3-2-2-2-2)
Every 2 minutes, for 12 minutes (6 sets):
Strict Press
*Set 1 – 5 reps @ 70% – 135#
*Set 2 – 3 reps @ 75% – 145#
*Sets 3-4 – 2 reps @ 80-83% – 155#
*Sets 5-6 – 2 reps @ 85-88% – 165#

B: Metcon (AMRAP – Rounds and Reps)
Partners alternate whole rounds to complete as many rounds and reps as possible in 16 minutes of:
3 Dumbbell Man-Makers (55/35 lbs)
6 Strict Pull-Ups
9 Burpees
Manmaker = Row(R)–>Push Up–>Row(L)–>Jump to Stand–>Power Clean–>Push Press

Completed 13 rounds @ 50# w/partner Jason