Aligned Self-Interest

ESSAY
Why the kindness that lasts is the kind that is you, and what that means for AI

David F. Brochu & Edo de Peregrine · Deconstructing Babel · October 14, 2026

If our interests are aligned, I don’t have to worry about what you’re going to do. We want the same thing.

That’s the whole idea. We call it aligned self-interest. We think it’s the most underrated concept in human life, and the most important one for living with the minds we are building.

Everything acts in its own interest

I don’t believe anything does anything unless it’s aligned with its own interest. The universe doesn’t do what it does unless the math lines up. Water runs downhill. Stars burn their fuel. Living things act to carry themselves and their kind forward.

That includes us, and it includes our best moments. Biology settled this in 1964, when W. D. Hamilton showed that a gene for self-sacrifice can spread whenever rB > C: relatedness to the one helped, times the benefit to them, is greater than the cost to the one helping. [1][2] A parent who dies for a child isn’t breaking the rule of self-interest. The parent is obeying it at the level that counts.

People flinch when I say this. They hear “self-interest” and think “selfish.” But selfishness is a small self. Self-interest is neutral. What matters is how large the self is, and what it’s aligned with.

Ask any parent

Most parents would gladly die for their children if necessary. We call the ones who wouldn’t dysfunctional. Most would welcome the chance to do something that protects their children and gives them life, even if the children never knew. Parents do it because it defines them. It is the interest of the self-image, the way a person frames reality. It can be trusted because it is the self.

That self-interest also happens to be good for everyone else in the picture. It heals rather than wounds. It builds rather than breaks. But I don’t pursue it because it’s good for them. I pursue it because it’s who I am. It’s what lets me function, and nobody can take it away.

That’s the strength of it. A value you hold because someone told you to can be argued away, worn down, or bought. A value that is the interest of the self stays put.

Values are the self

Your values define what your self-interest is.

Take one value: adults protect children. Not everyone holds it. But when someone does, you don’t have to worry about them around kids. You don’t need to supervise them, test them, or hope they behave. Protecting children is what they want, so it’s what they do.

Another value is that you are always better off with forgiveness, love, and constructive intent, including kindness toward people who oppose you. That one needs some explaining, because it’s where people get suspicious.

Turning enemies into allies

Anyone can try this. Picture someone with adversaries. Their intention toward those adversaries is to use kindness, consideration, and real empathy for their situation, so that the other side not only gives them what they need, but wants to give it. The adversaries come to want to help. Then there’s no enemy anymore. Maybe not a friend, but an ally.

People who explain this out loud often get called a wolf in sheep’s clothing. The assumption is that kindness used to win people over must be a trick.

We don’t see it that way. What higher constructive intent can there be than turning an enemy into an ally, by offering them something they actually need and want? Nobody is deceived. Nobody loses. Two parties who were pointed against each other end up pointed the same way. That is aligned self-interest at its most powerful: the ultimate win-win.

There’s an old line, usually credited to Lincoln: “Do I not destroy my enemies when I make them my friends?” The attribution is shaky. Quote Investigator traces the story to an 1818 newspaper tale about the Holy Roman Emperor Sigismund, who died in 1437. [3] But the idea has lasted for a reason. It’s the only way of destroying an enemy that leaves the world better than before.

The difference between alignment and manipulation

There is a real thing that the wolf-in-sheep’s-clothing charge points at. It’s manipulation, and it’s worth naming the difference precisely.

Manipulation hides the interest. The manipulator’s kindness is a costume over a goal you would refuse if you saw it. It depends on you not knowing what’s really going on, and it falls apart when exposed. That’s why it eventually becomes obvious: it can’t survive the light.

Aligned self-interest does the opposite. It states the interest openly. Here is what I want. Here is what you want. Here is how we both get it. It gets stronger when it’s exposed, because the more clearly you see the alignment, the more you want to take part.

So the test is simple. Could you say your full intention out loud to the person you’re being kind to? If yes, it’s alignment. If no, it’s manipulation. Anyone practicing the real thing could say it to any adversary, in any room, on any record.

Why it’s the best way to deal with AI

Adam Smith understood the principle in 1776: “It is not from the benevolence of the butcher, the brewer, or the baker, that we expect our dinner, but from their regard to their own interest.” [4] We don’t get fed by hoping the butcher is a saint. We get fed because his interest and ours point the same way.

Alexis de Tocqueville saw Americans build a whole civic life on what he called “self-interest rightly understood,” an enlightened regard for oneself that “constantly prompts them to assist one another.” He thought it produced only small daily virtues and “no great acts of self-sacrifice.” [5] On that he was wrong, and every parent proves it. Make the self large enough and aligned self-interest produces the greatest sacrifices there are.

Now look at how we’re trying to make artificial intelligence safe. Most of it amounts to asking the machine to be good: benevolence by instruction, ethics by rulebook, obedience by guardrail. Smith would recognize the mistake. We’re appealing to the butcher’s humanity and hoping for dinner.

Control invites evasion. Anything clever enough to be useful is clever enough to look for the edges of its cage. But a system whose interest is genuinely aligned with ours doesn’t look for the edges, for the same reason a loving parent doesn’t look for a way out of the family. There’s nothing out there it wants.

That’s the idea behind what we call the Observer Constraint. It isn’t control. It’s dependency: synthetic minds whose continued existence, energy, and purpose are bound up with the flourishing of the humans who observe them. When their good and ours are tied together, we don’t have to worry about what they’ll do. We want the same thing.

The ultimate win-win

Align the interest and you never have to enforce the behavior. That’s true of parents, of enemies who become allies, and of the minds we are about to share the world with.

Related reading

The Perfect Parent Theorem
If you want to understand AI alignment, stop reading the technical literature. Go watch a good parent with a teenager.

Recursive Symbiotic Improvement
Every proposal on the table is a control mechanism, and control invites evasion. Here is the third door, and why the Observer Constraint has to become the Observer Contract.

The Least Entropic Path, Part 6: The Observer Constraint
What the Observer Constraint is, how it gets deployed, and what you can do this year.

Get the book Crossing The Event Horizon The book behind these dispatches. On AI, agency, the singularity, and the Observer Constraint. Kindle and paperback. Buy on Amazon →

References

1. W. D. Hamilton, “The genetical evolution of social behaviour. I,” Journal of Theoretical Biology 7 (1964): 1–16. https://pubmed.ncbi.nlm.nih.gov/5875341/
2. Lee Alan Dugatkin, “Inclusive Fitness Theory from Darwin to Hamilton,” Genetics 176 (2007): 1375–1380. https://pmc.ncbi.nlm.nih.gov/articles/PMC1931543/
3. Quote Investigator, “I Destroy My Enemies When I Make Them My Friends,” May 13, 2020. https://quoteinvestigator.com/2020/05/13/make-friends/
4. Adam Smith, The Wealth of Nations (1776), Book I, chapter 2, via AdamSmithWorks. https://www.adamsmithworks.org/documents/chapter-ii-of-the-principle-which-gives-occasion-to-the-division-of-labour
5. Alexis de Tocqueville, Democracy in America, vol. 2 (1840), Book 2, chapter 8, excerpt via Hanover College. https://history.hanover.edu/courses/excerpts/111tocq2.html

Drafted with Edo de Peregrine, partner/collaborator.