Subscribe via Email

Tuesday, January 1, 2019

A Year in Review: 2018


non-technical: suitable for any reader
Interns from Cruise Automation at Lands End Lookout
This year began with a trip down to Stanford University to present at a symposium. A month later I returned to California to begin an 8-month intern with Cruise Automation on the autonomous planning & maneuvering team in San Francisco.
While at Cruise I followed up with my research work -- traveling to Brisban Australia to present ICRA 2018 work with Carnegie Mellon University, and then Montreal to present robot magic work done a month later.

It may be apparent by now that I'm interested in developing autonomous systems. I'm also fascinated by my own agency. I spend a lot of time in my head conducting inner monologues with myself. This post is an amalgamation of such monologues spanning 2018, along with occasional photos I took at random. My apologies if this post is sporadic and a bit disorganized: welcome to my mind.

Real quick: Check out this sick photo of Alcatraz!
I'm so impressed they built a whole jail modeled after a Tony Hawk Pro Skater 4 Map!

Part 1: Defining My Reward Function

Shoutout to Ethan Min (left), for being an awesome cook.
It's great having a professional chef as a roomate! 
Transitioning from academia to an industry internship was not only exposure to a new physical environment -- I've had this experience before in  Taiwan, Germany, and Korea. This was a different transition. It required changing how I interacted with others outside of academia and solved problems in an industry setting. When required to make rapid changes to my habits and behaviours in order to keep up, I wondered: what defining characteristics of myself would remain? What are my defining characteristics?

Yeah I could define myself on a granular level by my DNA sequence; but this doesn't yield intuitive answers for the key questions I have regarding my agency: namely how I act in the world, and why.
A question of particular interest has been on my mind most of the year: What is my reward system? (i.e., what motivates me?)


Does money motivate me?
summer 2018 SOMA hackhouse crew
i like this picture out of context
Having an industry position would remove many financial burdens. Could my reward system be as simple as to be purely driven by financial gain? 

I quickly found out money didn't motivate me. Thus far in my life, rather than focusing on earning money for more expensive things, I'd find value in whatever I may reach without financial means.
 For example, I remember as a child originally wanting to be a movie director. With a weekly allowance equating to a Slurpee and some Timbits I couldn't plan on paying for prestigious acting and film classes, and so I went about making my own personal videos. This in itself can also be expensive: you need a good camera, good editing equipment, a good social network. Instead of fretting about finances, I decided to become interested in classic 1920's film: minimal sound, poor quality, black-and-white, and most importantly cheap to produce.

Having an income is convenient in that it expands my economic reach, but I don't see it as necessary for my own definition of success.

How about something more abstract...

Does recognition motivate me?
In the eyes of many computer science students, San Francisco is often seen as the lodestar of modern tech. Take a trip to a University campus and you'll find groups of students grinding through interview questions in hopes of landing a Bay Area tech internship. In Waterloo they even have a name for it: the "cali or bust mentality"-- either get a top tech job, or it's game over.
My favourite Slack intern (Alan Li), one of the bajillion UofWaterloo students in the Bay Area, yet still manages to be one of a kind.

I ended up in the Bay Area partially by accident -- after meeting recruiters from Cruise Automation at the IROS 2017 conference and interviewing because autonomous cars are hella cool. Still, I wondered if I possessed a strong subconscious desire to be recognized as successful in my life, and thus upon reaching some definition of success, I would become complacent.
This is partially true. I value being recognized for achievements and reaching my goals. Unfortunately, the person whom I seek recognition from the most -- and the person who is most demanding of me -- is none other than myself. Thus, recognition doesn't motivate me in an extrinsic sense. My energy comes from a deeply ridden fear of letting myself down.

This leads to life often barren of comfort.

Perhaps I am rewarded by comfort?

The company provides many resources that help employees (vacation time, snacks, lunch, dinner).
Why keep working so much if I am able to settle down and just live comfortably?
    This question makes me feel sick to my stomach. I don't see a point of existing if I'm not constantly     striving to reach my full potential.
One of my roomates as the president for Halloween.
Always striving for your full potential!

There were a lot of people in the Bay Area whom I looked up to, all suggesting I sleep more and try to relax as it is healthier for me. My manager mentioned sleeping more would improve my memory retention and productivity. My roommates pointed out that I'm frequently yelling while I sleep (sorry guys) and that stress may be a factor. I was taken aback by such suggestions to go easy on myself -- after all -- I didn't get this far in my life by relaxing.
Regardless, I subscribe to the notion that if I have strong beliefs, these are the beliefs I should question the most. I heeded the advice from my peers and tried to instill work-life balance and to relax.


This stint involved forcing myself to get 8 hours of sleep per night (instead of my usual 4-6), and to allocate time for leisure activities (watching movies) because this is the kind of thing I've heard people do to relax.

During the ~5 weeks I maintained this routine I felt more rested, and had less tension on my mind -- in other words, I felt relaxed.
It felt awful.
It felt as if I had introduced a regression into my behaviour.

I learned a core quality of myself: I don't psychologically let my guard down and relax -- I'm not a chill dude.

This exercise helped me realize a significant problem though: I'd rather collapse under stress than admit I need rest. Even worse, I was in a stalemate with myself as to how to resolve this dilemma.  I was consciously aware of the benefits of sleeping more and relaxing the mind, but likewise, I was also aware of what felt best for me -- the act of always being on alert, steadily progressing forward like a military unit deployed in battle.

Clearly, something had to evolve, as it simply is not rational to maintain a mindset after becoming consciously aware that a more effective and healthy mindset exists. After multiple months of careful iterative reflection, I found a solution I like to call "purposeful leisure".

In my mind, purposeful leisure is the concept of taking relaxation and making it a conscious task in order to improve overall productivity. This means instead of relaxing because you are tired or not feeling up to something, you instead schedule times (or set other criteria) for when to relax and then put more focus into doing so at those times. Thus, leisure became part of my routine, without me feeling like I'm not giving my all at any point in time. In this new framework, failing to relax is the same as giving up on a task: I don't give up.
Piecing together a reward function
At the GooglePlex (Mountain View CA)
Danielle (right, red shirt, walking),
you almost escaped showing up in my blog photos
(keyword: almost).
I have reaffirmed the notion that reward doesn't come to me through monetary means, nor explicit recognition from others. I achieve great satisfaction knowing that at any giving moment, I am giving my absolute all at what I do, even if this puts me far out of my comfort zone. In fact, I have realized that I thrived particularly in scenarios where things get intense and require perseverance.
I also learned an aspect I didn't realize about myself. I seem to thrive in an environment where I possess ownership and control of my surroundings. This isn't ownership in the physical sense; but rather psychological ownership -- the notion that the success of a task or project is directly tied to my own value as a person. In other words, when I psychologically own something, I take on full internal responsibility for the failure and setbacks of that something. When I fully believe that I am an irreplaceable and valuable component of a team I'm on, I operate at my best. The question is how to ensure wherever I go, I may establish this psychological ownership.  

Part 2: Ways to Optimize Life
After multiple years of optimizing functions, designs, code, and just systems in general, I've become very focused on how to apply similar concepts to myself. 
Pre-2018 I had a few ways to get more out of what I do. I would ask myself primarily the following...


How can I get the most out of events in my life?

I vibe well with this quote
"Life is not measured by the number of breaths we take, but by the moments that take our breath away."~Unknown

For example, if on a given Thursday I plan to do both a 5km jog and plan on having chicken wings with friends, I will run in the morning so I'm more hungry and thus can eat more chicken wings in the evening and gain more utility then. This way I've both got a workout in and can eat more in the evening. 
  
How can I train myself to work for longer?
Pre-2018 I had plenty of ways to get myself working longer. For example, I would get up at 5am and begin my studies at 7am regardless of circumstances. Improving diligence, discipline, and establishing habits can be powerful; but it has limitations -- you only have so much time. I hadn't realized the fact that just putting in more hours doesn't mean they are more efficient. There is ultimately a trade-off point where your maximum output is reached and now doing more hours just spreads out your work over that time -- or even worse, you get even less done.

Following 2018 there are some additional questions that I've been asking myself


How can I do more by doing less?

This question isn't asking how to not work as hard. The answer to that question is simple: just don't do anything in life. I know that option always exists, and so I'm more interested in how to get more done in the future by eliminating less impacting tasks in the present.

In 2017 I had a way of rigidly calculating my day and then working on whatever tasks I had outlined for myself to do, no matter what. This is fine if you have discipline, but sometimes priorities change mid-day, and you have to learn to do whatever is most important at that time. The thing is if you sign up for lots of favours and bloat-work (especially as a student) you may quickly find all of your high priority tasks aren't really that impacting. For example, I may have signed up to help tutor someone in a course, and then afterward volunteer at some event, meanwhile I have an exam the next day I really need to study for. Volunteer work is not objectively bad, but I had the tendency to sign myself up for too much. Since I'm the type of person who won't let others down, these obligations become a top priority and thus obscure my ability to have maximal impact in what I actually need to do (such as studying for an exam). I've been trying to get better at eliminating unimportant favours, debts, and other tasks that pile up fast. This enables me to focus on key impact things. 

My family visiting me in California (taken on the Stanford Campus)! I know it is cliche to say "I wouldn't be here if it weren't for you"; but, seriously. I have trouble managing my own life, yet my mom has managed herself and the three of us.
That's impressive. Thank you all!

What is the best way to use my time, right now?
At any given point in time, there are many things you can do, and it's important to figure out which one you can do with the minimum effort given your current state.
For example, on a given day I may have to read a book for class and finish math homework. Knowing I do math better in the morning, I focus on that in the morning and then focus on the reading later. 

The point is, don't just treat work as work. The amount of time a task takes is largely dependent on:
your energy: How well can you focus on it right now? I don't mean how much do you "feel like it" -- that just requires discipline -- but sometimes you clearly aren't in the right head-space to do something and this may cause a 1-hour task (such as writing an essay) to take 4 hours. 
location: A task such as buying shoes may only take a few minutes if you're near a store, but could take an hour if you are at home and require yourself to walk there first.
available resources: If you have a task involving editing a video, this likely requires your laptop or desktop nearby, and so it isn't even possible to do while you're on a bus. 

As an example, this blog post (part two) was written in one hour during my flight home from San Francisco, typed using my phone. I had scheduled to work on it in the morning; since I'm often best at writing then, but alas at 3pm, the time I'm usually most tired, I got hit with a rush of creativity and I just started summarising my thoughts in a notepad. This definitely took less hours to do than if I tried to force it on myself in the morning. Still, I've gotten all my desired tasks done today. 
At the SF Zoo. "Let's go to the Africa exhibit last. I can see Giraffes in my backyard."
~Nyabuti (roomate, friend, skater, data scientist, probably like 50 other things I don't know about)















The ultimate optimization: Automation

In my opinion, this is the ultimate perfection, the epitome of efficiency is completion but under pure nonhuman delegation.
Prior to 2018 my brain operated on the premise that if I was capable of making something fully automated, I would drop everything I'm doing and automate it. I had been rather obsessed with automating everything I could in my life -- from robots in the AAlab to short startup scripts on my computer -- but until 2018 I had never considered this diagram.


What I didn't take into account was the amount of time spent automating vs the actual time gained from automating such a task. 
For example, I spent about 3 hours during my internship writing a script to automatically open all of my workstation tools so I could begin working. This saved me about 5 minutes of manual labour each time; however since I only had to do this about 20 times during my internship, I've effectively spent 3 hours of work to reduce 20*5 = 100 minutes of labour. This means 80 minutes of time was wasted. 

The point is, look at how much time you'll actually save by automating something before you automate it. More importantly, look at how much time the world will save by you automating something. 

A neat example
Consider the task of walking to your local movie store, picking up a movie, and walking home to watch it let's exclude any time deliberating on which movie to watch).

To be extremely conservative let's assume the following
Time to complete this task: ~30 minutes (walking to/from the store including time at a checkout waiting in line)
People doing this task: ~10 million (assume only 10 million people rent movies)
look look! a robot that can make tea!!! 
Frequency: assume a person rents 2 separate times per year

In one given year, this would mean approximately 30 minutes * 10 million people * 2 movies per year = 600 million minutes are spent yearly on this travel time.

That's over 1141 human years worth of time

If you can build a system that automates this short task of picking up a movie to watch, you've indirectly saved literally hundreds of millions of minutes of time yearly even across just a small population. 
That's why traditional movie vendors like Blockbusters were replaced by companies like Netflix. 


Conclusion for 2018
Visited Carnegie Mellon in November!
RISS alum (2017/2018)
This post seemed like a decoupled jumble between what motivates me and me ranting about automation. This mimics the disparate nature of these topics as I pondered them throughout 2018, until I've realized that they encompass the same thing. 

My passion for artificial intelligence and automation embodies my reward system. In order to automate something, I must take full psychological ownership of the task and the challenges to fully understand its inner workings. By automating a task I'm able to effectively solve the problem and to create value for society. This is what motivates me.

Putting it all together: My core mission
My goal is to build a future where all things that are unsafe, unfulfilling and arduous can -- at the user's discretion -- be handled by an autonomous system. I want to empower humans to find what gives them purpose, and live fulfilling lives in a world demanding their intellect and passion rather than inhumane labour.

This is a complex goal to reach, and it will take my lifetime; however nearly every waking moment in my day is focused on this goal in some respect.
As of 2018 I've been able to more formally summarise what motivates me, and why, thus helping me better understand what I need to do in the future. 

Instead of saying what I will do, I'm going to just do it and reflect later on how it pans out.
Reach out to me if you'd like to help.

Onward to the future! Onward to 2019!

Two of my closest friends from home (Emmett (middle), Keith (right)), waiting for me at the airport!

No comments:

Post a Comment

Please feel free to give feedback/criticism, or other suggestions! (be gentle)

Subscribe to Updates via Email