Should you marry her?

Weigh the costs against the benefits

Marriage is a contract: you promise not to break up with your partner even if you later end up wanting to, and they promise the same.¹ A wedding is a contract enforcement mechanism: by standing up in front of your community and declaring that you won’t leave each other, you make it harder to back out later. If you break up for frivolous or fixable reasons, your community will judge you. If you go to your friends for advice about your relationship struggles, they’ll be more oriented toward helping you make it work and less likely to suggest you leave.² Your acquaintances will be less likely to try to tempt you away from your spouse,³ partly because others in your community will think they’re shitty if they do.

The main benefit of marriage is investment.⁴ When both of you are much more confident that the other will stick around for the long haul,⁵ you are incentivized to make investments in a joint life that wouldn’t otherwise have been worth it. This could look like buying a house or having a kid, dividing labor like life admin and fixing things and cooking to make joint life more efficient, or putting in the work to understand each other better and meet each other’s needs better.

Beyond that, I’ve found the emotional security of commitment creates room for a more radical trust and love. You allow yourself to sink into each other and extend deep roots in each other’s souls.

The main cost, assuming you plan to get married eventually,⁶ is that you might have found someone better. So how do you decide whether to commit to your current partner? By popular request, I’ve made up a brief guide. While this advice is gender-neutral, women generally know what they want and are ready to commit sooner than men, so it’s addressed to the confused rat boys in my life.

Once your girlfriend is ready to get married, you have three basic choices at any given point in time:

• Commit to her now: pay the cost of forgone options, get the benefit of investment.

• Break up with her now: pay the cost of searching for someone new, get the benefit of a different (and hopefully better) marriage starting in a few years.

• Wait and see: pay the cost of delaying the start of marriage (and thus the accumulation of investment benefits), get the benefit of more information to make a better decision.

The first question is how long you should wait and see. A rough model is that you learn one more unit of information about your compatibility with your partner from every doubling of the amount of time you spend with them. My husband and I dated for about ~2.5 years before getting engaged, and we spent maybe 15 hours a week actively hanging out together over that period.⁷ So we’d already spent ~2000 full-time-equivalent hours together, or ~11 doublings starting from our first hour. A fifty year marriage would have been ~40,000 more hours, only another ~4-5 doublings on top of what we had already experienced (and as I’ll discuss below our marriage will probably be a fair bit shorter than that).

In general, if you’ve been with someone for a couple of years already, there’s not much point hanging around longer generically waiting for more information. You should have some understanding of what exactly you’re hoping to learn in another six months,⁸ and pay attention to make sure you’re actually learning something about that.

The second question is how your girlfriend compares to other women you could marry. Your wife will fulfill three main roles in your life:

• Cofounder: Your wife will build a life with you. At a minimum, you’ll handle household maintenance and life admin together. Beyond that, what exactly you build together will vary — my marriage has revolved around working together to have impact, many are oriented around raising children, some are about creating rich shared experiences — so what you need from a cofounder will vary. But broadly speaking, you’ll want to feel like she’s competent and on top of her shit, that she’s bringing a lot to the table and your skillsets complement each other well,⁹ and that you communicate well and resolve differences constructively.

• Friend: You’ll hang out with your wife more than you hang out with any other person. You’ll want to be able to laugh easily together, have a lot to talk about with each other, and do things together that you both enjoy.

• Lover: Your wife will be the primary outlet for your romantic and sexual drives. You’ll want to find her charming and hot, you’ll want to be in love with her.

You probably have more evidence lodged in your brain than you appreciate, and some focused and organized introspection can probably move you. For example:

• For the cofounder role, how excited would you be to literally cofound an organization with your girlfriend? Would you be excited to hire her?¹⁰ Would you be excited to work for her? How would she stack up among people you’ve worked closely with? What would her manager and coworkers say about her? Do you trust the way she makes important decisions? Would you trust her advice about your important decisions? What does it feel like to make decisions together and divide up the tasks of life together? If something happens to you, like you get hit by a car or your parent has a stroke and needs a lot of care, does it feel like a relief that she’s around to take care of things? Do you have similar aspirations for the life you want to build and a plan to achieve it together — if you want kids, are you on the same page about how much work each of you is going to put into childcare and do you think she’d be a good mother?

• For the friend role, how much would you hang out with your girlfriend if it had to just be platonic (where would she rank in the rotation of friends you see every now and then)? What interests and joys do you share? What are the points of connection that seem rarest and most precious? What are the interests and perspectives you don’t share with her that you really wish you did?

• For the lover role, how attracted were you to your girlfriend in the first twenty minutes of speaking to her, and how often have you had feelings of attraction that strong (and what’s the denominator — how many women have you spoken to for 20 minutes)? How’s the sex in this relationship compared to other sex you’ve had in your life? How strong is your felt sense of love and affection compared to what you felt in previous longer-term relationships?

• Make a big list of all the women you can think of, whether they’re single or not, and ask your friends to come up with names you might not have known. Based on what you know about them, think about how you think they’d do at each of these roles compared to your girlfriend. For the ones that seem like they would plausibly be better overall than your girlfriend, ask yourself how often that kind of woman tends to come onto the market,¹¹ and ask an honest friend if she’s out of your league. Go to parties and meet some of them, and see how it feels to be around them. Try to quickly falsify some of your hypotheses about them.

The last question is how short is life. How much do you value the next few years of your life, when you could either be married to your current partner or on the prowl for a new one, compared to the future span of your marriage?

Given AI timelines, I think the expected length of your marriage if you get married today is probably something like ten years.¹² Even if you want to set that aside,¹³ I’m skeptical you should think of the remainder of your natural life as worth fifty or sixty times as much as the next year. You probably have some intrinsic time preference (you-in-fifty-years is much less “you” than you-next-year) and you should discount the future to some extent because of uncertainty.¹⁴

I’ve tended to think of a normal lifespan as being worth about thirty times as much as next year. That means if it takes three years to find another partner, you’ve spent 10% of the rest of your life searching. This, combined with the investment benefit — which I think could easily be 10% or more over a ten year marriage — means that you need to find someone who’s better than your current partner by a pretty healthy margin to be worth the wait.

I admit I have a conflict of interest in writing this post. I love marriage (I wish I’d gotten married earlier) and want to fit in another wedding or two before the Singularity. And I feel a sense of sisterhood with your girlfriend — I’m not sure you should marry her, but for her sake, I think you can figure it out soon.

1. This is not an absolute promise, and nor should it be. You can and should get divorced if it becomes miserable to be with your partner over a sustained period. But you can’t break up just because you met someone better or otherwise start to think you can find someone better (for example, because you became higher status, or your partner became less attractive, or loses their job). And most importantly, you should stop investing mental energy into the question of whether you could find someone better.

2. And if you do end up getting a divorce, they’ll also know to treat this as a really big deal compared to a normal breakup, something you’re going to grieve over for years.

3. This also applies to secondary partners if you're poly. A poly friend said that being married made dating other people lower-stakes, because everyone was clearer about where they stood in the partner pecking order.

4. There is also an insurance benefit — if you become a much less attractive partner one day, you’ll get to lock in love and affection from someone who would now be way out of your league. But I think this is a smaller benefit in practice, and getting married primarily for the insurance benefit is not a healthy attitude to marriage. You mostly want marriage to make you invest more in becoming a better partner for your spouse, not let yourself go.

5. Just how long a haul that might be is an important question which I’ll get to in a minute.

6. If you don’t really mind being single forever, you can afford to have very high standards, which means a lot of this guide won’t apply.

7. This was unusually intensive, because we lived together for most of the time we were dating and half of that was during the COVID lockdown.

8. For example, maybe you need to have a lot of long conversations sorting out where you stand on kids, like my husband and I did. Or maybe your experience in the relationship so far has been seriously unrepresentative — e.g. you’ve been long distance for a while and just moved to the same city, one or both of you made a major career or lifestyle change, etc.

9. Complementarity is especially valuable if you want to have kids. They are a huge amount of work and it’s a really big deal if your partner is up for doing the lion’s share of that work.

10. You’ll have to modify this somewhat if you do something super technical that she doesn’t have expertise in. But even so, the best cofounders can add a lot in nontechnical roles.

11. In general, the kinds of women who have been married longer tend to be in higher demand and come onto the dating market more rarely.

12. Expectations are longer than medians, which I see as closer to five or six years for top-human-expert dominating AI.

13. For example, maybe you want to make a commitment that you’ll stay with your partner for a subjective period of time equivalent to one ordinary pre-Singularity marriage (e.g. fifty years), even if the Singularity does happen in between. I would strongly discourage you from making commitments that last much longer than that.

14. You shouldn't come into marriage with this attitude, but a relevant point for thinking about the appropriate discount is that not all marriages last fifty or sixty years.