Incomplete Thoughts
Human beings are excellent at beginning to think and horrible at completing their thoughts. Admittedly, this is not their fault, as evolution has encouraged reaction times intended for self-preservation rather than intellectual rigor. The genetic material that pondered the risk-reward ratios associated with fleeing predators probably fled ineffectively.
The benefit of thinking fast, rather than slow1 is evident - we are capable of making decisions without the paralysis of an overly analytical, rational mind. The cost is subtler, from our conversational patterns to the technology we develop to aid those interactions, we all but ensure that we never think complete thoughts.
Interrupting Thoughts
Long before the creation of text messages and doomscrolling, there were “thought-terminating cliches.” In the absence of anything particularly insightful to add to a conversation, awkward pauses would ensue. To address the risk of being perceived as dull or foolish, we created broadly-applicable cliches about teaching old dogs new tricks or that one is never too old to learn new things2 or any number of other platitudes that derail conversations. No one argues with the cliche, but neither does the train of thought continue to its previously-scheduled destination.
Sometimes thoughts are stopped by the stealthy appeal to authority that is a “thought-terminating reference.” This is the well-read academic who inquires “have you read Leviathan?” or notes that “you really should read Foucault’s discussion of the panopticon.” While a formal education has value, the requirement that one must consume its canon before the discussion can proceed does not, and again, curtails the thinking of the individual missing some relevant credential or reading. Essentially, rather than exploring an original thought, we’re asking the speaker to observe similarities between their idea and those written in books we barely remember rather than simply letting them complete their own thought.
The less pretentious form of this occurs via “thought-terminating confirmation.” This is the simple assessment of whether someone has access to the same reference that lives in your psyche. Taking the form of “you’ve seen the Seinfeld episode about festivus, right?” or “wait, you get my joke about my desk being made of rich mahogany?” Two long-time conversational partners would deploy shared references to accelerate the mutual exploration of an idea but two less-common conversational partners will break their own thought processes clarifying the other’s points of reference.
And of course, there are the “thought-terminating needs” that range from “I could use a snack” to “I’m tired” to “this is a decent essay, but I really need to pee.”3
For these and so many other failings of our humanity and our technology, most thoughts never reach their full potential in the form of their deserved conclusion.
So why does it matter?
Business and Philosophy
Often, we do not know where a thought will lead when we open our mouths to speak. Perhaps some extraordinary, species-advancing insight lingers just beyond the horizon of a line of thought never fully explored. We learn about opportunity costs in econ 101. Rarely, do we consider the opportunity cost of a good thought that we never complete.
Business are founded with a minimal sense of the commercial and aspirational thoughts they will ultimately represent. And most fail before those thoughts are completed.
Currently, among those in the worlds of finance and technology, one such thought incurred a staggering interruption. The idea of effective altruism began with the thought that time and money are scarce resources, and one should think rationally and carefully about how to deploy those resources to achieve the most possible good.
This led to discussions about longtermism, the relative importance of future quality-adjusted-life-years, and whether the potential to save more expected years of life by a slight diminution of the probability of AGI misalignment than by helping feed a hungry family tonight is morally justified.
Some rightly questioned the absence of a limiting principle or noted that the most visible proponents of this type of decision calculus were fortunate enough to enjoy comparatively comfortable lives in the near-term. And others, in contrast, noted that those with comfortable lives were also those with excess capital to donate, that perhaps you might maximize the life you can save by earning a ton of money and giving most of it away. Some debated what we owe the future, while others focused on optimizing for more immediate concerns.
This was a useful debate for a society continuing to grapple with the consequences and possibilities of its technological progress.
And then that thought was interrupted.
SBF
An affluent scion of academic royalty founded a cryptocurrency exchange (FTX), amassed a thirteen-figure fortune, and promised to give 99% of it away. He became the most famous and most prolific effective altruist on the planet.
He donated to political candidates, appeared on podcasts and magazine covers, and became a celebrity without even getting a haircut.
Then it turned out the collateral backing his exchange’s holdings was primarily a set of tokens (FTT) created by said exchange, the exchange was created so that his crypto hedge fund (Alameda) would have a better place to trade, and a backdoor in the former was channeling funds to the latter.
Yadda yadda yadda, run on the bank, billions of dollars vanish,4 the disgraced altruist starts texting and tweeting in a manner that conveys the insincerity of his previous assertions. We begin to ponder the end of effective altruism.
Except now, instead of continuing the incredibly productive society discussion about the relative merits of altruistic wealth vs. social inequality, current vs. future suffering, and how to build a better world, we’re getting ready to toss the baby out with the bong water.
When a scientist is discovered to have faked their data, we consider the scientist a fraud, but we do not indict the scientific method. Why? Because to do so would prevent the completion of any number of “thoughts” in the form of ongoing research.
Society requires difficult discussions and complex thoughts.
Society
Society improves as time elapses. Optimism is rational. But that doesn’t mean the process avoids mess and difficult moral discourse.
Hobbes5 would argue that, absent the absolute morals imposed by a sovereign, our society would crumble.6 Then, the utilitarians begin to challenge those moral absolutes with thought experiments about pushing fat people in front of trolleys to save a larger number of lives to question how absolute those moral limits may be. And then folks challenge the utilitarians by arguing that the long-term prognosis of a society lacking some norms7 is horrifying.
It’s hard. It’s messy. The folks with easy answers are either pointing to the worst sociopaths in human history, pontificating from the comfort of an ivory tower, or the type of arrogant jerk who folks avoid assiduously at parties.
You want to live in a modern society with smartphones and electric cars? Me too. Get comfortable with difficult questions, and in the process, do the hard work of trying to complete thoughts.
That means recognizing the interruptions and building external structures8 and internal temperaments that support complex, multi-faceted ideas.
Completion
Some thoughts, like the one you were having before you clicked on this essay, will never complete. Conversations will continue to be derailed by cliche, confirmation, and that double-espresso shot this morning. We will continue to research BCI technology that might one day capture states of mind for future use.
We will root for businesses to complete the thoughts their founders envisioned9, and ideally, the ideas their team members develop along the way.
We hope that the moral and ethical inquiries the effective altruism community posed will be further discussed rather than merely another on the list of interrupted thoughts.
And we’ll recognize that modern society and its morality will be messy. But one place to begin is to follow odd thoughts and conversational meanderings to their logical conclusions.
Whenever possible, do not aspire to have the most clever thoughts, nor the wisest thoughts, nor even the most appropriate thoughts, but rather, for complete thoughts.
1 Kahneman and Tversky won a Nobel Prize for their work and wrote a best-seller asserting not only cognitive biases, but the evolutionary rationales for why they exist.
2 It’s worth noting that these cliches point to wholly opposite conclusions, and yet both are equally well-accepted. Both convince their recipients, or at a minimum, mollify them and truncate their thinking.
3 The number of meetings I have attended where complete thoughts ceased to be even plausible because the morning cup of coffee was bearing down on my lower intestines is larger than I care to consider.
4 One could argue, in some strange sense, he might have done more damage than Madoff. To wit, Madoff didn’t steal $50B+. He received something like $17B in deposits, lied about increasing its value over decades, then came up something like $57B short when the Ponzi scheme collapsed. FTX had an $18B valuation in July 2021 and its collapse will probably suck value out of the balance sheets of any number of crypto-focused entities (even those not directly connected to FTX.
5 Yes, I know that I mentioned Hobbes as an example of interrupting thoughts by requiring prior reading just a few paragraphs ago. These footnotes should enhance thoughts rather than interrupt them, but if you find them to be detracting from the broader message, don’t read ‘em. I won’t be offended.
6 Comments about lives becoming “solitary, nasty, brutish, and short” come to mind.
7 Maybe we don’t need the moral absolutes, but we’d probably prefer a society that frowns on killing and maiming even if some optimization algorithm suggests it’s a net positive.
8 Ideally, these external structures are extensions of our thoughts, allowing us to be present and focused (completing our thoughts) without interruption. Currently, we’re limited by our inability to multi-tread, constraints of short-term memory, and so on. Perhaps BCI will ultimately offer the remedy to many of these issues.
9 AE being one such thought.