In light of graduation season, this week’s material has been focused on a mix of inspiration and general course of action to what I would characterize as an overarching sense of thriving.
When I’m not yapping about trading or options, much of my writing or curation revolves around learning, motivation, and creativity. These 3 factors are in a continuous conversation with each other. That internal conversation has far-reaching effects because it guides our actions which in turn feed back to this internal conversation. We’re these complicated black boxes hosting billions of hormonal and electrical collisions that mean on some level we’re all the same in that we can be described generally as such a machine, but profoundly different since the infinite combinations of those reactions make us unique as, well, precious snowflakes.
Any perspective that fails to appreciate both how general and specific we are is an incomplete description of our condition. Which brings us to the fundamental tension of advice. It’s tempting to give it because on one level we’re not that different, but on another level it’s impossible for any general advice to be right-sized for any individual.
It is with this uneasy marriage of humility and hubris that I dare to offer a few thoughts for those pulling up to the staging areas of their careers. It is the same answer I give to those insistent enough to figuratively sign a waiver acknowledging all the disclaimers that an honest person would give before proffering advice.
The magical words that indicate such an insistence, an insistence which places earnestness above omniscience, which may be the best we can do in wicked domains, are “what would you tell your kids to do?”
In the face of such an approach, I’m cornered. But since I know that you know my kids are not your kids, I can trust you will be able to adapt the advice to your own or your loved ones’ situation.
What should a new grad pursue?
The number one thing a new grad should optimize for is rapid learning.
Womp, womp.
Kris, you put that behind a paywall?
Hang on.
It’s quite clear to me that this is either not as obvious as it appears OR people don’t even know what rapid learning means.
Let’s start with what it means to learn.
Actually, let’s start with what it is not.
Learning is not trivia. It’s not most of what you ever did in school. To be as charitable as possible to the institution of school, it is a survey of the pu pu platter of subjects that have achieved a depth of scholarly history for which the degree of specialization required to contribute to our collective understanding offers an ample capacity to absorb a motivated student’s talent. There is no end to biology. If it has grabbed your imagination, there is a lifelong pursuit awaiting you.
However, most people will not become academics, scholars, or scientists. Most will not chase anything formal school is instrumental for, or for which access to leading professors is important. The learning you get from school beyond the years of arithmetic and reading acquisition is mostly trivia.
[This is not a knock on trivia or, by implication, the subject of history. Thinking needs raw material to operate on. There are extreme points of view that don’t see the benefit of learning anything, since you can look stuff up. If I had to steelman that position, I’d argue that basic knowledge should be reclassified as something to acquire on demand since there’s an opportunity cost of onboarding knowledge you’ll never actually need.
But I think the extreme view is a bridge too far because there’s an irreducible arbitrariness in what will turn out to be useful knowledge. The project known as humanity seems to be an instance of a cosmic explore/exploit problem. I think there’s plenty of room to update curricula to better balance pragmatism and imagination, but as we get to the steep part of any learning ROI curve, the risk of premature optimization would overwhelm the gain in local efficiency. But insofar as I think curricula can use an OS upgrade, we’re pretty far from that point.]
So for most people, school mostly teaches trivia that they will mostly forget. The forgetting part is neither the student or the school’s fault. It’s actually a symptom of the uselessness. Knowledge is use it or lose it. If you’re losing it, it’s because you’re not using it. What odds are you willing to lay on your ability to do long division with 3-digit numbers? Can you multiply the year the Magna Carta by the atomic weight of oxygen? That “Are you smarter than a 5th grader?” was a fun show that proves the point. You might not know what a 5th grader knows, but you can still multiply 2 numbers. Use it or lose it.
Great, we’ve established that most of what we were told is learning, is actually not learning, at least according to any definition of learning worth having. Ok, here’s a definition worth having:
The assimilation of information that leads to adaptive changes in behavior
It’s only a few words, but it orients “learning” towards identifiable goals.
To go all 5th grader on you, subject:
“Assimilation of information” is not just about onboarding but also about accessible retention. You need to be able to draw the arrow from your quiver when appropriate.
Predicate:
“Leads to adaptive changes in behavior” implies that the knowledge is applied towards growth which is something that happens in the physical world, even if it starts internally. Growth is a goal. It can be as fuzzy as imagination or brain fodder, but these remain intermediate steps to change.
That change should be in the direction of desire.
This deserves a couple of comments:
- Getting better at crime is a desire and anti-desire depending on who you ask, so the notion of growth is neutral from a non-moral master view.
- The recursive nature of learning is also on display. Our desires are inputs that dictate where we spend time to learn, but also outputs that can change based on learning.
Learning is ultimately about outcomes. The concept, being morally neutral, is compatible with any metaphysical view. But being as we are literal material composed of nature’s Legos, our bodies and minds have maintenance requirements throughout our lives. Those requirements should include a surplus capacity for growth in the early years of our lives and a surplus to slow the decay in our back nine.
Rapid learning is the abstract foundation of this surplus. You can modify the terms based on context. Make it more concrete if you prefer. Savings. Redundancy. Time. All of which return to the abstract. To a benevolence at a human scale. A nest for babies, children, and the next generation to grow. To learn.
[It’s possible that this value-neutral concept of learning would not lead to benevolence but annihilation. Perhaps the pace of advance into god-like technology intercepts nature’s learning algorithm, evolution, at a time scale in which it cannot recover its goal of replication. We don’t even know if passive investing is a Taleb turkey problem, good luck with “was the die cast on our own demise when a proto-human first used a stone as a bludgeon?”]
We defined learning. We established why it’s important. It follows that, given its importance and our finite time, that faster is better (holding quality constant). We laid all this out but did not talk about how.
How do we learn?
Well, the how is critical because it’s the advice for new grads follows very naturally from the how.
We learn from risk. From bearing the cost of trial-and-error. School trains us to appease. Generally, the expedient way for a student to meet there 2 biggest goals, namely to get good grades while minimizing the teacher’s interference in the student’s affairs, is to flatter whatever dogmatic attachments the teacher harbors.
In elementary school, the answer the teacher wants and the right answer are aligned. Spelling, remedial comprehension, basic math operations. As children mature, the classes become more subjective. Risk becomes more inconvenient for the student as there is no marginal payoff vs reciting the “acceptable” answer since the ceiling is an A+ and is designed to be accessible. In real life, separation requires a combination of excellence and risk. I suspect many reading this will find it self-evident from experience that good grades had such steep conditions.
When you graduate, you want to optimize for learning and learning velocity. You want to prioritize for:
- Risk and accountability
- Working with exceptional people
The problem is these are not always straightforward to evaluate from the outside and the shortcuts that help evaluate don’t often work where you want them to.
Risk and accountability
You want responsibility quickly. In fact, this single fact was a major factor in choosing SIG over a bank when I graduated, despite far less up front pay. I wanted acceleration not a tantalizing y-intercept.
But zooming out a bit, I wanted trading more than banking because the whole field meant the chance to sink or swim sooner. I can’t explain it better than Paul Graham does in his classic essay How To Make Wealth:
Economically, you can think of a startup as a way to compress your whole working life into a few years. Instead of working at a low intensity for forty years, you work as hard as you possibly can for four. This pays especially well in technology, where you earn a premium for working fast…
Companies are not set up to reward people who want to do this. You can’t go to your boss and say, I’d like to start working ten times as hard, so will you please pay me ten times as much? For one thing, the official fiction is that you are already working as hard as you can. But a more serious problem is that the company has no way of measuring the value of your work.
Trading is not technology, but it’s similar in that you aren’t spending your time stapling and copying. But say you are not in trading or a start-up, how fast do you get to be at a client dinner or? How fast do you get to make decisions that impact the company’s p/l by choosing from a set of possibilities where there is a dispersion of value over the replacement or default decision? You want to ask the interviewer these questions, although I might try being more indirect unless you think impatience will reflect well.
Tactically, this means either trying to work for a smaller company, a startup in the limit case, or a company whose functional work teams are lean such that its members’ contributions are critical. Whatever the equivalent of a pod is in the industry you are entertaining.
Finally, there’s the canonical case in trading of someone getting fired and the boss tapping the trading assistant who is most familiar with the book to step up to manage the position. If you are talented and hang around the rim of any position where talent matters, you’re a 1 night away from all the responsibility you can handle.
My favorite such story is my friend Tony thrust into hosting Around The Horn for the first time when Max Kellerman suddenly left the show. That day was the start of an almost unheard-of 21-year run as an ESPN host. Talk about being thrown into the fire, that was the day after the Janet Jackson nipplegate Super Bowl.
Working with exceptional people
This one is hard to judge from a few rounds of interviews. The more experience you have with people, the better your judgment gets. You’ve seen how first impressions cash out, you recognize red flags, and the reps build intuition. But job-matching is naturally adversarial since the stakes are high. You’re trying to find who you’ll apprentice under, and that person who emerges as your eventual mentor might never have been in the room during interviews.
You’ll need to reach for heuristics. Unfortunately, the good ones are weak and the convenient ones are broken. Credentialism has some signal. I don’t think I’ve ever met an MIT grad who wasn’t smarter than me. But the error bars on the signal path from even honest credentials to efficacy are wide. The Stanford PhD who can’t reason out loud and the state-school kid who dismantles your argument in five minutes are data points your priors lack.
In fact, the state-school stud example is a common version of the credential trap. Berkson’s paradox tells us that inside any selected pool, the visible signals stop telling you what you think they tell you. If you see a short kid start on one of the best varsity basketball teams in the state, get ready to watch an absolute savage. The population correlation will correctly predict that taller people are better basketball players, but when you restrict the range to people already on a basketball team, the correlation loses predictive value and can even invert at the extreme. The Taleb version of this idea is that you don’t want a surgeon who looks like someone who plays a surgeon on TV. I say it a lot, but once you are familiar with Berkson’s, you see it everywhere. When you are in a wicked info landscape knowing these little anomalies can counteract the System 1 snap judgements trained on population-level observation.
Let’s move to company reputation. Reputation, the literal outside view, is a convenient heuristic. With enough samples can sketch a reasonable portrayal of a company. But it won’t work where my first piece of advice sends you — (not AI dash, sigh) to a lean team where your contribution is critical. Not enough samples. You’d be lucky to find even one. The places everyone has opinions about are, almost by definition, big and legible ones, therefore the least likely to give you the responsibility you actually want.
Unfortunately, there’s no shortcut to spotting exceptional colleagues. It is the bane of employers and job-seekers alike and it’s not for lack of effort. There is tremendous leverage not only in hiring well, but in the spirit of all this advice, to learning extremely fast if you find the right mentor.
An exception
We’re giving preference to going to a small company where you are pressured to learn fast and compress your timeline. But there’s a specific class of bigger institutions that earns an exemption. These are the places that have managed to formalize their the hard-won, domain-specific practical wisdom, or metis, that usually lives only within individuals. The default in companies, especially those that have a star culture, is that knowledge stays tacit and exits when good people leave.
Most organizations cannot or will not invest in the preservation and transfer of knowledge because it’s expensive. Some companies may see training as a box to check, while those with a long-term view will see it as an investment in a compounding edge. If you are talking to a firm in the second category, you are at the doorstep of rapid learning even if the org chart is a pyramid.
A few concrete examples of orgs known for their training:
- Navy SEALs are experts in decentralized command under pressure. They fuse a culture of risk assessment and extreme preparation. The government spares no expense in manufacturing these soldiers. (My favorite book during every man’s signature special ops rabbit hole is Lone Survivor which my older kid also loved.)
- SIG and Jane Street fancy themselves the SEALs of discretionary and quant trading, respectively.
- Goldman is in the fancy client business. I imagine it’s hard not to walk away without some durable closing skills and a smidge of dark arts.
- P&G built the academy that trained a generation of CMOs in brand management with their alumni in leadership roles across hundreds of giant consumer companies. The playbook for selling a commoditized product existed well before the era of digital marketing.
- Finally, Disney Institute and Chick-fil-A are widely considered the template for customer service.
Blood from a stone
It can’t be overemphasized. The purpose of a new grad’s job is rapid learning. Hopefully, in the field they want to be in (although the messy process of learning can also mean self-discovery and detours which are fine if not expected). But in the very real possibility that the field isn’t for them, it’s more important that the rapid learning is for skills that mesh with their talents and can find purpose in whatever direction their careers evolve.
Since it’s unlikely that one’s first job is their last, let me offer a final bit of advice. It’s time to change roles or companies once you feel your learning rate slow to a crawl. When professional gains feel like blood from a stone, it’s time to poke around. Explore. Maybe develop new skills at night while you bide your time. There’s even a chance these explorations rejuvenate your current role as AI has done for many.
I mention this because there can be a disconnect between learning rate and how things are currently going for you. This is definitely a trap. Results are a lagging indicator of effective learning. The cost of slow growth shows up in the future, delayed by the comfortable harvest seeded by your previous high learning rate.
I leave that as a warning. I will tell my kids the same.
Wrapping up
- Optimize for learning rate. Not pay, not prestige, not the y-intercept. You want acceleration.
- Actual learning requires risk and accountability
- When you find an exceptional colleague, help them and if you can learn from them you’ll compress time
- If opting for a big company, focus on those who make serious investments in training that will serve your career especially since you are unlikley to stay at the first big company you work for
- When the learning becomes blood from a stone, you need to make a change. Your own stagnation follows on a lag.
When in doubt, ask yourself where you will learn the fastest.