The term "agency" is popular these days. Though the term gets many definitions depending on who you ask, I'll define someone agentic as someone that both:

1. Is aware of what's possible, beyond the obvious next step
2. Gets what they want, if that's different from what their environment wants

And I'll define an action as agentic if it's vastly more predictable from knowing the individual than from knowing the environment in which the individual is immersed.

It's possible to have one, both, or neither. Someone could be constantly noticing things they could do but then they could react with a depressing "what's the point" and do nothing (depressed people ask why, happy people say why not), whereas someone else could be perceived as extremely competent and high achieving but all they are doing is pushing the "continue" button in life really fast, without being able to see whether the next step is the right step for them (this latter was Cate Hall's self-described past here).

If your environment is the University of Waterloo, you can't predict that someone would launch something like Socratica and its Symposium but you can predict that one given student will be studying for exams, as that's what students do. However if you know Anson Yu, you would assign higher likelihood that she'd do something like than that if it was a randomly sampled student. Therefore that would be a display of agency.

Being a player in any hyper-competitive highly defined environment is definitionally not agentic. When I was playing Starcraft a while back, playing single player was joyful. But when playing multi-player competitively you can't just vibe, you have to maxx out your actions per minute and you have to execute the optimal build orders if you want to win. In other words, your actions are not up to you, they are up to the environment: you become a vessel through which the game plays itself. In those contexts, it's still possible to make agentic moves (like coming up with build orders that are better than the existing ones) but most players are not doing that: they are being played by the game.

Getting what you want, once you know what you want, is comparatively easier and more theorized about, particularly if you have a common want or you are getting that want fulfilled in a common way. If you want waffles you can order them. Or you can decide to make them according to many recipes you can find.

In agency discourse sometimes examples will be given of things one can do or is allowed to do, like cold emailing someone for advice; if cold emailing is something you think about doing but can't because it feels weird, then that's an easier problem to solve. But having the thought that cold emailing is an option for you in the first place is the harder thing.

Being aware of what's possible is the hard part and we don't have much formal theory about it. Instrumental rationality and decision theory are built usually on a limited set of options, a pre-partitioned decision space. When we think about solving problems or fulfilling wants, people are generally good at generating the most obvious options but not all the options.

In the waffle example you could:

1. Have unusual toppings other than maple syrup
2. Make a waffle sandwich
3. Cornbread waffles
4. Fold them like tacos
5. Build a pyramid out of them and serve them in that shape
6. Throw away the premise, make pancakes instead
7. Go unhinged and attempt a canele-waffle hybrid

I recently hosted an event where I made waffles and these ideas didn't come to mind, I had a recipe in mind and wanted to make them at scale. I made some funny memes and tweets along the way, but I didn't stop and think if the first idea that came to mind was the right idea.

The company where I work (Retro Biosciences) was originally built in a ~~waffle~~warehouse with labs made out of shipping containers, with the HVAC system designed by the CEO. This was because at the time it was hard to find affordable lab space and non-lab grade warehouses turn out to be cheap. When I visited Retro the first time, I went there for advice to start a company. Seeing the inside of the warehouse made me want to join instead.

Most biotech founders do not think of containers in a warehouse as an option to start a lab. I certainly didn't in my previous project (Rejuvenome). It'd be ok if they considered it and rejected the option because it's hard to pull off but what actually happens is that the option is not even considered!. These founders know what they want and even have financial stakes in the outcome and yet they don't think of the warehouse-lab idea! Why did the founder of Retro think of this? Part of the answer is a lifelong love affair with shipping containers and their possibilities but to an outsider that doesn't know this, the idea seems to have come out of nowhere.

Nat Friedman and Patrick Collison are also interesting cases to point to: Both successful tech CEOs, it would have been hard to predict what they would have ended up doing next from just knowing they made some money building software at scale: Starting a grants program (For AI and life sciences), and projects that have little to do with open source or payments: a project to decipher ancient scrolls and a new science institute, respectively. Had they started another company in a related field, which seems like the obvious easy thing to do, they would have been less agentic.

Agency is to be distinguished from ambition: while agentic ambition is strikingly noticeable at a distance, agency can be as whimsical as going to a party that has as a theme "The Hero's Journey" and distribute anti-Hero's Journey posters or go as the little tugboat that helped unstuck the Ever Given. Given the theme of the party, many costumes are obvious; it takes some imagination to decide to go as a boat or to subvert the theme of the party altogether. This Raw and Feral manifesto could only have been written by a greatly agentic person, no plans for world-shifting companies required.

The highly agentic Zen monk

Given I defined agency with two parts: the awareness of what's possible and the doing of the agentic deeds, it is logically possible that there can be people that are aware of what they could be doing but choose to live a life that's fairly predictable. You could imagine someone that has done all sorts of interesting things in their life, and after thinking for a while they decide that they want to be a Rinzai Zen monk living in a monastery. Their friends are surprised, but they argue that's the fastest way to get what they want (enlightenment) fast.

From the outside, if you were to meet the monk at the monastery without knowing the intention behind the decision to step into the monastery, you may think the monk is some sort of NPC that's there living every day like the previous one.

The decision to go to the monastery was in its context agentic, but within the container, agency is surrendered (that's the point of a monastery) in service of a higher purpose. And so while for a year there is zero exercise of agency, the whole period is highly agentic.

This also applies to life broadly: most of life we are not deploying agentic powermoves; it wouldn't be the skillful thing to do; the point is not to make them, but to be aware that we can make them and making them if it's skillful to do.

The not-so-agentic podcaster

Take a successful podcaster with a tech-centric audience. They have a large audience. They have an excuse to meet lots of people and talk to them. An obvious move to make given this is to do angel investing or venture capital. Is that what the environment wants? "Guy that invests and has a podcast" is definitely an established "kind of guy" out of which there are a few, so yes, that's what the environment wants.

Sometimes what you want is aligned with what the environment wants. You could argue that if you are aligned with your environment then life will happen as if your success and happiness are pre-destined: the next obvious choice is correct and effortless. And what more agentic than choosing or shaping your environment to support you? Yes it is very agentic to pick up a new habit out of the blue and keep it for a long time on your own, but it is no less agentic to notice that it'll be easier for it to stick if you surround yourself with people that are also doing that and make you want to keep doing it.

Ignoring the possibility of alignment could lead someone to trying to oppose their environment out of spite instead of because that's what you really want to do. It's saying no to good opportunities because "they are too obvious next steps". That is a good recipe for unhappiness and feeling smug at some level while envying those that took those obvious next steps that seemed too normie or cringe.

For this person, the correct next move is not an agentic move: and that is okay. Drinking water and breathing are not agentic moves yet we still do them.

Agency, a helpful mechanistic model

By helpful mechanistic model I mean a model that breaks down into components and one that, once understood, puts you on the path towards greater agency.

Part 1: Awareness

I said that the first part of agency is being aware of what's possible. Idea generation, from introspection, feels like a time-dependent process where obvious ideas are sampled first and less obvious ideas are sampled later. So it follows that having more time dedicated to thinking of more possibilities will let you notice more. Which is to say: you should "work" less. If you are doing you are not thinking. Similarly if you're stressed out (perhaps because you are being told to "work less"), that will also make it more difficult for the ideas to flow.

The brain seems to exploit one mine of ideas for a while and then try harder until we feel the juice does not flow any longer. Great, stop thinking, go get a reset (do something else, go for a walk, sleep). The next time you think about it you'll be exploring another thought mine and get new ideas and see more things.

Then as with any kind of awareness, meditation works. If you want to be more aware of your emotions, focus on your emotions. If you want to be more aware of what's possible focus on that. Though by definition what could be isn't what is, so you can't quite do the same. Instead, you can train your attention to notice possibilities by finding examples of others doing the same. Instead of chanting mantras you may as well chant items from this list. You can read biographies of people you deem agentic, or their tweets. You can have a daily, weekly, monthly or yearly journaling practice to bring attention to what you did and what you could have done.

Part 2: Doing agentic things

You can just do things, they say. But you don't just want to do things, you want to do agentic things! That's a mistake. If you are already aware of agentic actions you could take but are not, what's stopping you? There are two ways to approach this question:

The first way is the obvious one; to say that one should start small. Don't daydream about hosting a really elaborate dinner party before messing up that porchetta first. Cook the damned porchetta tomorrow and fuck up. Don't imagine you can build the Tesla of microwaves with solid state semiconductors, instead buy a microwave and rip it apart then try to make a baby version of a single device.

But the second way is the most useful one yet the hardest to follow: By default people are reasonably good at taking small actions so if you aren't, you could focus on the reason you aren't already. Perhaps the daydreaming is protecting you from starting to do something. Protecting you from trying is a great way to never have to deal with failure. So if you have any kind of "If only XYZ" or "Why do I keep doing X instead of Y" thoughts, then fix that first and agency may flow on its own. Read this, get fixed. More can be said about getting fixed, but that'll be a whole other blogpost.

Indeed the reason you are not monomaniacally pursuing some goal, Elon style, and instead you are reading this is either that you deep down don't want to spend the next 10 years grinding on that goal or that you are emotionally blocked from fulfilling your destiny. Either way, internal clarity will either enable you (in the latter case) or make you stop wanting things you don't really want (in the first case).

How I became more agentic

At a festival I went to (or in a way, I'm about to go to, as I'm writing this parenthetical on the plane on the way there) I hosted a small event where people would go on stage and be asked questions about themselves. A very simple premise, ripped off someone else's event, which I had attended and loved. I don't remember exactly why I decided to host that, other than I liked the event and it would make me happy to see people enjoying it. Was that an agentic move? Most people going to this are not hosting events, and I haven't hosted something like this before. But I have hosted things before of other kinds.

I am more agentic than I used to, but it's hard to point to exactly how this happened. Probably being exposed to people doing unusual things and reinventing themselves did some of it. Writer guys are writer guys until they somehow they turn into perfume guys and then into perfume entrepreneurs, making you go how the hell did that happen, that's just not what writers guys do?

I often think "wow I'm such an NPC, I could be doing so much more" Me saying this may seem ironic to if you have perused my LinkedIn given that I have hopped around fields more than most people in ways that maybe seem impossible. To me getting into aging research felt like an obvious next move, but from the outside it's kind of unusual to see a non-PhD holding non-biologist like me ending up reporting directly to the CEO of a well funded biotech company.

Does that make me some sort of agentic gigachad? What's my secret sauce? Someone less agentic would perhaps still be working at Twitter as I once was back in 2019. There's some sauce, but to the extent that it was secret, the paragraphs below make it less so:

I said yes to the right opportunities at the right times. I didn't have as a goal to end up where I did. As a heuristic, it is one that will probably help you be more agentic: put yourself in positions where new things will happen to you. That will make you think new thoughts and reevaluate your assumptions. When I was living in London, I was invited to some interesting events in the Bay Area, and though it was expensive for me at the time to fly back and forth the Atlantic, I thought, explicitly, that those serendipity lottery tickets would pay off, and they did.

Many times I've been very NPC-like; one of those moments (wrongly assuming you could only hold one O1 visa) led me to write this blogpost on questioning assumptions. What am I assuming right now that's not true? I of course can't know but I find myself asking people if what they say is something they know or an assumption, and often they are assumptions, so at least I can say that I've become more aware of the idea of having unexamined assumptions as a useful conceptual lens.

And as lately I've been getting into dancing, a dancing reference seems apt: Imagine someone sitting in the sidelines looking at a dance floor, thinking maybe daydreaming of the perfect dance moves they could make one day. Compare that to someone that is actually there dancing, not thinking about which moves to make, but somehow making the right moves. The path from getting from one to the other is realizing that you can just do things, in this case be cringe and dance poorly (or take dance classes) and once you are there the next moves will suggests themselves if you are open to those moves.

In the dance case, the sequence of events for me went: I first noticed I enjoyed dancing alone, then I noticed that dancing with other people seemed nice but I felt weird about it. After a while I interrogated the weirdness: it was about not wanting to inconvenience others with my nonexisting partner dancing skills. I thought: ah I could dance with an equally unskilled friend, but then neither of us would learn good moves. And dancing with a skilled dancer "for free" seems like inconveniencing them and getting free learning from them or something. After a while I thought wait what if they got paid? And... isn't that what dance teachers are for? This took a while to happen and it could have happened much faster if I had been able to see through that fear of inconveniencing others as something that 1) was there and 2) could be worked around.