A conversation with Satvik Beri

On Friday, Sep. 20, I sat down with Satvik Beri to talk about his perspective on how to maximize earnings, if I want to do a lot of good by donating a lot to highly effective charities. I’ve typed up notes from that conversation below.

Satvik’s main thesis

Today’s economic landscape is marked by large and increasing volatility—industries appearing and disappearing within a couple years, rapid and accelerating technological change, etc. He expects our generation to produce multiple hundred-billionaires who gained their wealth from industries that don’t exist at the moment, analogous to Bill Gates/Microsoft creating the PC industry.

In order to thrive with this increasing volatility, generalized and transferable resources are extremely important. The main components of this are

  1. A general set of highly useful skills at the cutting edge of what’s currently possible. (In Satvik’s case, he started out narrowly with strong math and Excel skills, but was able to quickly branch out to machine learning and analytics when he saw these areas were becoming high-demand and popular.)
  2. The general ability to understand people and markets, identify real problems and find efficient solutions. Being able to understand and predict others “helps you work on more valuable problems, and working on the right problems is a multiplier on having good problem-solving skills.” (After automating away his job at a large bank, Satvik didn’t ask for more work, but instead spent a long time talking to people across the company trying to understand and solve real business problems. He ended up solving some million-dollar issues in a few weeks.)
  3. The related ability to parlay this ability into reputation by networking and selling yourself well, in order to acquire career capital and become a magnet for good opportunities. (Satvik successfully used negotiation techniques to get an 80% raise at his bank. Since he started working his salary has gone up by a factor of 4-6.)

On choosing jobs

For developing a broad set of useful skills and a strong network, it’s important to find a job with a lot of autonomy that cuts across many different concerns. Statistically, CEOs of large companies tend to have worked in multiple areas during their early career at a rate far higher than baseline. By working in multiple areas, you’re gaining more skills, which is important since you don’t know which ones will be helpful later—for example, Satvik learned machine learning, which was a huge winner, but also natural language processing and Lisp, which weren’t.

Working in different fields also gives you a much larger and more diverse network, which helps with finding new opportunities as well as connecting problems with solutions. Satvik was able to solve one tough data-duplication problem by connecting his client with someone else who happened to have written the right algorithm for the task. Plus, all of his major salary increases came from networking and selling himself—for example, the latest jump, due to a new job, was offered to him by a former coworker.

For working in multiple areas, look for companies with rotation programs, or startups. Conversely, specialized jobs like becoming a doctor or a COBOL engineer are less valuable—they may offer high earnings to start, but they have less potential to take advantage of high volatility for explosive career growth.

On choosing areas

You don’t have to predict how valuable skills will be very far in advance. Usually you can get a good impression of what’s valuable just by looking at a bunch of job postings, tracking which keywords appear with what frequency, etc. (Having a broad network and talking to people a lot helps also.)

Areas that are very new are great for young people, because no one has experience so you’re not at a disadvantage. For example, in 25 years, nobody would come to someone Satvik’s age with a data science problem, because there will be people with 25 years of data science experience to go to instead. But when Satvik started, he had as much experience as anyone, so it was much easier to get reputation and career capital in the field.

On how to work effectively

The core of his method of career growth is to get really good at understanding people and markets. Lots of very smart people make the mistake of spending 100% of their work time on their “actual job”—they finish their assigned work early because they’re competent, and then they just ask for more. What they should do instead is split more like 75% “real job” and 25% talking to coworkers, trying to understand and solve other business problems with side projects. When Satvik worked at a bank, he automated his job away and then pretty much entirely worked on side projects, and those side projects ended up providing massively more value than his work on his “real job”.

Potential issues and responses

Specialization as an alternate path

I asked whether specializing in skills that were very difficult (e.g. doing difficult math at a hedge fund) might be an alternate path to maximize earnings. Satvik pointed out that this exposes you to volatility (e.g. the risk that your hedge fund goes under or the markets become much more efficient for some reason). It also cuts off the potential for explosive growth, since explosive growth is mostly in sectors that don’t currently exist, rather than ones that have had time to specialize and become difficult.


I noted that this path seems to require many things to go well. You need to:

Satvik replied that actually, you only had to become a magnet for opportunities, and then you can make as many attempts at solving important problems as you want. He’s learned many skills that didn’t end up being relevant (like natural language processing and Lisp), and has failed to solve relevant problems many times. But he’s really good at the core skill of understanding and predicting people and markets, so he has many opportunities to do important and useful things.


What would I do if I wanted to try out Satvik’s method of career growth? I asked for two scenarios: spending a summer internship on it, and spending a full year.

Satvik suggested that if I were at an internship, since there’s not much time to understand the whole business, I should focus on something I understood well: their intern program. As someone who went through it from the other side, I would probably have a better perspective than any of the full-timers, and might be in a position to improve the efficiency of interns or something like that.

If I had a year, Satvik suggested that I try to do a startup, but instead of focusing on tech development first, spend the first 3-6 months really understanding a market (a very narrow one, if necessary), then identifying a very pressing need and solving it—manually at first (to cut down on upfront costs), then scaling it with tech.

To get fast feedback on component skills, like negotiation, I could practice negotiating on utility bills and credit card limits, or other small things. (Satvik is working on developing a course on negotiation also.)

He also recommended several books to read to figure out how market-understanding skills work (see “further reading” below).

My next steps

Immediate things I can do include reading some of the books Satvik suggested, to get a feel for how the skill of understanding markets. I can also start doing negotiation practice now with e.g. credit limits (hooray for being young with no credit history!).

I’m already planning to do a fair amount of interviewing for jobs/internships this year, so I can use this for an opportunity for negotiation practice (once I get offers), as well as “keeping my ear to the ground” to see what skills are becoming important.

Further reading

In an email, quoted below, Satvik referred me to several books for developing the core market-understanding and self-marketing skills:

This is based on my theory that the most important business skill is the ability to model1 people, organizations, markets, and economies effectively. Modeling others helps you work on more valuable problems, and working on the right problems is a multiplier on having good problem-solving skills. These are ranked in order of priority/effectiveness and I suggest reading them more or less in order:

  1. The Charisma Myth, particularly with the sections on “presence”. The tools you’ll learn from this book are extremely effective at getting people to open up and share the actual problems they’re facing, not just the low-level proposed solution. I wrote a Hacker News comment that illustrates this problem in a bit more detail: https://news.ycombinator.com/item?id=4997703
  2. Ramit Sethi’s book I Will Teach You To Be Rich has a few good scripts on negotiation (especially salary negotiation) which I used to negotiate an 80% raise at one point. It’s worth reading every script in the book, as it will help you model a lot of different business people in different situations effectively. You can skip the rest of the book unless you find it particularly interesting, the scripts are the best part.
  3. The Copywriter’s Handbook by Bob Bly also gives you a lot of useful practice modeling individuals. Ostensibly the book is about writing effectively in order to sell stuff, but even if you never intend to do this the lessons on understanding how people perceive writing are extremely valuable.
  4. The Personal MBA: Mastering the Art of Business is a fairly good introduction to modeling businesses as a set of systems, and light-weight enough that you can get through it in a few hours while still getting a lot of value.
  5. The Innovator’s Dilemma takes it three steps further. It goes into huge amounts of detail on treating organizations as systems, and introduces you to thinking of economies and markets as systems as well. Unlike most such books it’s heavily backed by data and will teach you a lot.
  6. How to Win Friends and Influence People (the old edition, not the new watered-down crap) is an excellent, slightly cynical guide to understanding and influencing others. Don’t follow the suggested actions directly-they’re slightly evil. Even so, the author clearly understands psychology and reading the book will teach you a lot about influencing others.

We’d also mentioned self-determination theory. Cal Newport’s So Good They Can’t Ignore You is a good introduction in case you want to learn more about the research behind happiness at work.

  1. To “model” someone or something is to understand and predict its behavior well. This use comes from the sense of having some mathematical model of the system. ↩︎


email me replies

format comments in markdown.

Your comment has been submitted! It should appear here within 30 minutes.
Austin Parish

Excellent conversation that I found very useful and enjoyable. I’d love to know when Satvik makes his negotiation course available.

email me replies

format comments in markdown.

Your comment has been submitted! It should appear here within 30 minutes.