Hide table of contents

Comment Permalink

These are all good reasons.

I think 5he biggest reason you've missed against starting a for profit is risk of failure. It's really really high, maybe 5x or 10x that of a non profit.

Again I really like your argument, and think that despite the risk of failure, more EAs should be trying their hand at for profits.

Wave is a beautiful company.

Just don't name it Ajar AI or Anthropomorphic :D

calebpApr 211

The focus of the article was on reasons to start a for-profit, but I'm not sure that the risk of failure for for-profits is significantly higher than that of non-profits (ambition adjusted). Why do you think it's so much higher?

See in context

Consider starting a for-profit company instead

by calebp

Apr 212 min read 39

122

Career choiceEffective givingOpinionEarning to giveEntrepreneurship

Frontpage

Consider starting a for-profit company instead

Scalability

Standards

Learning Potential

Earning to Give Potential

38 comments

I speak to many entrepreneurial people trying to do a large amount of good by starting a nonprofit organisation. I think this is often an error for four reasons:

Scalability
Capital counterfactuals
Standards
Learning potential
Earning to give potential

These arguments are most applicable to people aiming to grow very quickly (startups).^[1]

Scalability

There is a lot of capital available for startups, and established mechanisms exist to continue raising funds if the ROI appears high. It seems extremely difficult to operate a nonprofit with a budget of more than $30M per year (e.g., with approximately 150 people), but this is not particularly unusual for for-profit organisations.

Capital Counterfactuals

I generally believe that value-aligned funders are spending their money reasonably well, while for-profit investors are spending theirs extremely poorly (on altruistic grounds). If you can redirect that funding towards high-altruism value work, you could potentially create a much larger delta between your use of funding and the counterfactual of someone else receiving those funds. You also won’t be reliant on constantly convincing donors to give you money, once you’re generating revenue.

Standards

Nonprofits have significantly weaker feedback mechanisms compared to for-profits. They are often difficult to evaluate and lack a natural kill function. Few people are going to complain that you provided bad service when it didn’t cost them anything. Most nonprofits are not very ambitious, despite having large moral ambitions. It’s challenging to find talented people willing to accept a substantial pay cut to work with you. For-profits are considerably more likely to create something that people actually want.

Learning Potential

Most people should be trying to put themselves in a better position to do useful work later on. People often report learning a great deal from working at high-growth companies, building interesting connections, and gaining legible experience, which opens up interesting opportunities to do high-impact work. Often, non-profit work is less fast-paced and interesting.

Earning to Give Potential

Many nonprofit organisations are funding-constrained. I am more optimistic about value-aligned people becoming very wealthy and then funding nonprofit organisations than about efforts to persuade those with generational wealth. Convincing individuals who have already built successful companies to donate a significant fraction of their wealth seems promising, but I suspect one of the key bottlenecks is the scarcity of people who have built successful companies and could serve as compelling advocates. Most importantly, powerful AI systems are on the horizon, creating many interesting opportunities for new companies.

^{^}
There are, of course, reasons not to start companies, such as worries about value drift, poor personal fit, or the specific project being a poor fit for for-profit work and being exceptionally valuable. I often feel like people overstate these, but I don’t argue against them in this essay.

122 Reactions

Comments38

Sorted by

New & upvoted

Click to highlight new comments since: Today at 9:33 PM

huwApr 2281

I actually made this exact decision, just in the opposite direction! Last year, I had a pending offer from AIM’s Founding to Give programme, and another to be CTO at Kaya Guides (my current role).

Odds & magnitude of success

For AIM incubatees, using historical data, I calculated that:

30% of CE incubatees received funding from a top funder (OP or GiveWell)
About 2/3 of funding for CE incubatees came from OP and GiveWell, so we can multiply by 3/2 to get the total funding amount they would’ve received
Based on real-world grant data from OP and GiveWell, the median AIM charity could expect to receive US$1.4M in lifetime funding, the mean could get US$20M, and the top handful (i.e. a GiveWell Top Charity) could get hundreds of millions

For FTG incubatees, I borrowed from AIM’s own BOTEC, but substituted more real-world data when I had it to model distributions. This models:

The rate of incubator (ex. YC) acceptance should be about 20–60%, given that ~40% of CE charities don’t languish.
The historical odds of a YC company getting a substantial exit is about 11–16%
AIM then model out company valuations and founder shares at exit.
The median YC startup might yield about US$2M donations at exit, and the mean about US$56M
The median YC healthtech startup might yield about US$2M at exit, and the mean about US$4M. (Much less of a long tail)

A healthtech startup might also have significant positive impact through its work, so I decided to model that in too. It increased the potential value by about 5× on average.

Mutliplying through, the summary statistics I came up with for each option (in value-equivalent US dollars):

CE incubatee

p5: ~$100k
Median: ~$500K
Mean: ~$6M
p95: ~$23M

FTG incubatee

p5: ~$0 (actually negative in the model, but intuitively, many startups fail, make no money, and have no positive impact)
Median: ~$2M
Mean: ~$56M
p95: ~$11B

FTG incubatee (healthtech)

p5: ~$6M
Median: ~$14M
Mean: ~$21M
p95: ~$64M

For Kaya Guides specifically, I also added a counterfactual impact factor that bumped it up a little bit, because I think we have an unusually unique opportunity to pioneer a new intervention globally.

Here’s the notebook I used to calculate all this. Maybe I should write this up as a full post someday.

Personal fit / corporate pressure

I spoke to a great deal of people, and one thing that came up repeatedly is that if you’re in the for-profit game to have a direct impact, you should probably forget about it. Opinions differ a lot here, but I’ve seen the inside of a handful of tech companies—and many more products from the outside—and I do generally believe that at the point you’re big enough to have a significant impact, it’s quite likely that your investors will pressure you to squeeze money out of it in a way that will likely ruin said impact. This is especially true if you’re relying on some of your impact coming from donated exit money, which will, in all likelihood, reduce your control over the company.

The rest of my decision here is personal, but I’m still not fully sold on Earning to Give. I tried it earlier in my career, and I found that the toll of not doing direct work was so mentally debilitating that I wasn’t able to effectively work that hard. I think that being a highly successful startup founder is probably really hard, and I’m evidently not constitutionally motivated enough to just make money.

Conclusion

Broadly, I would see the decision like this: I think founding a startup and founding a charity have highly transferrable skillsets, but require extremely different personalities and constitutions. For me, then, this narrowed the FTG paths I would be effective at to just healthtech, which had a much lower ceiling. Frankly, my model for healthtech wasn’t meaningfully larger than my model for Kaya Guides, and at this point, I just went with my gut feeling.

(I know this comment is a bit structurally messy but I hope it elucidates something of value and provides some harder numbers to Nick’s intuitions)

calebpApr 229

I'm not sure I follow your analysis - what does these numbers represent?

CE incubatee

p5: ~$100k
Median: ~$500K
Mean: ~$6M
p95: ~$23M

I do generally believe that at the point you’re big enough to have a significant impact, it’s quite likely that your investors will pressure you to squeeze money out of it in a way that will likely ruin said impact.

I think this could be a crux. Overall, I'm not very convinced that this effect is particularly strong relative to incentives that apply in the nonprofit case. Some rough heuristics for why I think it's often overstated:
* Once a company is operating at scale (e.g. Series B) it's rare to see large pivots, and if a change isn't large why would it cost a lot of impact?
* Most incubators want people to work on things they are excited about, there's only so much you can push people to change directions before the founder gets fed up.
* I suspect that investors might place a lot of pressure on companies to change the way they operate - but overall (relative to nonprofits) I think that's probably a good thing.

I think I'm also less convinced than you are that the skillsets for for-profit and non-profit founders are highly transferable. I guess it depends a lot on the particular organisations.

huwApr 226

I looked at a handful of mental health startups to inform my guesses on impact. I looked the most deeply into BetterHelp, but you can clearly see through their numbers that their prices have almost doubled in 5 years (and steadily, too, this wasn’t a COVID thing). From the research I did, my sense was that it wasn’t getting passed back onto their counsellors, nor fuelling an increase in growth spending. There’s no way it got twice as expensive to deliver therapy.

I think if we had to point to a single mechanism, once you run out of user growth—as BetterHelp have—your investors push you to find an equilibrium price. That price is necessarily going to be higher than the price that guarantees the broadest impact, and likely to be higher than the price for the most (breadth x depth) impact.

My best guess is regional pricing can act as a crude form of means-testing, but this probably comes with a perverse incentive to ignore the cheaper regions (as BetterHelp have—almost all of their users are in the U.S.).

(All of that goes out the window if you don’t go direct to consumers—I think deeper forms of healthtech might be quite value-aligned!)

James HerbertApr 222

Thanks for sharing your analysis! Very interesting to read.

I do generally believe that at the point you’re big enough to have a significant impact, it’s quite likely that your investors will pressure you to squeeze money out of it in a way that will likely ruin said impact.

Did you look into whether steward-ownership is a viable strategy for mitigating against this risk?

huwApr 223

I didn’t. It evidently works—as do cooperatives, which I was also excited to found—but I think the big worry is up the top end. It’s very hard to imagine a FAANG company structured this way. And some of the average-case calculations above are skewed upwards by a handful of top success stories.

Knight LeeApr 281

At Google, Larry Page and Sergey Brin control 51% of the shareholder vote thanks to their supervoting stock.
At Meta/Facebook, Zuckerberg controls 61% of the vote.
Anthropic is a public benefit corporation.
OpenAI was supposed to be controlled by a nonprofit board though Sam Altman is trying to convert it into a public benefit corporation.
Shareholders are also unlikely to remove Elon Musk from Tesla even if he does a lot of things against Tesla's interests.

Executives are under intense pressure to make profit to prevent the business from going bankrupt, and maybe to get bonuses or reputation, but the pressure to avoid getting voted out by shareholders is relatively less.

Charities have a lot of the same pressures (minus the bonuses).

I don't have any expertise, I may be totally wrong.

huwApr 222

what does these numbers represent?

How much money a CE charity might be able to raise on average. (This makes the assumption that deployed cash from a charity is roughly equivalent to donated cash from Founding to Give which is what the other numbers represent).

JasonApr 2224

Did you adjust for the likelihood that some of the funding secured by the charity would likely would have gone to other effective charities in the same cause area?

huwApr 236

No, but I think that’s reasonable in most cases (although hard to figure out exactly how to allocate it).

calebpApr 2312

I don’t quite understand why you’re comparing donated money from people earning to give to money consumed by a charity.

I think a better comparison could be to just convert everything into impact adjusted dollars imagining that you’re able to sell off your impact equity. In this scheme it’s clearer that taking more money from value aligned funders is bad, whilst taking money from non-aligned funders is roughly neutral and donating a bunch of money is very good. The charity has to essentially “pay back” in impact the value aligned funder to get itself out of a hole, whereas the for profit doesn’t need to worry about that (on impact grounds).

To be clear, I think many nonprofits do a lot of good and are well worth funding - I spend most of my time trying to fund them - but it is harder to work out net positive if you’re consuming a bunch of fungible and value aligned resources.

huwApr 239

Thanks for the feedback! I’m not really smart enough to figure something like that out tbh, and by the point I’d seen that my realistic options were within an order of magnitude of each other (and both high-risk with high overlap) I was pretty satisfied that my decision was likely gonna hinge on something else.

Maybe your adjustment would take it outside that range but I think at the point of extreme success these charities would be selling impact at competitive rates (so funders would be getting marginal value out of them), and more than likely going to counterfactual funders (ex. government). Maybe this is truer in GHD and especially mental health than in animal welfare, which seems more concentrated. But yeah, at this point I was pretty satisfied that Kaya Guides had minimal risk of substantial funding displacement in a success scenario (I can’t be too specific about this in public), so I picked it.

(Maybe again—I’m just highlighting my specific scenario, there’s definitely an attempt to generalise here but I didn’t think too hard through it)

calebpApr 2318

Makes sense. Sorry I didn’t say this before but I really appreciate your comments on this post and the high effort modelling that you did when working out what to work on. I think it’s a great example to set for the community and shows how seriously you can take these decisions (if you want to).

NickLaingApr 242

I think this point is really important. If the charity raises 20 million from GiveWell that's great, but counterfactual impact is unlikely to be very big. If they raise it though from other foundations or individuals, I would say the counterfactul impact per dollar might be 5-10x as much.

But in general agree with almost everythng @huw nicely points out here.

And yes don't expevt @huw to account for this in a calculation. I think it would be nice to have a super well-researched estimate of "Value per dollar" of givewell funding vs. other foundations, and am surprised this hasn't been done.

JasonApr 244

If the data were available, the amount an CE charity might be able to raise on average from funders other than highly-aligned funders might work better if someone were deploying your analysis for a different decision about whether to found a CE charity vs. earn to give. You've mentioned that you were "satisfied that Kaya Guides had minimal risk of substantial funding displacement in a success scenario," so it makes sense that you wouldn't adjust for this when making your specific decision.

(The working, rough assumption here is that the average CE charity can put a dollar to use roughly as well as the average GiveWell grantee or ACE-recommended charity -- so moving $1 from the latter to the former produces neither a net gain nor a net loss. That's unlikely to be particularly correct, but it's probably closer to the actual effect than not adjusting for where the money went counterfactually).

Mo PuteraApr 227

Maybe I should write this up as a full post someday.

I hope so! This was a super-illuminating comment, thanks for writing it up.

NickLaingApr 21*12

These are all good reasons.

I think 5he biggest reason you've missed against starting a for profit is risk of failure. It's really really high, maybe 5x or 10x that of a non profit.

Again I really like your argument, and think that despite the risk of failure, more EAs should be trying their hand at for profits.

Wave is a beautiful company.

Just don't name it Ajar AI or Anthropomorphic :D

calebpApr 211

tylermjohnApr 238

I think this is a good thing to have in the toolkit and has been underleveraged in the past, so I'm glad you posted this. But imo the stronger considerations for most EAs are that they are likely a poor personal fit for for-profit work (especially given that prior experience is the biggest predictor of success) and capital incentives are very hard to align with most impactful aims.

calebpApr 234

I mostly agree with this. I wasn't really aiming to give a balanced take here on whether people should start for-profit instead of nonprofit - I just meant to list a few (imo) underrated by EAs features of for-profits.

I'm more confused about the poor personal fit point. I suspect that many EAs are also a bad fit (at least initially) for starting nonprofits, but the EA ecosystem makes it somewhat easier for nonprofits to survive (which imo is probably a good thing overall).

The version of your claim I most agree with is:
* An EAs comparative advantage is identifying the most overlooked and big deal moral considerations, and access to minimal viable nonprofit capital
* This selects for highly neglected things where you might be able to have an impact at a small scale, with not particularly high standards
* So even if EAs aren't sufficiently competent to build high-growth companies, they can still have an outsized impact via founding nonprofits

tylermjohnApr 233

I think this is true, but also having a successful for profit that achieves some of the goals you set out is an inherently narrower set of skills because you need to do market research, product market fit, customer relations, p/l, find ways to scale teams and products, etc. These are skills that need to be learned whereas for nonprofit work you can just do your research or whatever. Some of them involve a bunch of soft skills and types of scale/customer mindset I don't commonly see in EA.

calebpApr 232

Yeah, I think starting research nonprofits is an especially different skillset. Good point.

MiriApr 228

I find this post interesting, and probably largely agree with many of your points, but they don’t fully align with my experience or intuitions as a nonprofit entrepreneur.

I broadly agree with the Scalability and Capital Counterfactuals sections, though I think some of your concerns can be mitigated by the reality of planning and running a non-profit. Many funders at scale are not necessarily value-aligned (to EAs), and many non-profit founders (at least when going through the CE incubation program) start out with a clear view of possible endgames and payers at scale - these don´t assume that growing the organisation´s budget and staff as much as possible is the goal, and they don´t always assume that the charity will bear all costs for the entirety of the program forever.

I found the Learning Potential section so vague that my initial (defensive, sure) gut reaction was that it felt like motivated reasoning. While I don´t actually believe this was your intent, in my experience this applies to founding a non profit to a very large extent:

People often report learning a great deal from working at high-growth companies, building interesting connections, and gaining legible experience, which opens up interesting opportunities to do high-impact work.

I don´t have any way to evaluate which side is

less fast-paced and interesting.

Lastly, I find it hard to disentangle what the average for-profit and what the average non-profit is. For example, while the Standards argument might apply to a majority of all non-profits, in my (biased) opinion it is not representative of the average CE-incubated charity. Similarly, I would expect a great deal of for-profits to make things no one wants and burn a lot of capital in the process of going out of business, though that may not be a majority of the for-profits started by your average EA.

calebpApr 22*4

Similarly, I would expect a great deal of for-profits to make things no one wants and burn a lot of capital in the process of going out of business, though that may not be a majority of the for-profits started by your average EA.

There are strong reasons to believe that profitable companies are producing goods or services that people want (which is not to say they are morally worthwhile). In most cases, the end user will simply not pay if the service they provide is bad. This just isn't true in non-profits. I don't think being an EA changes the calculus very much on the for-profit side, but importantly for profits have a strong kill function.

Matt_SharpApr 217

Related - see Ambitious Impact's Founding to Give program and intro post here. I wonder how the first cohort are getting on...

calebpApr 216

Yeah, I think that program is cool. Although I'd probably be even more excited to see EAs participate in YC (or Entrepreneur First if they don't have a cofounder).

huwApr 2211

The plan for FTG is to get the founders into those programmes or to good seed funding, it’s trying to fill a different niche in the market which is value-aligned co-founder matching (harder than you think!)

Brad West🔸Apr 254

I think that another aspect to consider in starting a new organization, non-profit or for-profit, is how many of your deficits (organizational, research, mathematical, etc.) can be addressed or alleviated by AI tools today. I think historically there were a number of qualities the absence of which would make starting new things very difficult. I think using AI tools could dramatically lower the bar if you have good idea for a business or a nonprofit.

Chris LeongApr 224

A lot of these considerations feel more compelling if AI timelines are long, or at least not short (with capital being the one consideration going the other way).

calebpApr 222

I'm not sure why you wouldn't care about standards if timelines are short. I feel like you should care more about people actually using your work, which might be more likely if you are forced to get product-market fit.

Chris LeongApr 230

Let's look at what you wrote under this section and not just the headline.

"They are often difficult to evaluate and lack a natural kill function"

That seems to me like a longer term issue.

calebpApr 234

Ok, I did also write:
* Nonprofits have significantly weaker feedback mechanisms compared to for-profits
* Few people are going to complain that you provided bad service when it didn’t cost them anything.
* Most nonprofits are not very ambitious, despite having large moral ambitions.
* It’s challenging to find talented people willing to accept a substantial pay cut to work with you.
* For-profits are considerably more likely to create something that people actually want.

Which I think are all differentially more useful in short timelines than in long timelines worlds as you don't have less time to mess around, and you instead need to very quickly work on a useful thing. If you disagree with this maybe we're talking past other and I misunderstand your perspective.

Fwiw I'm sympathetic to nonprofits being better on net than for profit in short-timelines for reasons that aren't discussed in this post e.g. greater freedom to focus narrowly on useful work, particularly in cases where there isn't a viable business model.

Chris LeongApr 232

Points 1, 2 and 5: These all seem like variants of feedback being good. Seems like if timelines are short, you probably want to take a shot directly at the goal/what needs to be done^[1], even if the feedback mechanism isn't that good. If you don't take the shot, there's no guarantee that anyone else will. Whilst if timelines are longer, your risk tolerance will likely be lower and feedback mechanisms are one key way of reducing this.

Point 3: I expect a large proportion of this to be a founder selection effect.

Point 4: Seems to fall more under more capital which I already acknowledged as going the other way.

^{^}
I suppose this lines up with "greater freedom to focus narrowly on useful work" which you consider outside the scope of the original article, whilst I see this as directly tied to how much we care about feedback.

Robi RahmanApr 234

@Ben Kuhn has a great presentation on this topic. Relatedly, nonprofits have worse names: see org name bingo

quinnApr 233

In AI safety adoption is also a reason. I expect labs to adopt stuff that they pay for, whereas research / open source public goods they'd be like "oh that's cool" then never use it.

calebpApr 232

Fwiw, I am a fan of open-source public goods. I think more nonprofits should try to build products rather than papers (or at least be in that mindset). I am a little worried that doing research is just more accepted than producing products that people might use and that creates unhelpful barriers. Relatedly, I quite like the move towards essay websites more care goes towards thinking about the UX and trying to actually interact with users.

peterhartreeApr 223

On "Standards": I often link people to Jason Crawford's blog post "Organisational metabolism: Why anything that can be a for profit should he a for profit."

calebpApr 22*2

This is great, thanks for sharing. I hadn’t come across it, but it matches my intuitions well.

HelloHelloApr 221

What about a mix of both... a B2B company with EA clients! That's what we're doing. I believe that for-profits have different incentives and attract different kinds of people. I reckon our new agency, Rogue Impact, will be paid by money which ultimately comes from donors, but make that money go further by working faster and to a better quality than many nonprofit employees. I posted about it the other day - would appreciate some feedback on the idea!