“We have seen that one of the chief causes of the danger of disastrous consequences in the use of the learning machine is that man and machine operate on two distinct time scales, so that the machine is much faster than man and the two do not gear together without serious difficulties. Problems of the same sort arise whenever two control operators on very different time scales act together, irrespective of which system is the faster and which system is the slower.
This leaves us the much more directly moral question: What are the moral problems when man as an individual operates in connection with the controlled process of a much slower time scale.”
The above quote is from a piece called Some Moral and Technical Consequences of Automation from 1960. It covers many AI adjacent topics and broadly discusses the idea of disparity of time scales between two different individuals.1
When I read things like this, I am reminded once again that perhaps the main thing to think about in life is the idea of what time horizon you’re operating on and what you think is valid to navigate through the dark forest to.
Put another way, people have different certainty requirements at different time horizons and every consequential decision requires us to decide how much of the future must be clear before we are willing to act.
The Dark Forests of Technology & Startups
In technology, we think about this on a spectrum. It ranges from “move fast and break things”, a short-termist view that the immediate ability to create change is what matters in order to make breakthroughs, and that the fast changes aren’t consequential enough to slow down for, to “choose good quests” which look at the self and the impact of the quest in our lifetimes, all the way to longer duration post-individual goals, perhaps like Elon’s view of going to Mars or something.2
Importantly, I believe longer time horizons don’t necessarily mean higher certainty requirements. You can have immense conviction in a distant outcome of a Good Quest for example, while knowing almost nothing about the mid-term path to get to the end of the quest, and it can still be the right thing to do.
These ideologies impact how people think about company building and legitimacy building and even humanity, as best shown today via how many are operating on perhaps the shortest-term time horizon in the history of technology, due to just how close and dark the “dark forest” of technological progression truly is (and the resulting existentialism of this).
The Dark Forests of Politics
We can also look at this through the lens of government, policy, and politics which have their own strange disparity of time scales.
Governments make decisions whose consequences can compound for decades, politicians operate on election cycles of a few years, and the people making those decisions operate on the finite horizon of their own lives. These decisions are also made with extreme uncertainty and varying belief systems of what is or is not existentially important for citizens and a country.
In this essay which discusses how to get big things done in government one of the core ideas is that meaningful reform often has immediate costs but benefits that take years to materialize. Move too slowly and the political consensus can collapse before anyone gets to see whether the underlying policy actually worked. In other words, the feedback loop of politics may operate faster than the feedback loop of policy.
This also makes me think about age limits in politics.3
I find it disturbing that many of the consequential choices being made with many cascading effects for decades, are enacted by politicians that will be long gone before they ever see these impacts.4 There is a narrow window (of decades) where wisdom intersects with the reality of having to live with the outcomes of your morals and wisdom. I’d like politicians to serve in that window and not after.
The Dark Forests of Investing & Love
It is naturally, also worth thinking about this when it comes to investing.
In venture, a core criticism of Compound has always been “how do you know what will happen” and really the underlying question being “why do you bother to try to know?”
People basically think some form of : you will be wrong about so many things, this is false confidence, false understanding, false precision, false realities that you are building upon each other and thus you shouldn’t try to model it nor spend the time doing so. Instead, you should focus all of your time on understanding networks, founder psychology, etc.
Things that are “single shot” and feel durable, not that are some layered multi-step system with tons of uncertainty and seemingly endless acres of poorly lit paths in the forest.
This is, in my mind, an unwillingness to navigate two dark forests very deeply: The first, “what will happen in the world such that a company today will be valuable in 8-12 years” and the second “Will your process create good investments, make you a good investor, etc.”
A lot of investing (life?) is about wandering, and perhaps staking your career and income on how well you understand the world creates far more of an emotional toll that rolls up directly to who you are as a person/investor and your intellect instead of “we bet on the wrong person, it happens.” which is strangely far easier to stomach despite just being wrong on a different axis. The power law is comforting in many ways.
“For the individual scientist… [this] requires an imaginative forward glance at history which is difficult, exacting, and only limitedly achievable.” - Some Moral and Technical Consequences of Automation (again)
In public markets, there is another criticism which is that the forest is darker for you versus others: “How can you possibly know about this company with greater precision than someone who stares at it at all the time.”
A brief romantic detour.
In life we think about this concept of navigating the dark forest and disparity of timescales in how we discuss family values or just how we think about what must be figured out, what should be vaguely figured out, and what we will figure out over time with our partner or loved ones.
So many relationships hinge/falter on some sense of alignment in what parts of the dark forest are reasonable to leave dark versus not before commitment.
What kind of parent do you want to be? What do you want your career to look like? Where should we live? What kind of childhood should we give our kids? How will our marriage be? etc. etc.
An endless list of questions that are proxies for love, for compatibility, for the life we are so certain we want to live for the next 50+ years.
Some are naively thought to be answerable and we look to these long checklists to check off before we get comfort with a person and a future life. Others are naively thought to not matter, and we look up only to realize the excitement or feelings of love might not be enough relative to the alignment that was far less durable than we could have ever thought.
Ok back to public markets.
Markets are romantic relationships of buyers and sellers.5 There are a set of conditions that we all believe about a given company based on some set of beliefs (these might be called “facts” or “data” but they also might be narratives or projections). It is the job of a great investor to look at what are the areas of the dark forest they believe matters to navigate, and on what time horizon they are willing to navigate it at.
As you go further out on the time horizon, the surface area for disagreement and variant view changes.
Quarter by quarter? Very shallow/small dark forest and you must have very precise beliefs that are better than others.
Year by Year? Decade by Decade? I might be willing to theorize that the year 4 part of the dark forest will matter so much more than years 1 and 2 that it doesn’t actually matter if we know all of the things other investors do who are staring at a single company all day, every day, perhaps missing the forest for the trees.
I will spare you from an onslaught of the many cliche examples but technology is littered with these so instead a hypothetical conversation:
“Look, it’s sweet that you think it’s really important that we align on sending our kid to french elementary school before we get married, and I can understand why you think that’s important, but I’m pretty confident that no matter what, our kid will be great and so I don’t think we should worry about deciding now or really even try to know about the possible schools they might be going to.”
…fast forward 7 years and our child (Amazon Web Services) is a great, well adjusted kid that we love very much (projecting $1T/year in revenue in the next few years), and none of this precision about elementary school (e-commerce GMV) mattered to figure out.6
My sense is this concept of knowing what matters to know in investing is even more important at a time in which we feel obligated to utilize all of the ever-growing tools and information at our disposal in order to seek comfort in our edge, when in reality it likely should be used to make sure our imagination is truly unique. That we can indeed think out of distribution for our time horizon.
To quote Wiener for a third time:
“We must always exert the full strength of our imagination to examine where the full use of our new modalities may lead us.”
To me, the important part of this is not that imagination eventually illuminates the whole path but that it can allow us to decide which ideas or parts of our imagination we need to parse before moving forward. How much certainty do we require before we keep walking in the darkness? And are we willing to try to find our way, with some stumbles I’m sure, in the half-light, to get to a future that we believe in.
In investing and in life, as we learn more, we also learn that the forest is always dark.
In the modern interpretation of the essay’s case, man and machine.
I guess some exhausting people are thinking of post-humanism too.
Which we execute on horribly
There is a ridiculous, but maybe worthwhile argument for why some parts of government should be forced to have teenagers or young adults in the loop or something. I haven’t thought much about this but am always reminded of Rudolf’s essay discussing youth involvement in politics wrt AI: “There are many AI youth activists (in the sense of specialising in the topic of AI, being AIs, or both), who try to use their position to advance youth interests. The problem is that they don't have concrete policy asks beyond "allocate more money to us", which puts them at odds with every other interest group in society, many of which (e.g. the retired) outnumber them in raw numbers as a voting bloc as well as in terms of resources, power, and influence.”
Is this not one of the worst and most cringe sentences I’ve ever written?
I am so sorry for this writing, really.
>How much certainty do we require before we keep walking in the darkness?
As a recent grad trying to figure out what my next steps are, this was a great metaphor/reminder. The path doesn't need to be fully illuminated in order for me to start walking!