Returns for performance are superlinear in business
- Superlinear returns for performance are a feature of the world, not an artifact of rules we’ve invented. We see the same pattern in fame, power, military victories, knowledge, and even benefit to humanity. In all of these, the rich get richer.
- The most obvious case of superlinear returns is when you’re working on something that grows exponentially.
- Startups can also grow exponentially. If your product is only half as good as your competitor’s, you don’t get half as many customers. You get no customers, and you go out of business.
- The most obvious way to take advantage of superlinear returns for performance is by doing exceptionally good work. To do something exceptionally well, you have to be interested in it. Mere diligence is not enough. So in a world with superlinear returns, it’s even more valuable to know what you’re interested in, and to find ways to work on it.
- Choose work you have a natural aptitude for and a deep interest in. Develop a habit of working on your own projects; it doesn’t matter what they are so long as you find them excitingly ambitious. Work as hard as you can without burning out, and this will eventually bring you to one of the frontiers of knowledge. These look smooth from a distance, but up close they’re full of gaps. Notice and explore such gaps, and if you’re lucky one will expand into a whole new field. Take as much risk as you can afford; if you’re not failing occasionally you’re probably being too conservative. Seek out the best colleagues. Develop good taste and learn from the best examples. Be honest, especially with yourself. Exercise and eat and sleep well and avoid the more dangerous drugs. When in doubt, follow your curiosity. It never lies, and it knows more than you do about what’s worth paying attention to.
Humans are not automatically strategic We do not automatically —
- Ask ourselves what we’re trying to achieve
- Ask ourselves how we could tell if we achieved it (“what does it look like to be a good comedian?”) and how we can track progress
- Find ourselves strongly, intrinsically curious about information that would help us achieve our goal
- Gather that information (e.g., by asking as how folks commonly achieve our goal, or similar goals, or by tallying which strategies have and haven’t worked for us in the past)
- Systematically test many different conjectures for how to achieve the goals, including methods that aren’t habitual for us, while tracking which ones do and don’t work
- Focus most of the energy that isn’t going into systematic exploration, on the methods that work best
- Make sure that our “goal” is really our goal, that we coherently want it and are not constrained by fears or by uncertainty as to whether it is worth the effort, and that we have thought through any questions and decisions in advance so they won’t continually sap our energies
- Use environmental cues and social contexts to bolster our motivation, so we can keep working effectively in the face of intermittent frustrations, or temptations based in hyperbolic discounting
Tips for debugging (cc: )
- Reproduce with the smallest example. In the simplest environment.
- Read and re-read the error statement. Read the stack trace. Add more logging if you don’t know where the error is thrown.
- Change one thing at a time.
- Divide and conquer. Sometimes that means binary search on good/bad commits. Other times isolating the problem.
- Be open to debugging in different environments.
- State your assumptions.
- Get a second set of eyes on it.
- If you’re debugging some stateful code, think about how you ended up at that state. A recipe that (reproducibly) gets you to that state is often the path to fixing it.
- Look at the logs (all the logs).
- When in doubt, start with the most recent changes (especially dependencies changes).
- Make sure the code you’re looking at is actually getting executed (deployed code, external dependencies, etc.)
- Reduce the feedback loop — write a test or figure out the quickest way to test your debugging hypotheses.
- Look for patterns. Look for symmetry.
- Redundancy tends to hide bugs.
- Validate your inputs.
- Trust, but verify behavior.
Others
- People like beautiful things and for rational reasons. Because, what does a beautiful thing tell you? It tells you the person who made it really cared, and you can observe some superficial details, but probably they didn’t only care about those and did everything in else in slapdash way.
- People root for underdogs, but pick favorites when stakes are on the line.
- Improve health by working out five days a week, getting 8 hours of sleep, consuming 50 grams of fiber, and consuming 1 gram of protein for every pound of body weight.
- GPT-4 is catching up in speed, closing the latency gap with GPT 3.5. The median request latencies are stable across both models, staying below 1 ms per token. In the 99th percentile, latencies have reduced by more than half in just three months.
Substack edition
Returns for performance are superlinear in business
Superlinear returns for performance are a feature of the world, not an artifact of rules we've invented. We see the same pattern in fame, power, military victories, knowledge, and even benefit to humanity. In all of these, the rich get richer.
The most obvious case of superlinear returns is when you're working on something that grows exponentially.
Startups can also grow exponentially. If your product is only half as good as your competitor's, you don't get half as many customers. You get no customers, and you go out of business.
The most obvious way to take advantage of superlinear returns for performance is by doing exceptionally good work. To do something exceptionally well, you have to be interested in it. Mere diligence is not enough. So in a world with superlinear returns, it's even more valuable to know what you're interested in, and to find ways to work on it.
Choose work you have a natural aptitude for and a deep interest in. Develop a habit of working on your own projects; it doesn't matter what they are so long as you find them excitingly ambitious. Work as hard as you can without burning out, and this will eventually bring you to one of the frontiers of knowledge. These look smooth from a distance, but up close they're full of gaps. Notice and explore such gaps, and if you're lucky one will expand into a whole new field. Take as much risk as you can afford; if you're not failing occasionally you're probably being too conservative. Seek out the best colleagues. Develop good taste and learn from the best examples. Be honest, especially with yourself. Exercise and eat and sleep well and avoid the more dangerous drugs. When in doubt, follow your curiosity. It never lies, and it knows more than you do about what's worth paying attention to.
Humans are not automatically strategic
We do not automatically —
Ask ourselves what we’re trying to achieve
Ask ourselves how we could tell if we achieved it (“what does it look like to be a good comedian?”) and how we can track progress
Find ourselves strongly, intrinsically curious about information that would help us achieve our goal
Gather that information (e.g., by asking as how folks commonly achieve our goal, or similar goals, or by tallying which strategies have and haven’t worked for us in the past)
Systematically test many different conjectures for how to achieve the goals, including methods that aren’t habitual for us, while tracking which ones do and don’t work
Focus most of the energy that isn’t going into systematic exploration, on the methods that work best
Make sure that our "goal" is really our goal, that we coherently want it and are not constrained by fears or by uncertainty as to whether it is worth the effort, and that we have thought through any questions and decisions in advance so they won't continually sap our energies
Use environmental cues and social contexts to bolster our motivation, so we can keep working effectively in the face of intermittent frustrations, or temptations based in hyperbolic discounting
Tips for debugging
Reproduce with the smallest example. In the simplest environment.
Read and re-read the error statement. Read the stack trace. Add more logging if you don’t know where the error is thrown.
Change one thing at a time.
Divide and conquer. Sometimes that means binary search on good/bad commits. Other times isolating the problem.
Be open to debugging in different environments.
State your assumptions.
Get a second set of eyes on it.
If you're debugging some stateful code, think about how you ended up at that state. A recipe that (reproducibly) gets you to that state is often the path to fixing it.
Look at the logs (all the logs).
When in doubt, start with the most recent changes (especially dependencies changes).
Make sure the code you’re looking at is actually getting executed (deployed code, external dependencies, etc.)
Reduce the feedback loop — write a test or figure out the quickest way to test your debugging hypotheses.
Look for patterns. Look for symmetry.
Redundancy tends to hide bugs.
Validate your inputs.
Trust, but verify behavior.