Writing about my job: Internet Blogger

Response to Aaron Gertler’s You should write about your job.


I’ve been writing since September 1st, 2020, initially about voting and mechanism design, then about an increasingly varied assortment of topics ranging from the importance of economic growth within an EA framework, to the organization of research institutions and more generic career advice.

The blog has been moderately successful in terms of attracting attention from people I respect without causing any major scandals or other negative effects.

I occasionally have some interruptions, but mostly work on the blog full time.


Some skills I’ve developed include:

  • Self-management: I have no deadlines, no manager, and generally speaking, no accountability. If I don’t choose to do something, it won’t get done. The sub-skills include finding good ideas for posts, prioritizing them correctly, avoiding distractions, and actually executing and “shipping”. Anecdotally, many of the people I talk to seem to be held back here, whether they’re blogging, starting a company or just trying to take a hobby more seriously. If all I got out of the last 9 months was this skill, it all would have been worth it.

  • Patience: It’s one thing to build intuitions for exponential growth, another to actually follow through and make investments on long time scales. Since we’re systematically over-exposed to successful blog posts, your view of success is likely distorted, and it will take far longer than you think to become a good writer and to get noticed.

  • Writing: This sounds obvious, but it’s worth noting that you don’t already have to be a good writer. The critical thing is not just practice, but having feedback loops, mentorship and goals. Many bloggers have public contact info, and will happily read your draft.

  • Talking to people: I started blogging in part because I hated lockdown-era Zoom calls, and just wanted to avoid meetings and work alone in peace. Recently, as I’ve ramped up on more rigorous research projects, I’ve had to proactively reach out to more senior researchers, ask them for introductions and email authors for clarification or feedback. I was pretty bad at this initially, and would just publish without talking to a single person, even if I was a total amateur in a field with several readily-accessible experts. Since then, I’ve gotten a lot better at figuring out who to talk to, which questions to ask them, and then actually taking the time to do it.

These are all skills I’ve developed during the course of blogging, but you can also see them as (very soft) pre-requisites. If you’re really terrible at self-management, blogging might not be a good career. The degree to which this is true depends on your views on growth mindset, your own learning ability, etc. I wrote here that several prominent bloggers were “losers” in some sense in their previous endeavors, and so you shouldn’t let failure in some other domain discourage you.

Career Growth

Blogging can be an end-unto-itself, but can also be a useful and low-cost way to earn a formal role at a research or media organization. You quickly build up a portfolio of past writing projects, as well as an audience and potentially connections. Some potential next steps could include:

I haven’t applied for any of these myself, but have talked to people selecting for these roles, and have some sense that they believe blogging is a reasonable entry point. Of course, that depends a lot on what kind of blogging you end up doing, and how well it fits with the interests of those programs.

Path to Impact

Scott Alexander famously wrote “The less useful, and more controversial, a post here is, the more likely it is to get me lots of page views.” In one view, this means you should try to:

  • Write some controversial and popular posts, even if they’re useless
  • Do more useful writing, leveraging your newfound audience as a path to impact

I don’t think Scott is endorsing this strategy, and I wouldn’t either. As tempting as it is, the problem is that readers are not fungible. You might end up with 10,000 subscribers, but it doesn’t help if they’re exclusively the kind of people attractive to useless controversy.

It’s difficult to formalize, but my own theory of change is closer to:

  • Publish good writing, often useful, almost always in good faith
  • That aligns with my intrinsic interests
  • That aligns with the interests of people I consider to be influential
  • Try to correct moral or epistemic errors within that community of readers

The tricky part is “people I consider to be influential”. This can mean people with money, or people with large audiences, or people those people respect and listen to. To be clear, this is not really an explicit strategy on my part, but it is how I justify my particular approach to writing.

Other possible paths to impact include:

  • Solve specific problems in an important domain, using blogging as a faster and more dynamic alternative to conventional research.
  • Write for a popular outlet like Future Perfect and try to slightly shift the behavior, beliefs and values of a million readers.
  • Provide independent and sometimes contrarian viewpoints that lend perspective to an existing community.

This last point is somewhat contentious, and can obviously go astray. You also have to play the balancing act of remaining close enough to the community to be trusted, but not so close that you share all their assumptions.


Per week:

  • 20 hours: Writing, doing small bits of research for a specific writing project. Writing long replies to emails or commenting on blog post drafts.
  • 8 hours: Reading blogs, papers. I don’t have a particular news source I follow, and don’t curate any feeds. I mostly just get sent articles from various friends, follow the hyperlinks, and then end up with a bunch of bookmarks to work through.
  • 1 hour: occasional phone call, often informal chats with someone who just wanted to talk without a particular agenda.

All those numbers might be +/- 50%, depending on how I’m feeling. I’ve also taken a couple months of vacation since September.

I received a small amount of funding from Emergent Ventures. From what I understand, grants go as high as $50,000, but that’s not confirmed. You could also get around $80,000k/year from EA Grants, or seek out private donors. I haven’t asked the Survival and Flourishing, but historically they seem to give out around $50k for individual grantees. You could also explore Patreon and Substack.


Though it’s hard work with uncertain rewards, there are benefits:

  • Meet cool people: If you like football, tough luck, you’ll still never meet Tom Brady. If you like weird internet blogs, good news! You can very quickly get in touch with the people you admire, and have a decent chance of getting to hang out with them. This is fun in some kind of unhealthy parasocial sense, but it is genuinely nice to meet people doing work you’re interested in, and nice to have those people be interested in your work too.

  • Flexibility: You have to be careful with this, but no real accountability also means you can do whatever you want! That’s scary, but also very fun, especially post-vaccine.

  • Ride the Hedonic Treadmill: It’s not the most popular carnival attraction, but it is the most universal. At some point, you will get your first 10 followers, and it will feel unreasonably good. Of course, there are downsides, but it’s not clear to me that you really do “pay back” the happiness when you return to baseline. The weird thing about exponential functions is that their derivatives are also exponential!

  • Productivity: When I had a day job, I felt languid, tired and unmotivated constantly. This led to doing poor work, and feeling bad about myself. As a blogger, I have a lot of personal accountability and have found it exceptionally motivating. If I don’t do work, it won’t get done. Accordingly, I work fairly hard, but this doesn’t take the form of longer hours so much as getting way more done per hour.


As always, you’re welcome to email me. If you have questions you think other people would be interested in, please post them on the EA Forum discussion.

See Also

Holden KarnofskyMy current impressions on career choice for longtermists
Alexey Guzey - Why You Should Start a Blog Right Now
Nadia EghbalReimagining the PhD

And previously on my blog:

Generating Intuitions for Exponential Growth

You’ve probably heard of the Rule of 70:

To estimate the doubling time of an exponential function, just divide 70 by the growth rate.

For some rates, this works really well. At 2% annual growth, the rule gives 35 years, and the actual value is 35.003 years. Other times it fails horribly. At 70% growth, the rule predicts doubling in one time step, but it actually takes 1.3.

How does the heuristic perform in general? Not that well. It’s accurate at 2% growth, but then quickly converges to being off by 0.3 timesteps.

An alternative, the Rule of 72, performs a bit better, converging to being off by 0.28, which is still not great:

So why was the rule ever popular? An early version is attributed to 15th century Italian accountant Luca Pacioli. Coincidentally the same guy who failed to teach Leonardo DaVinci math. In Summa de arithmetica, he writes:

In wanting to know of any capital, at a given yearly percentage, in how many years it will double adding the interest to the capital, keep as a rule [the number] 72 in mind, which you will always divide by the interest, and what results, in that many years it will be doubled. Example: When the interest is 6 percent per year, I say that one divides 72 by 6; 12 results, and in 12 years the capital will be doubled.

For 6 percent, the error is only 0.1, which is not yet too bad. In general, it’s helpful to think of the Rule of 72 as a heuristic that works decently for a certain range of values.

That might sound like a crippling limitation, but for typical investments it’s actually a decent range. Hedge funds average around 7.5%, and the S&P has returned an annualized 9.81% since 1994. Unless you’ve invested your money with Byrne Hobart, your returns are likely in this modest range.

Luca Pacioli, a good mathematician, bad teacher, and responsible investor who failed to foresee the rise of meme stocks.

I’m giving heuristics a bad name. Truth be told, they’re far far better than your intuition.

From Stango and Zinman’s publication on Exponential Growth Bias and Household Finance:

Exponential growth bias is the pervasive tendency to linearize exponential functions when assessing them intuitively.

This has real implications:

exponential growth bias can explain two stylized facts in household finance: the tendency to underestimate an interest rate given other loan terms, and the tendency to underestimate a future value given other investment terms. Bias matters empirically: More-biased households borrow more, save less, favor shorter maturities, and use and benefit more from financial advice, conditional on a rich set of household characteristics.

It’s not that being off by a small timestep will ruin you. It’s that the errors compound, such that the longer your time horizon, the more horrendously skewed your linearized intuition gets:

If you work in startups, finance or anything adjacent, this might be your time to feel smug. Middle America makes poor financial decisions, but surely your experiences have improved your savvy? A paper on Misperception of exponential growth would tend to disagree:

This group of professional decision makers did not show less underestimation than naive subjects… Underestimation appears to be a general effect which is not reduced by daily experience with growing processes.

Equivalent papers are likely being written as we speak on the inability for both lay and expert analysts to properly evaluate and intuit viral growth during the COVID-19 crisis. [1] Already, we have the London School of Economics on the public’s inability to understand log scales. [2]

Being able to wrap our minds around growth matters a lot, and will matter increasingly in the wacky world of pandemics, tech startups and bitcoin. Exponential growth rules everything around me, and none of us can make sense of it. Sam Altman was right: “Everyone’s intuition for exponential growth sucks, so do the math.”

Still, I can’t break out a spreadsheet every time I want to think about a decision. It’s useful to have tools for thought we can actually fit in our heads. The charts above suggest a slight correction to the Rule of 72:

To estimate the doubling time of growth greater than 12%, divide 70 by the growth rate. Then add 0.3.

This gives us a heuristic that works decently well, never exceeding an error of 0.1:

But how does the error compound? It might not sound like a big deal to be off by a tenth of a timestep. But again, in the world of exponentials, that can be a big deal. Off by one timestep could mean off by an entire doubling, easily the difference between riches and ruin.

This chart also makes it clear that the heuristic is not quite as consistent as previously suspected. Remember when I said it converges? That was a lie. Here’s a more cosmopolitan view of error for really big growth rates:

It does actually start to level off at the end, so if you have to deal with growth rates over 1000%, there’s another correction you could make.

This might all sound silly, but in a post-Covid world, it shouldn’t. Daily cases in the UK grew 400% this last month. Depending on your timescale, you might think of that as 38% weekly growth, or as 24,000,000,000% annual growth. [3] [4]

So far we’ve been looking solely from the perspective “given a growth rate, what’s the doubling time?” But these problems can come from a few different angles:

  1. Given trends with different initial values and growth rates, how long until one overtakes the other?
  2. If a growth rate increases, how much does doubling time shorten?
  3. Given a doubling time, what’s the 1000x time?

#3 is trivial (just multiply by 10), #2 just requires using a heuristic twice and comparing, #1 is really hard to do in your head. I shared some weird examples of this in The Unreasonable Effectiveness of Starting Over. Honestly, the best “heuristic” is to have Wolfram Alpha at the ready, and input α0x - α1x = Δ, where Δ is the ratio of initial values, α0 and α1 are growth rates, and x is the number of timesteps until they cross over.

There are also cases where you’re trying to estimate growth, in conjunction with other factors. To write Golden Handcuffs, I had to model a case where:

  • A software engineer makes $X/year
  • Generates savings after taxes and cost of living
  • Invests that capital at some growth rate
  • Does so continuously over several years

In this case, you’re not just considering the growth rate of some lump sum. You’re considering the growth rate of a growing pool of capital, modulated by a dynamic tax rate. If you want to play with this yourself, I have a model for State + Federal + FICA taxes here.

This is a key consideration for the Financial Independence Early Retirement community. Over at Scattered Thoughts, Jamie Brandon has a great visualization of the non-linearities in saving:

The point being, when you grow your savings from 0.2M to 0.4M, your runway doesn’t just double, it skyrockets! At 30k in annual spending, a 0.2M nest egg will only get you 8 years, but 0.4M extends your runway to 57 years, and at 0.41M you’re financially self-sustaining indefinitely!

One upshot of this discussion is that while intuitions are okay, heuristics better, and math unreasonably effective, what we really need is tools to rapidly model weirder and more complex scenarios.

Or maybe one day we’ll just develop better notation. As Alexey Guzey quotes Bret Victor:

…back in the days of Roman numerals, basic multiplication was considered this incredibly technical concept that only official mathematicians could handle … But then once Arabic numerals came around, you could actually do arithmetic on paper, and we found that 7-year-olds can understand multiplication. It’s not that multiplication itself was difficult. It was just that the representation of numbers — the interface — was wrong.

Can we run this process in any kind of systematic fashion? Michael Nielsen has an extraordinary example, alongside a longer exploration with Andy Matuschak.

But that’s a story for a different post.

See Also
Neil Hacker – Compounding


[1] The Stango/Zinman paper was written in 2009 just after the financial collapse. Similarly, the Misconceptions came out in 1975, catching the tail end of the 1970s recession and oil crisis when unemployment soared to 9%, then considered a historic level.

[2] Full paper available here, and a response from Andrew Gelman: Let them log scale.

[3] I wasn’t expecting a number that big either, but I guess that’s the whole point of the post. Sanity check: 5x M/M growth = 512 Y/Y growth = 244,140,625x = 24,414,062,499%.

[4] Of course, eventually Covid runs out of people to infect.

If you want to play around with heuristics and growth rates, you can clone the model here.

Become a Billionaire Part II: You're Not Even Trying

Follow up to Life Advice: Become a Billionaire.

On Reddit, the comments are skeptical. Respondents suggest that perhaps there are more important things in life than money, and even if you do start a company, selling it for $10 million is better than risking it all for a chance at a billion dollar exit.

Which is funny, because those are precisely my reasons for optimism.

When you hear that the success rate for startups is low, or that very few founders succeed in hitting billion dollar valuations, remember that you’re including the entire population of people who, it turns out, actively don’t want to become billionaires in the first place.

What you should be asking is, what are the odds of becoming a billionaire, conditional on actually wanting to? Conditional on even trying? Conditional on not machine gunning yourself in the foot?

As it turns out, the mere willingness to not sell matters a lot. Here’s Peter Thiel:

The most important moment, in my mind, in the history of Facebook, occurred in July of 2006. The company had been around for 2 years, it was still just a college site. Maybe 8 or 9 million people on the site. The revenues were tracking to about 30 million, no profits. And we received an acquisition offer from Yahoo for a billion dollars.

…full disclosure, I think that both Breyer and myself thought that on balance we should take the money and run. But Zuckerberg started the meeting, and the first thing he said was “it’s kind of a formality, we have to have a quick board meeting, shouldn’t take more than 10 minutes. We’re obviously not going to sell“.

So sure, we can do the math, write out some probability distributions, factor in conditional risk. But here’s the bottom line: The more unreasonable it is to become a billionaire, the less competition there is, and the easier you should expect it to be.

This is perverse logic, and you could argue that it proves too much and could justify any bad decision. But it applies here anyway. It’s not that becoming a billionaire is actually irrational. It’s that people look at the failure rates, infer that it’s hard, that the rewards aren’t worth it, do some moral or hedonic calculus, and give up prematurely.

You could object that Reddit commenters are not the same population as YC founders, and this is all a tremendously unfair comparison. That’s possible. But I wouldn’t be too surprised if it turned out most people are just looking for an easy exit. After all, that’s the reasonable thing to do, isn’t it?

So stop complaining about the “risk of failure” when you’re not even trying to succeed.

This set of posts is not my most rigorous, but let’s run some numbers anyway.

Y Combinator reports a $400+ billion valuation for its top 100 companies, out of around 2000 it’s ever funded. So at first approximation, that’s $200 million in market cap per startup, or around $100 million per founder. But founders don’t retain all equity. What’s worse, dilution occurs as a function of rounds raised, so the bigger the pie, the less likely you are to own a large share of it. In practice, it’s not that bad. Stripe’s founders reportedly own around 23% of the company, and Airbnb’s founders collectively own somewhere around 40% of their company. That works out to 11.5% and 13% ownership per founder.

Still pretty good!

But remember again that we’re talking about a snapshot at a moment in time. Many of these companies are still growing rapidly. Using archive.org, I was able to go back and compile some data:

I also compiled data on the cumulatively number of startups funded. Taking a simple average gets us:

Again, it’s a power law, so even if the average is really high, the median outcome is probably $0. But if you’re a risk-neutral hits-based utilitarian minded person, that doesn’t matter. It’s pure expected value.

So stop complaining that it’s a poor bet. You have no idea how good it is, and it’s getting better every year.

Finally, let’s take a harder look at the happiness data.

I originally shared this chart from Matthew Killingsworth (2021):

On EA Forum, Julian Hazell and Michael Plant plot the data without the log scale and z-scores, and get a much more pessimistic interpretation:

With a linear scale, it’s easier to see how hard returns drop off.

…but wait a minute, they don’t just plateau harder, they actually dip down! Compare again to the original chart. This isn’t just a matter of axis-choice, it’s a bizarre discrepancy.

Rohin Shah asked the same question, prompting this response from Michael:

there was a discrepancy between the data provided for the paper and the graph in the paper itself. The graph plotted above used the data provided.  I’m not sure what else to say without contacting the journal itself.

Kieran Healy noticed the same problem and produced a similar plot:

As the original paper explains:

Mean levels of experienced well-being (real-time feeling reports on a good–bad continuum) and evaluative well-being (overall life satisfaction) for each income band. Income axis is log transformed. Figure includes only data from people who completed both measures.

Healy is unsure, but offers the following explanation:

The z-score means in the replication package are, presumably, calculated from all the observations for each measure. But if the figure is showing a subset of the two (i.e. only observations from people who answered both questions) then the z-score means across income levels will be slightly different, depending on who is excluded… That might well be just measurement error, given the vagaries of income reporting and small-n noisiness at very high incomes, but it would directly cut against the main claim of the paper.

To summarize:

  • It’s unclear what’s actually happening in the original paper
  • It’s possible z-scores are being calculated at each income band amongst different subpopulations
  • This creates two separate interpretations (“wellbeing continues to increase, just more slowly”, “wellbeing caps out at $400k”) depending on which methodology you choose
  • In either case, Life Satisfaction continues to increase with income

This mirrors the Kahneman and Deaton (2010) finding that measures of wellbeing plateau, but measures of life satisfaction do not.

The supplement to Killingsworth (2021) provides some additional useful context, including an interesting section titled “why the current results might differ from past results showing a plateau in experienced well-being”:

Examining Figure 1 in the 2010 paper finding a plateau (3) shows that positive feelings (“positive affect”) appeared to have been at the response ceiling in slightly more than 70% of responses at the lowest income level, and around 87-88% of responses at upper income levels. Accordingly, the vast majority of participants in that study were indicating the highest possible level of positive feelings the scale allowed at incomes of $75,000, limiting the ability to detect further improvements in people with incomes above $75,00

You could counter-argue that this is not a statistical ceiling effect. People just genuinely have the best lives possible. I firmly disagree. As noted earlier:

The Cantril Ladder depends upon the capacity to imagine a better possible life. At the moment, it is difficult to conceive of a world in which diseases are eradicated, although such a world would make us much happier. Conversely, we can imagine someone from the distant past reporting “I’ve lost two children to disease, lost my wife to childbirth, lost half my friends to war, but the harvest is good and we have a good chance of surviving the winter months, so maybe a 7/10”. We should not take this as strong evidence that their life is nearly as good as possible.

We still have disease, we’re still superstitious and ignorant, still caught up in tribal violence.

So you don’t want to be a billionaire. Fine. But just stop fetishing poverty. Stop acting like you can shield yourself from moral corruption of the market so long as you achieve the right work-life balance. If you’re going to pretend to be anti-capitalist, at least quit your day job and do something with your life.