List of Advice

Over the years, I’ve gotten a lot of advice from friends, colleagues and reading that stuck with me. I started writing some of it down — sharing here in case it’s useful to others. I’ll keep adding to it as I go.

"Patience."
1. This single word is the parting advice I got from a manager earlier in my career whom I really respected. I still haven't developed a lot of patience (ppl who work/worked with me can probably testify). I'm not proud of it.
2. Shortest path is not the fastest, see "proof" here.
"Life is unfair."
1. If you feel you didn't get what you want, did you ask for it?
"It is always my fault."
1. "Other than health and a few specific issues in your life, if things go wrong it is always your fault". If this is shocking to you (at least it was shocking to me initially), take a minute to examine a few examples where things went wrong and you feel like it is other ppl's fault. Then think really hard if there really is nothing, like absolutely nothing, you could have done to influence the outcome? I bet not.
2. Thinking this way make you have agency in making changes. "Be the change you seek".
3. Never have the mindset of a "victim" - it is an active choice.
"Take actions on surprises."
1. Test result is most useful when it surprised you, that is the only time you should NOT ignore it.
2. You probe a system you get surprised, that is how you learn, and build better mental model and intuition so in the future your guess will be more correct; intuition saves you time and makes you efficient, good ideas are often motivated/initiated by intuitions.
Relate to the above, if someone's behavior surprised you, take notice, it might be the time to ask for feedback.
"What did you do about it."
1. Stop complaining, take actions or accept things the way it is.
"Don't help them."
1. Give people space to fail and figure it out.
"Be honest."
1. Sooner or later you have to be.
"The world is not that malleable, you are."
1. Probably not completely true, but a fresh perspective when you feel stuck - what are some assumptions, behavior on your end that could be changed to make the situation better?
"Nothing lasts and it doesn't matter" is how homeostasis works.
1. All your joy and anguish will come and go. If you feel stuck, just think about this a bit. It's biology.
"Life is not a dress rehearsal—this is probably it. Make it count. Time is extremely limited and goes by fast. Do what makes you happy and fulfilled—few people get remembered hundreds of years after they die anyway. Don’t do stuff that doesn’t make you happy (this happens most often when other people want you to do something). Don’t spend time trying to maintain relationships with people you don’t like, and cut negative people out of your life. Negativity is really bad. Don’t let yourself make excuses for not doing the things you want to do." - Sam Altman
"The opposite of freewill is not determinism, it's compulsion." - Josha Bach
1. The quote was originally about addiction, but it's a good reminder for cases where compulsive behavior takes away our time and life.
2. Many people, including myself, have some form of OCD that manifests as having to do certain things or do them in certain ways. Hope this quote serves as a good check when we go too far - are we acting from choice or compulsion?
"Intuitions."
1. "The universe is so hard to understand because there is nothing to compare it to" - some geophysics Professor in UC Irvine I think.
2. To a large part, we learn by analogy. We compare a new thing to something we already know and understand well. Understanding means we have a mental model of how it works. That is called intuition.
3. Some of the strongest eng and product leaders have great intuitions, they seem to guess things right a lot without doing much detailed work
"Be curious."
1. Curiosity is the only lasting motivation.
A few project management things:
1. Front-load risk mitigations, focus on the unknown/reducing variance
  1. Dealing with the unknown/variance is what project management is all about
  2. Bring risk mitigation related work items to the early phase of a project, i.e. front loading
  3. Acknowledge and be explicit about risks we are taking on for new project, known and unknown ones
  4. Have a plan to mitigate them (or at least work towards learning how to mitigate)
2. Aggressively shorten time to the first test (you learn a lot and can adjust your plan)
3. Explicitly manage dependencies
  1. A lot of the time, dependencies are in the subtext or as context outside of the main discussions, the main discussions are often about the work your team is doing, but for the project to be successful, on time, your team depends on other teams, how do you hold them accountable?
  2. Dependencies especially items on the critical path, need to be managed as if they are your own work items
4. Learn about basic project management practices, what is a "critical path", scrum vs waterfall what's the pitfall of each?
5. It's important to have a central person holding a lot of context in their head and ensure the conceptual integrity of the overall project plan. Having those context spread over 10 people, nothing gets done.
"When stuck, take a break, and try the opposite of your default mode."
"Real respects real." The first order of business is doing good work, all else are secondary.
"People closest to the work should take the strongest positions."
"Don't shout from the back of the bus."
1. Have real accountability / skin in the game.
Don't work too hard, don't ignore your dreams - sometimes we are very emotionally invested in our job, you forget your own life priorities, and you forget people you work with are also a human being
1. This is from one of Paul Graham's essay
2. I think there is an "operating distance" to work, too close you lose sight/perspective of bigger picture (e.g. bigger problems to solve, maybe someone who frustrates you in a project is a great human being and someone you generally want to be friends with and learn from etc), too far you probably aren't that engaged to be productive and are doing your team a disservice.
3. So don't work too hard, don't be too close (or too far) from your work
When you say something (good or bad) about someone tell it to them directly, and be genuine about it.
1. Good == you should tell your colleague they did well, those positive feedback loops add up
2. Bad == only by telling someone something is bad, things can be improved/changed
"Pick up the garbage on the floor."
1. A team/organization is a group of people held together by a few amorphous things: common practice, processes, eng culture, trust and friendship; all of these will decay if we don't upkeep them, and your team falls apart.
2. At one point, I threw around a number: "eng process will decay by 7% per month if no one maintains it". So for every process/eng culture that is working well, there are a group of ppl behind the scene actively maintaining, improving and enforcing it.
3. A message I shared to EM and TLs on one of the team I supported: a lot of things we do to improve team and reduce entropy will seem unscalable. But they will work, there is no silver bullet, just lots of lead bullets. "Continuous improvement is a shared responsibility. Action is an independent one.", so do your part and inspire other engineers to do theirs.
"Deliver results - focus on outcome, then work backwards."
1. I still think Amazon is the best company on this planet... jkjk
2. But I generally enjoyed my time there and really liked a couple of Amazon's leadership principles, and they are highly correlated: "obsession about customers" isn't really about customers per se, it is about the company caring about winning, winning over customers is the outcome, we are always obsessed about it and then we say "we work backwards", we should be crystal clear of what we want to achieve as an outcome then propose solutions to get there while always checking
"Planning is useful; plans are not that useful."
1. Planning makes everyone on the same page about the tradeoffs and constraints of the system and be prepared to make tradeoffs / adjust as new info comes in
2. Always try to shorten the time to the first test of your hypothesis so you can adjust your plan
"When in doubt, optimize for the global."
1. Optimizing for local and having too much selfish interests will generate solutions/situations that can't really last
Work begets more work.
1. Once you made a mildly wrong decision to start some unimportant work, there will be more unimportant work started around the initial set of unimportant work. Very quickly there will be jobs created around those work and this is how bureaucracy is created.
2. Set the bar high for starting a project - "Hell yeah or No". And, regularly put ongoing projects on trial to see if some need to be shut down.
"Have backbone, and think on your own."
"Refining craft in itself is motivating."
1. Motivations wax and wane,
2. The act of trying to be the best in class, motivates you more consistently - that's why I always liked asking eng teams to publish papers, to show that they are doing something world-class;
3. That sense of superiority in being the best is contagious to other aspects of our work; by being the best, you would not want to disappoint the world and would want to have rigor, and uphold the highest standard
4. Once you have that kind of standard, it is so much easier for everyone to give and take feedback, mediocre standard is prone to be debated, world class standard doesn't need to be debated, I/we/you are most likely not there, period, improve.
"Focus on learning early in your career."
1. "Mortgage your youth to learn"; later, learning gets expensive (opportunity cost and the actual cost of making mistakes)
"Assume the responsibility, do the work first, 'title' comes later."
1. I hate to generalize, but myself and ppl from my own culture seem to not feel empowered to do stuff until there is a title change; we even have a saying to this 名不正言不顺 (if you don't have the title, you can't get the message thru). There must be some truth to it.
2. The opposite of this thinking is just assume it is all your fault and start improving things. "Ask for forgiveness not permissions"
"Being worried about someone doing something, despite good intentions could give that someone an impression you don't trust them."
"Strong opinions, soft delivery, private venue."
1. Slow down your delivery
2. Turn frustration into structure
3. Criticize privately, and in person / vc, not async
"Buy time."
1. Create space for you, others, team to have perspective, gather data so we can make informed decisions
"Take moving averages."
1. Mentally or physically have a record of observations over time, taking a moving average is the way to improve precision of assessment and feedback, while avoiding whiplash
"Don't share negative feedback alone, always offer a balanced view (and have perspective)."
"Make mistakes 20% of the time, don't ping me, you'll get better on your own." is what you should tell your team/reports.
1. It's probably better expressed in this quote "Showing people that you trust them to do well is very powerful and encouraging. And part of that is accepting imperfection where it really doesn’t matter."
"Taking no action sometimes is the best action."
1. This is like empty space sometimes creates the strongest compositions in photos and paintings, doing nothing is sometimes the highest form of action you can take, but it goes against your poor intuition
(More to come, hopefully...)