In this episode, I sit down with Philip Su, a former Distinguished Engineer at Meta - one of the rarest engineering titles in the industry. Philip worked directly with Mark Zuckerberg, helped ship foundational products like Web Messenger, and later joined OpenAI in its earliest days.
🚀 JOIN MY TECH COMMUNITY: SWE LAUNCHPAD Inside, you'll learn How to break Into tech with practical resources, job prep guidance, and real industry insights you won’t find in school. → https://swelaunchpad.com
✉️ TECH NEWSLETTER Discover the latest tech trends shaping the industry, unlock career tips, and access exclusive resources to stay competitive in tech. → https://sajjaad.substack.com/
🧑💻 LEARN TO CODE Learn to code with bite-sized projects and an AI-bot to help answer your queries. I recommend you start with their python module, then check out java and javascript. → https://coddy.tech/ (It's FREE, but you can use the code SAJYT for 20% off of premium features!)
0:00 - Working with Zuckerberg & pushing first-principles thinking 0:44 – Who is Philip Su and how he became a Distinguished SWE Engineer 2:14 - How he reached one of the highest engineering levels at Meta 2:54 - When did Philip join Facebook? 3:24 - What do Distinguished SWE Engineers actually do? 4:47 - Why most engineers don’t think like owners 5:53 - How to stand out as a junior engineer? 9:00 - Compensation at the top: salary vs stock reality 11:01 - What Mark Zuckerberg was actually like to work with 12:51 - Learning directness, feedback, and hard conversations from Zuck 13:20 - What can aspiring SWE apply from Zuck? 14:52 - How Philip made hiring decisions 16:16 - What AI will realistically do to engineering jobs 17:08 - Staying in the top fraction as the bar rises 18:23 - Are computer science majors cooked? 21:18 - New roles that emerge as AI replaces code writing 22:32 - Why humans may struggle to compete with AI 23:41 - Should students still pursue computer science? 24:55 - If you were a student right now.. What would you focus on? 26:39 - Will there be jobs after the "AI Takeover" 28:13 - Where human engineers still beat AI today 29:40 - One career decision Philip would change looking back
AI Generated·May contain errors
Working with Zuck was amazing.
He was always brutally honest, like very straightforward.
Let me give you a concrete example of where he was just trying to push the boundaries.
Today, I'm interviewing Philip Su, a former distinguished engineer at Meta,
which is one of the highest levels of software engineering across the industry.
It's four whole levels above being a senior.
He got to work directly with Mark Zuckerberg and Meta CTOs
before he decided to leave and join OpenAI in its earliest days.
There, he got exclusive firsthand experience on the future of AI.
And in this video, we're going to be diving into his experience,
his insights on AI and what you as a computer science or software
engineering student need to do to navigate this new reality.
Are computer science majors cooked?
Is AI taking over?
I think in the long run, it feels like yes.
With that being said, let's get right into the video.
You were a distinguished software engineer at Meta,
which is one of the highest levels of software engineering at that company
and at most companies like there are juniors, there are seniors,
there are principals, and way above that, we got a distinguished engineer.
You were one of eight people who got that honor.
Can you tell me how you got there?
Yeah, great question.
So I had been working in industry by that point, probably 20 years.
And so it had been a long, slow journey. Right.
But I also think I joined Facebook at a relatively early time
when there were many opportunities to do big things.
And so, for instance, my first feature, Video Calling,
was written with just two other engineers,
and it was the entire feature end to end.
And so product reviews were done with Zuck directly and things like that.
And so I think because of that, there were just a lot of opportunities.
And I started in industry, for instance, I spent all my time at work,
like literally, you know, probably 18 plus hours.
Right. I had a sleeping bag in my office.
Like, I just worked all hours.
And that's definitely not something that I recommend
unless a person is really committed to that being what they want. Right.
But I do think like those long hours were a big thing.
And I think a lot of it also is just being in the right company
at the right time. Right.
You can be doing great work, but if the company is not itself
expanding into bigger opportunities, then your own opportunities
will be limited by that.
Whereas I think when I joined Facebook, it was expanding very quickly.
And so there were so many domains that you could do interesting work in.
Were there any key projects or technical artifacts
that you developed, per se,
that really pushed you to get to the distinguished level?
I think the work that really helped grow me over time, you know,
we did a rewrite of Web Messenger, which was really fascinating.
Did the video calling feature, worked on Workplace a bit
and a few other products, worked in the sharing team for Newsfeed
and stuff like that.
So there were just many opportunities.
Whereas at Microsoft, for instance, when I worked there,
a person who switches teams maybe every two to five years
at Facebook, it was very common to switch teams every six months,
you know, because products were shipping so quickly.
You probably shipped your product in six months.
And so you moved on to other things.
And at what point did you join Facebook to have all these opportunities?
When I joined the company was 500 engineers globally.
Oh, OK, OK. So you got to put in your imprint.
Yeah, because it was very small back then.
And so with 500 engineers, if you think about the people
that maybe senior or principal level, it would be realistic to know all of them.
You know, and so the work is very direct.
The level of trust between teams is very high because, you know,
the individuals. Yeah. OK, cool, cool.
And so I know as junior software engineers
and even to a certain level, senior software engineers, we write code.
We create design documents, sometimes talk to customers.
Once you start getting past that distinguished engineer
or even like certain levels around there, what is the nature of your work?
Yeah, this is a great question,
because I think it's different from some jobs.
Like, for instance, take professional basketball. Right.
If you start off as a basketball player and you eventually get drafted
into the NBA, the skills that got you there are still going to be the skills
that will make you great.
Like you just have to be great at an exceptional level. Yeah. Right.
I don't think that's true for most software developers, meaning like
the things that got you there aren't going to get you to the next level.
So when you first start, it's a lot about being able to implement things
well and correctly and quickly. Right.
But eventually everybody gets to an asymptotic point
where their personal contribution code wise can only be so much.
A lot of your impact later on will be
helping the team choose the right things to work on being a huge thing.
Convincing the team about technical vision or even about business vision
is going to be a huge thing.
So I think those skills are often not the ones that most software developers
have when they first start.
Interesting. Yeah.
So I think especially junior software engineers were like code, code, code.
Right. We're barely thinking about the full scope of the work
that we're working on.
We might not even own a feature of a product. Right.
So how does one start thinking like that?
I think a huge thing many people miss is most people don't think like owners,
meaning owners of the business, not owners of a feature
like you've got to think to yourself always, if I ran this company,
would I want myself to be working on this thing right now in this way?
Most people don't think that way.
Do you have an example of it?
Let me give you super concrete things. Right.
Like like, for instance, most people, if their workplace
lets them fly business class, they'll do all sorts of things to fly business.
Right. Yeah.
Like people at Microsoft used to do funny things like reroute the flight
to have a leg that was long enough to qualify for business.
So they did this. Right.
If you own the business, you wouldn't do that.
If it was your money, you wouldn't do that. Yeah. Right.
So then if you don't act like an owner, you'll never be made an owner. Right.
So I feel like most people don't embrace the larger question of like,
should I be working on this if I own this place?
Would I spend money this way?
Would I spend my time this way?
I think the more you can think from an overall holistic perspective
and think in a way that's aligned with what all the strategic people
in the company would agree with, the easier time you will get in growing.
I guess I have a follow up question on that is as a junior engineer,
if you even want to think like an owner, right, sometimes you're given
such little scope in terms of where you can make a contribution
such that if a senior engineer or principal engineer is like saying something,
this is how I'll be.
Why would they evaluate your opinions over theirs?
So then how would you ever be able to make that?
Yeah, that is a really deep and profound question.
So I think I will cover it on a few fronts.
First is, as you know, most tech companies still claim
they cannot hire all the right people they need.
Most companies, even despite layoffs, they will continue to tell you
they cannot hire all those people.
So by definition, the team that you're on, most likely,
if they could get a more experienced person on it, they would love that.
And then the question is, why aren't you that person?
Right. If they claim they need it.
Yeah. And yet you want that person. What's stopping it?
I think there are a few things that might be stopping it.
One is simply we don't all have the skills
we will eventually have one day in our career.
So some of it is you've got to earn that respect. Right.
And earning that respect means like Dr.
Martin Luther King Jr.
said, like, if all you can be is a bush, be the best little bush. Right.
Like if you cannot be trusted with small things,
no one's going to trust you with big things.
And so when you start on a mundane thing, like it's very standard
when you first join a team, you probably fix bugs. Right.
Yeah. Will be the best bug fixer there is. Right.
Fix them faster than anybody else. Right.
You got to earn that before people give it to you.
But the second thing about it is a lot of times we're so heads down
that we don't make the time to spend to think about the larger things
that would impact the team.
So so often we're so busy trying to meet our, you know, bug fixing criteria.
We're like, we're going to ship this product. Right.
That very few people are taking the time to, for instance,
consider what industry competitors might be doing at this moment,
what the business might need that might be bigger than the thing.
So so much of our time is spent trying to micro optimize
and be efficient in our day to day work instead of making the biggest
efficiency gain, which is choosing the right things to work on.
So I guess start in your small spheres of influence.
Do the best that you can. Yes.
And then it'll start growing.
But still always think kind of like an owner.
Yes. As if you this is your product that you're shipping out,
not Zuckerberg's product or anyone else. Yeah, totally.
Like imagine if you ran a small team of five developers. Right.
And one of them is like 17 years old.
The prodigy is joined very little sort of work experience. Right. Yeah.
But if he or she was always saying smart things,
there's no way you would ignore what they say because you know their age.
Like that is just not a realistic thing.
What is more likely is they aren't saying smart things. Right.
Or they're saying them in a way that is abrasive or hard to accept
or they're saying it at the wrong time when we're super focused.
They're talking about these big ideas. Right.
Yeah. So I think most people do not ignore
obviously smart things that are being said.
And then the question is, how can you become the person
that says those smart things?
First, you've got to get the smarts, right?
You've got to put in the hours, get the smarts.
But then you also have to know when and how to say things in a way
that is timed correctly and that is accepted by the right people
to make the impact that you want to make. OK, cool, cool.
All right. And so now for probably one of the most juiciest questions
as a distinguished software engineer at big tech company like Metta,
what is the salary compensation looking like?
Really quick, if you want to become a software engineer, listen up.
If you're not using AI to code, you're already behind.
Not because you're bad at coding, but because everyone else
is moving faster than you.
People are shipping in a weekend what used to take weeks.
Average engineers are suddenly outputting like seniors
and the bar for coding is quietly rising.
But here's the problem with AI nobody talks about.
The real problem isn't AI writing code, it's executing it reliably.
That's where Temporal shines.
Temporal gives your AI workflow a backbone.
If a step fails, it retries automatically.
If a service goes down, it resumes exactly where it left off.
If your agent crashes halfway through, Temporal picks up right where it left off.
This idea is called durable execution,
and it's what top AI companies are using to build production ready applications.
So, yes, use AI to code faster.
But if you actually want to be a great software engineer,
learn how to make AI code reliable in production.
So if you're ready to write code as a failure doesn't exist,
try out Temporal link down below in the description.
And now back to the video.
My data on that is going to be about 10 plus years old, right?
But I do think it's realistic to think somewhere between one
and five million dollars per year based a lot on performance level.
OK, cool, cool.
And what's that breakdown of like salary stock bonus?
It's overwhelmingly stock.
It's because as you get more impactful in a company, more senior,
the expectation is that you will affect the company's fortunes more directly.
If you just join a company as a college hire straight out,
nobody expects you to move the stock price.
Yeah. But if you are the lead of one of the company's large offices,
it is expected that if you perform well, that the company will do better.
And in that case, giving you stock as your primary compensation
makes a lot more sense.
Yeah. OK, cool, cool.
That's nice to know.
And it's very incentivizing.
Another perk of being at such a high level at Metta
was you got to work directly with Zuckerberg on certain things.
What did you learn from him and what were the things you worked on with him?
Yeah, great question.
And I'll start with leading with one observation, which is
you don't get to work with those people because you are that level.
You became that level because you worked with them.
Oh, if that makes sense.
Yeah. People don't listen to you because you're a certain level.
Yeah. OK.
You earn that level by being a person who says things that people listen to.
And then they eventually think, oh, this guy says smart things.
And then they automatically bump you up.
Totally. Yeah. Working with Zuck was amazing.
He is super focused. Right.
I think he also is I found him genuinely well-intentioned,
meaning like he genuinely seems to want what is best. Right.
I also found it was great that he was always brutally honest,
like very straightforward, like would just tell you when he disagreed
with something, would make some observations.
Let me give you a concrete example of where he was just trying to push
the boundaries. Right.
So I, along with maybe three or four other engineers,
were responsible for the rewrite of Web Messenger back in the day.
And back then it integrated with email, if you can believe it.
So Zuck actually asked, like, hey, why don't we go up against Gmail? Right.
And we thought, what? Gmail?
Like we have like four engineers. How is this going to work? Right.
But we discussed it for like three minutes.
And then he said, well, I wasn't really thinking that we'd do it,
but I just wanted to figure out what the what the boundaries. Yeah.
And so I do think he's willing to consider much more broadly
the solution space than I would have myself considered. OK, cool. Cool.
So he was trying to push the limits of what you could work on.
And just to think from first principles. Right.
Like, like, why is it that you can't compete with with Gmail?
Well, it turns out there were a lot of reasons.
Facebook was that small that you couldn't. Right.
But at least, you know, it's worth thinking about and articulate.
And were there are certain things that he taught you
that helped you later on in your career?
I do think I still struggle to be very direct with people.
So I think engaging in hard conversations was one thing that both he
and Sheryl Sandberg focused a lot on encouraging people to do
in the company, encouraging people to give feedback openly to people, you know.
So when you submit feedback to attach your name to the feedback,
this is something that many people feel uncomfortable with.
But it's something that I learned over time and I push myself to do.
And were there any things that you learned from him
that could be applicable to people who are trying to break into tech?
Always be building. Right.
In college, he was building before he dropped out. Right.
And so I think you give yourself more shots at goal
if you're actually doing things right.
Many people think, oh, well, I'll start building once you pay me to build.
Yeah, yeah, yeah.
I think like with the leveling, right, it goes the opposite direction.
You build first and then people pay you because they love what you build.
The best way to even get hired as a software engineer.
People encourage just build things and post about what you're building.
Yes. People will reach out to you.
Yes, totally. And the opposite thing can happen.
Like some people will not take an internship somewhere because it pays
too little. Yeah.
And what I told my son, who is doing an internship this summer,
what I told him is the best places to work.
You should be willing to pay them. Yeah.
You don't want to like because you want to gain the experience.
Absolutely.
Like if you had a chance to work with a luminary, like in any field.
Right. In fashion, in basketball, in whatever.
You should be willing to pay for that experience.
And the type of person that says, I will work for you once you pay me.
Yeah, that is a person that will always be the wage earner. Right.
Because they are thinking not like an owner.
They are thinking like a paid, a hired hand.
And I'm just curious about your time at Meta.
You were a software engineer on that path.
You also went to being a manager, director.
You influence a lot of hiring decisions. Right.
When it came to interns and software engineers,
a lot of our audience are interested in getting hired as a software
engineering intern or software engineer.
What were some hiring decisions you made that we could drive some insights
to help them? I make hiring decisions very differently.
I think from how most companies I've worked for want me to hire.
So I need to put that on there. Right.
But I buy a strongly toward figuring out things
you want that are hard to change.
So many companies, they'll do like a coding interview
and then we'll do a design interview.
And I feel like most of those things honestly can be easily taught
like for a reasonably sharp person.
I feel like the things that cannot be changed are very difficult to change.
But you actually want in a person.
One is self-starting, right.
An initiative and a drive to do things on their own.
A ability to receive feedback and to really be insightful about themselves.
That's the thing that many people struggle with.
And if you don't have that, you're not going to build that in someone.
Right. For sure.
Third is a bent toward collaborative working.
Right. Like a lot of people, especially the stronger
they are as an individual, the easier it is for them to be the lone
frontiersman on the Western front, like defending their own territory.
Right. But it becomes harder to find people that truly make groups better.
Yeah. They don't want to work with others.
Exactly. So you end up like I end up looking for those characteristics
a lot more than like, can you code this thing?
Can you design this thing?
Which I feel like for most smart people are easily teachable.
What are things that current people who want to make it
into software engineering should focus on? Yeah.
I feel it's really tough with A.I. coming on.
Right. Yeah.
The thing that I tell people a lot is, you know, A.I.
is not going to put everybody out of a job, but it might put,
you know, 30 percent of people out of a job. Right.
Well, I think over time higher than that.
But but but the the percentage is not all or none.
It's not like one day you wake up and there are no software developers.
But I do think the tide is rising quickly.
Winter is coming.
And you will survive if you are always in the top fraction of people.
Right. Yeah.
So nobody like will go to zero on all the hiring.
People tend to want to keep the best people.
Yeah, I would say even if you consistently try to be above average,
just the top 50 percent, that should keep you safe for a long time.
And what are those defining characteristics that people can work on?
Yeah, well, I would say if you even just work harder than the average person.
Right. Many people just want to work
the minimum amount that they can get paid their salary
for these slow quitting people or whatever.
Yeah, like like there are many people that think it's a it's this huge life
hack to like the undiscovered like laziness, right?
Like you're you're you're just like outsourcing all your work, right?
If you just work hard, you're already going to be way ahead of most people.
But the other thing is, don't strap yourself to things that are obviously
disappearing of what's obvious.
Like I would say my personal opinion right now is if you choose to join
a company that refuses to let you use AI as part of the day's coding,
you are joining a thing that is going to be an outmoded business soon.
Yeah. And like I right now refuse to join a company
that would not let me use the latest technology that I can use to produce code.
So that that's a good example
where you just don't want to strap yourself to to sinking ships either.
Yeah. And I guess part of that is potentially hopping companies
exploring what's latest in the trends.
Yeah. Yeah.
You have to be insightful about whether your company is one of these sinking ships.
Right. Yeah. And and if it is, you should find a way out.
I got to ask our computer science majors cooked.
Is AI taking over?
Well, I think in the long run, it feels like yes.
Now, the question is, how long?
OK, the question is, is it's a library.
You just need a lot of people.
So there are many people that will say to you things like, well,
we thought the ATM would replace bank tellers,
but now we have more bank tellers than we ever did. Right. Yeah.
So that is the majority opinion.
I think it will displace most jobs that can be done systematically.
And the thing with software is that the best software is built systematically.
Right. It's and the the semantics of coding languages are well defined.
And so it's unlike asking software to write the the great American novel.
Right. Where the semantics aren't as strongly defined.
So I do think it's a structured job.
You take a job like paralegals.
Right. Who might have the biggest thing they might do on a weekend
is read 5000 pages and summarize. Right.
This is something that AI is perfect at. Yeah.
So are we still going to have those jobs?
Almost definitely not. Right.
So I do think software falls somewhere in that bucket of like, hmm,
it will probably replace most software developers.
But then the question of how fast this is really hard to say.
Right. There are many people that say like, oh, it's easy to get to 80 percent,
but it's hard to go from 80 to 90 and 90 to 100.
And that may well be happening here as well.
And so then I have to ask you, are there more roles
that you think will emerge as a result of AI
that potentially software engineers can go into? Yeah.
So if you consider, for instance, tech meetups, right,
one of the things that annoy me most about tech meetups is a lot of them
you'll go to. It's always I'm a CEO that's looking for a CTO. Right.
This is very classic.
I have this brilliant idea. I'm looking for a CTO. Right.
That sort of person now is super empowered. Right.
They no longer need a technical person.
They can probably vibe code their first prototype
to at least have a proof of concept work.
So just the one man CEO can. Yeah.
So I think it democratizes, as you said, that as well
is the accessibility of building custom software.
Like the beauty with a computer is that it is a universal tool.
If you can express to it what you want done, it can do it.
So this is a generic power, a superpower that technical people have
that soon everybody will have that.
I think overall is a great thing.
I think software engineers that really love that work.
I think there will always be some aspect because someone's got to design
the next version. Right.
So like there's always going to be a small subset of people working on that.
But I think most software developers will need to go into domains
where their technical expertise is not at producing the code itself.
Maybe you're great at software architecture.
Maybe you have a brilliant idea for a new game mechanic
you want for your mobile game. Right.
Like those are the types of domains that I think you have to go into.
You can't just remain.
You will not sit there like competing with this thing in code generation.
Like it's hard for me to believe that that will be your specialty.
Yeah, you'll beat it.
OK, so we're going to become architects and the AI is going to be the code
generation.
I mean, eventually, won't the AI be able to architect it?
They'll identify the problems, architect the solution, code a solution.
Yeah. I mean, this is the opinion that I have that most of my friends
don't agree with me on is I feel like as the level of IQ of A.I.
goes up at some point, it will displace so many jobs
that it will become difficult for the average human to do something
that capitalism would compensate them for.
Like you, you need to add value in some way that people are willing to buy
your service. Right.
And I think as I get smarter and smarter,
you know, people are going to have to invent new ways to add that sort of value.
We used to think with the Jetsons, right, that the robots would clean our houses
while we did art and music.
Yeah. And it turned out that we are cleaning our own houses
because the Roomba still can't vacuum worth anything. Right.
But the robots are taking the the music jobs.
They're taking the art jobs.
They're doing these crazy videos. Yeah. Right.
All of the all of the stuff we thought we would have free time to go do
because Rosie was vacuuming the house or whatever is not happening.
It's going the other way. Right.
And so I do think that the tide is rising in a way that at some point,
most human jobs that require intellect,
you won't be able to beat a computer price wise or speed wise.
OK. When I will take over software engineering. Yeah.
Bill Gates is famous for saying something to the effect of we always overestimate
how quickly technology will revolutionize things in the next few years.
And we always underestimate how much it will in 10 years.
I think that will happen here is most people are calling shots like by 2026,
it will be like ASI and it's going to launch nuclear war and all this stuff.
Right. I think that honestly is probably a little bit too, too optimistic.
But I think in the 10 year time frame,
will it radically reshape what it means to be a worker in society?
Yeah, I think it will. OK. OK.
So by 2035, our world's going to look very different.
I have no idea what the right timeline is. Right.
Because a lot of other things can go nonlinear. Right.
Once the code starts modifying itself,
things can go like second order very quickly and things get weird.
So I definitely don't want to make random predictions.
But my intuition is that we are over enthused about some things that it can do.
Right. But in the long run, we're too pessimistic about how quickly
exponentials work. True, true.
And then I got to ask you if
we have a lot of computer science students, software engineering enthusiasts,
should they not pursue that degree anymore or what should they do to adapt?
Yeah. Great question. What what else should they do?
Yeah. Right. Like because right now people are saying the obvious thing,
which is like, oh, you should go into plumbing because like plumbing,
you know, that's physical and it's surfaces, blah, blah, blah, blah.
How long do you think a plumber is going to have their job?
You know what I mean?
Like if you look at where the optimist is going, where AI training can often
you now have robots that can catch like crazy balloon balls or whatever.
Right. That is obviously increasing and improving at the same pace. Right.
And so my question is, like, you can't just call this one job of software
development and say, like, oh, they're cooked.
We're going to have to go to, you know, you're going to have to become a fast
order cook. Right.
Because I have a hard time imagining what other jobs, you know, get done.
I feel like the bigger thing we need to do is make sure that our society's rules,
that legislatively, that we are anticipating this future
and that we aren't going to leave a bunch of people with an inability
to make work for themselves.
Yeah. But I mean, that's I guess one side of it.
The government needs to intercept and ensure people have food on the table.
But if you were a student right now, what would you focus on?
Yeah, it depends how good of a computer scientist I am.
Both good relative to my peers, as well as how much do I have a natural passion for it?
Yeah. Right.
I would never counsel a person to become a professional pianist,
even though a very small number of people do make it as a professional. Right.
And those people have the passion for it.
And they're willing to work 12 hours a day at piano to do it. Right.
If you're that sort of computer scientist,
you should absolutely stay in computer science because you will love it. Right.
If you're the type of person that doesn't need to be paid
and you're still writing code, you should still do it
because the robot might go faster than you, but you love writing the code.
So just do the thing. You're like an artist that paints. Yeah.
I think if you're the type of person that was a business major,
but you went into computer science because you saw all the dollar signs,
you should get out because because the person with a passion is going to out
compete you. And if you didn't have the natural propensity toward it,
you're never going to be as good as as someone who sees the tide
rising and is outrunning you. Right.
So so I think that it depends a lot on your personal commitment,
your personal passion and an objective look at how you do relative to your peers.
For sure. For sure.
I think the passion and natural curiosity will drive so much more innovation
and education. Yes. The passion of dollar signs.
Yeah. And if you take a look at America,
one of the many reasons we're a great country, right, is that in America you can be
you can host the the world's largest ball of twine,
you know, which is somewhere in middle America. Right.
And you can make a living just running a museum shop that does that. Right.
This is the beauty of America. Right.
Is like if you have a passion for something,
you could definitely make a living doing it.
You just got to be sure that you love twine more than anybody else.
You're willing to do the twine. Right.
So I think that the most important thing is be excellent at the thing you choose
to do and out execute other people in that domain.
In America, you'll always have a job based on that.
Even after the AI takeover,
I suspect we will invent things that people will value simply because a human
produced them. For instance,
you could Spotify yourself the best string quartet for your wedding,
but you could Spotify and it'll play better than any string quartet you can afford
to hire for your wedding. People are still hiring string quartets. Why?
Because they love the fact that the Peruvian woman is hand weaving this
carpet. Right. Like they, they,
they like the idea that human sweat was behind something.
I think a human labor will always be valued for the fact that it is human
labor.
Yeah. I guess in a sense, uh, if you go to like New York city, right,
obviously there are a lot of cars on the street,
but then there are some people who they have like the little horse carriages and
then they, they, they sit on them. It's very,
very like spotty here and there in the city,
but they're able to charge like a hundred dollars just to take you across the
street. Uh, because people just value that like a almost nostalgic human
experience.
Absolutely. Like, um,
if you consider young people these days buying records again, right. It's like,
why would you have a clearly inferior product that gets worse every time you
play it? Like, um,
and the answer is because people have begun to see that perfection is one type
of excellence, but this idea of the physically tangible,
this idea of the warmth of seeing the music actually spin and play is something
that has a quality all its own.
Yeah. I mean, in that regard,
so we've dabbled a little on like what types are there anything within computer
science you think the human element is still necessary for?
I think there are many computer science jobs that don't focus on raw code
output, right? If you think about staged, um,
changes in a deployed system that upgrade you from one design to another,
that's a whole skill in itself. Like, um, how do you do, uh,
sorry, that one thing, can you break that down to our audience?
The kind of, I think we've got a young audience here.
So for instance,
let's say you work at a company that has a large web service like Amazon,
right? Um, and you have code that's like 20 years old. Um,
if you have a new design for a new feature that you want to build,
oftentimes the question is how do you keep the site running while you make the
changes that are going to enable your new feature?
Yeah. Without regressive bugs.
Right. That is something that AI is right now not very good at.
Now one day it will probably be good at it, right? But right now it isn't.
And it's something that human experience can make you very good at.
It's a type of decision-making, right? Um,
you want to be the type of person who is not trying to primarily bang out code,
but the type of person who knows what code would be valuable,
knows what traps to avoid, that sort of person. Um, and as you know,
the AI is often, very often wrong, right?
And so you want to be the type of person who spots wrong things quickly. Yeah.
Right. And so those skills I think will still be very valuable.
If you could reflect back a little bit on your own career,
but now that you know everything that's going on with AI,
what's one piece of advice you would give?
I think I would have encouraged my younger self to be a little more thoughtful
about what I really wanted in my next job before I
just went to a job.
I think I often chose jobs because they sounded exciting or because it,
it was, you know, working with a friend or, or, or it paid much. Right. Um,
but I do think that over time, it feels like, you know,
the biggest optimization you can make with your life on a day-to-day basis is
probably find the right partner, sleep on the right bed. Right.
And then also probably choose a job and a profession and a company that those
are probably the biggest optimizations that you could make.
And I think many people, for instance, take the company decision.
Many people don't reflect more than a week on what company that they should
join. Right. And I think that's the highest offer. Right. Or, or,
or the only like if you, right. So,
so I do think that those are the right optimizations to really make.
Cool. Well, Philip, thank you so much for being on this video.
I really hope that you guys got a lot of value out of it. Uh,
and if you did make sure to hit the like button, subscribe,
if you haven't already, um,
do you have any links that you'd want people to click on share about?
So one thing I'm working on right now that I would love your feedback about is
I'm working on a podcast player for iOS called Superphonic.
It does things no other podcast player does. So as an example, for instance,
right now, no podcast player in the world allows you to subscribe to a topic.
You have to subscribe to a feed, right? I think that's nonsensical.
Superphonic lets you subscribe to topics. It does all sorts of stuff.
So we'd love your feedback on that. So please, uh, tell me what you think.
Yeah, for sure. We'll, we'll have the links down below in the description.
Be sure to check it out and I'll catch you all next time.