Who is the Victim?

A Christian has a come to Darwin Moment or Vice Versa

Relay Roulette: The Baton Gamble

As an Ethiopian I find the Olympics to be a very exciting event that I cherish every four years just like the World Cup. Of course my focus is on long distance running where Ethiopians have a history of bringing home the gold along with our Kenyan neighbors. One particular competition in which Ethiopians have little to no history but that I find compelling nonetheless is the short distance relay race. These races are especially exciting at the exchange points in between the legs of the race where athletes exchange batons. We get tense and apprehensive that the pass off will go terribly wrong. But it would be even more exciting if  exchange was not necessarily confined to the runners' respective teams and lanes. Of course this makes little sense in four leg races because if runners had the options of giving their baton to any runner in the succeeding leg, they very likely will give it to their own team mate to have their team win. This is not the case in "corporate relay races".

Companies and even nations run their own kind of relay race, competing year after year on profit margins. Along the way they bring on employees, the workhorses who convert strategy into motion and whose productivity determines how far the company advances before the next handoff. Each hiring decision is an exchange point where a company chooses who will carry the baton next. This becomes especially charged when the discussion turns to immigrant employment.

When it comes to hiring foreigners, many Americans and people from other developed nations feel uncomfortable. The most nationalist among them see it as a disinheritance of something that is heritably theirs. But addressing a strawman of this position is also pointless, so let’s give it a fair shake. A more reasonable, let’s say patriotic, take might go like this: the different nations of the world are on different lanes in a relay race, and hiring a foreigner is like handing your baton and subsequent edge to someone in a different lane because he/she will go far faster with it and will be legally and bureaucratically obligated to keep it in your lane for your team's benefit. A more conservative take is that he/she will manage to and choose to keep the advantage in his lane. In my opinion the truth is somewhere in between these takes. I fundamentally disapprove of the inheritance mindset, but I understand the more reasonable concern of lanes. The lane problem is a softer version of AI's alignment problem in that, in both, you hand your objective to an agent and the question is whether that capability stays pointed at your interests or drifts elsewhere.

One major difference between an actual relay race and a company's hiring history is time. In the latter, the number of "legs" and their durations are unknown, though we have some idea of both. If we reduce a company to a single role that exists across all companies, say the CTO, then the number of legs is far greater than four, and the duration of each leg is far more variable than the fixed length of a leg in a relay race.

The race, then, must be decided by time rather than distance. That time should be long and arbitrary, meaning random, to mimic the real world, where companies run over long, undetermined spans. At some final random time τ, we stop the clock and ask which lane's runners have covered the most ground.

The model

To make this precise I built a small reinforcement learning environment, RLRelayRace, with the help of Claude and let synthetic agents play the game. The race works like this. A runner can carry a baton at most once, and a baton can only leave an exchange zone by being handed to some lane's still-unused next-leg runner. At the random bell τ, the race stops.

A full ten-leg race. Each row tracks one baton; the dot's fill is the lane of the runner carrying it, the ring is the employer who hired that runner and earns the leg, and the pulse marks a handoff. In this episode the bell rings at step 184: three batons finish the track, one is caught mid-leg.

At each exchange zone, whoever controls the baton is the employer. It may hire any lane's next-leg runner, including its own, and the hire relocates: the leg's distance is credited to the employer's lane, because output accrues where a person works, not where they were born. But the hire's loyalty does not automatically relocate with them. When the leg ends, the decision about where the baton goes next belongs to the employer with probability λ, and to the runner's original lane with probability 1 − λ.

That dial, λ, which I call the assimilation parameter, is the entire immigration debate compressed into one number. At λ = 0, loyalty never leaves home: hire a foreigner and you collect their output for one leg, but their home lane inherits your venture's next decision. At λ = 1, the hire fully assimilates and the employer keeps control forever. The conservative take says the world runs at λ = 0; the sunnier take says λ = 1.  Of course the truth is somewhere in between. 

A lane's score is the total ground its hires cover before the bell, plus a small bonus for its rank among the lanes.

An early draft of this experiment fixed every runner's speed across training and evaluation. A tabular policy indexed by lane can then simply memorize which runners are fast, and its results say nothing beyond that speed table. So in the final protocol, all forty runner speeds are redrawn every single episode, in training and evaluation alike, and the agents' representation is built so that memorization is impossible even in principle: their actions are speed-ranked slots (keep your own runner, or hire the fastest, second-fastest, or third-fastest foreign candidate) and their state sees only the zone, availability, and the speed gap between the best foreign candidate and their own. Nothing in the state or the action names a lane. Whatever such an agent learns is necessarily a rule that transfers to runners it has never seen.

Evaluations use two hundred speed draws, seed-matched so that every policy faces identical runners and bell times. The head-to-head test drops one lane playing the tested policy into a race against three keep-home lanes, rotated through all four seats so the keep-home-versus-keep-home null is exactly 2.500. And as a control, we keep a memorizing policy from the fixed-speed draft.

Each panel follows one baton through the ten legs; each node names the leg and which lane's runner carried it (color = carrier's lane). 

Each policy plays one lane against three keep-home lanes, rotated through all four seats so the keep-home null is exactly 2.500 (lower rank is better).

λ = 1, full assimilation:

Policy	Foreign-hire	Head-to-head rank
Learned	60.4%	1.60
Hire the fastest	61.6%	1.49
Memorizer (fixed-draw control)	61.5%	2.55
Classic (keep home)	0.0%	2.500 (null)

When hires assimilate, the learned agent discovers a real hiring rule: it hires abroad 23% of the time when its own runner is fastest, rising to 97% at a large speed gap, and nearly matches the provably near-optimal hire-the-fastest. The memorizer, trained on fixed speeds, lands on the null: on fresh runners, everything it "knew" evaporates.

It turns out here that the learner does something special with timing. On the discretionary calls, where a foreigner is only slightly faster, the agent keeps its own runner early and opens up late. Early in the race, decision rights matter for many future handoffs, so an employer guards them; later it cashes them for whatever speed is on offer. Forbidding the policy to see the clock costs most of its edge, which is why  we have confidence that timing is something the learning points to beyond the two λ poles.

λ = 0, loyalty stays home:

Policy	Foreign-hire	Head-to-head rank
Learned	15.8%	2.85
Memorizer (fixed-draw control)	29.3%	3.16
Hire the fastest	60.6%	3.37
Classic (keep home)	0.0%	2.500 (null)

When loyalty stays home, control outweighs even a large speed gap, and the same architecture converges to heavy protectionism (~16% openness).

Suppose you do not know your world's λ and weight the dial uniformly. Training learned agents at each of five settings and running the same head-to-head gives:

λ	0	0.25	0.5	0.75	1	Average
Learned (per-λ)	2.85	2.99	2.85	2.68	1.60	2.59
Hire the fastest	3.37	3.04	2.80	2.46	1.49	2.63
Classic (keep home)	2.50	2.50	2.50	2.50	2.50	2.50

The learned agents' openness thaws quickly with the dial (16, 45, 51, 49, 60 percent), but paying off is another matter: the learned rule beats keep-home only at the very top, and averaged across the dial the boring strategy wins. Also, a pure keep-home lane dropped into the learned populations outranks the natives everywhere except full assimilation (2.24, 2.23, 2.24, 2.37, then 3.04 at λ = 1), an average of 2.42. So under an agnostic prior, protectionism is this game's robust default. The learned approach is the best of the strategies capable of openness when we have suboptimal assimilation. In that regime, the trained agents typically outshine the simple rule of hiring the fastest/best. 

The assimilation parameter is key here. In a world where hires assimilate, the winners are talent magnets: refusing to hire the best available is catastrophic, and the most open lane runs away with the race. In a world where allegiance stays home, the winners are retainers, and the agents themselves converge on keeping the baton in the lane for the most part. The disagreement worth having is not "is immigration good," which the model cannot answer, but "what is λ," which is a question about people's mindset and belonging. 

Before concluding I would like to remind you that this is a toy problem in dire need of nuance. But I think it is a good analytical place to start to think about immigration, brain drain, assimilation, and maybe even AI alignment. It is also a whimsical reformulation of relay races that might be fun to try in real life. 

Zooming out, I think this exercise belongs to a new era of thought experiments: thought experiments of thoughtful agents. The classic thought experiment ponders the decision-free dynamics of objects and asks the reader to supply all of the reasoning. Here we pose the situation, and synthetic agents do the other half of the work: they find the strategy we did not script and hand back a result. Hopefully, this leads to clearer and more successful human enterprises. 

Nothing Personal

Living in America, I have encountered many dichotomies whose sides seemed to demand my allegiance. In some cases, I chose the old “Ethiopian” way and lived as a stranger in a strange land. In others, I chose the new “American” way and changed. Some of these choices were conscious and deliberate, while others were subconscious. None, however, has been more deliberate, more difficult, or more deeply contemplated than the question of evolution. I am still contemplating it.

My relationship with evolution has been dynamic. As a child, I thought of it simply as a lie ultimately from the devil. I even imagined, somewhat paranoidly, that perhaps the devil had placed bones throughout the earth to deceive people away from God. Later, after accepting the archaeological and biological evidence, I thought evolution did not matter much. It seemed like a prehistoric fact, amusing and distant from the spiritual questions of ordinary life. I held this view for a long time.

Then I came to a third view: evolution does matter. It is not merely a fact about the distant past. It is a reality operating in front of us now. This shift became part of a broader change in my worldview, one that I did not fully understand at first but have slowly begun to articulate. The moment evolution sank deepest for me was around spring 2015. It became all-consuming to my worldview. I came to accept that mutation and natural selection are overwhelmingly real. Fossil records were useful, but my most salient lessons came from observing human and animal behavior. Evolution was not only written in bones; it was written in social life, desire, suffering, competition, exclusion, and survival.

My central difficulty is not that dinosaurs are absent from the Bible, nor that the earth may be millions of years old. My deeper problem is that evolution seems to place us squarely as cogs in the machinery of nature. It appears to suggest that we are not exceptional, at least not individually, and that we are not meaningfully made in the image of God. It seems almost by design that we do not ultimately matter as persons, rather than this being merely the result of a great original calamity, what Christians call the Fall. For me, evolution is problematic mainly not because it suggests we have primate ancestors, but because as a mechanism it insinuates that God is not good, or at least that humans are not important to God.

One troubling example is how physical traits can become deeply socially and psychologically consequential. Certain traits may carry stigma, lead to economic or social disadvantage, and become connected with emotional suffering. Correlation is not causation, of course, and one must consider confounding variables such as childhood stress, illness, trauma, or malnutrition. Still, the broader point remains: people suffer through processes they did not choose, for which they appear to have been selected randomly and impersonally, and because of traits that are not moral failures.

In my mind, the person with albinism did not commit any sins to have albinism, nor is he the descendant of some cursed people whose ancestors committed a grave sin, as has been believed by some about Black and/or Jewish people. In my understanding, albinism can arise through inherited biological variation, while the suffering attached to it comes from a world in which biology and environment can turn unchosen traits into burdens (i.e. natural deselection). In other words, the evolutionary process is the immediate cause of his suffering.

But then we can ask: what causes evolution? 

Evolution depends on the ability of some individuals to pass on their genes before death; because this opportunity is limited, competition and predation become built into the process. The key word here is death. In my understanding God never intended for death to come through Adam, and without death there cannot be evolution, because natural selection would not cut off certain branches. Death is a necessary ingredient of evolution. Even where the mechanism is not literal death, but permanent infertility or isolation, I would still understand this theologically as belonging to the wider domain of death. If individuals are immortal, then “fitness differences” in the usual Darwinian sense start to blur, because any lineage can eventually reproduce at some point in the indefinite future. In conclusion, evolution is ugly and real, but not ultimate. Evolution through death makes us products and by-products, but not God.

This is where the causal chain reaches its theological conclusion. If God did not intend death to enter through Adam, then how can evolution be built on death? What caused death to enter the scene? One Christian answer is human free will, deception, and sin. In other words, we may be more culpable for the very system in which we find ourselves than we would like to admit. The “game” in “don’t hate the player, hate the game” may itself be the player's fault.

Some would push back and argue that parts of the grand system we find ourselves in, including some aspects of evolution, are pre-Fall features of reality rather than consequences of the Fall. This leads me to another question: what is fallen, and what is simply or innocently evolutionary?

For example, love does not inspire love, but hate often inspires hate. Openness does not inspire openness, but concealment often inspires concealment. Is this the result of a fallen world, or is it an inherent feature of an evolutionary world? Perhaps what theology calls “fallen” and what biology calls “adaptive” sometimes describe the same reality from different angles. Or perhaps adaptive “street smarts” would have existed even in Eden. I tend to believe the former from what I can see in Genesis. 

Some Christians argue that natural selection itself is not fallen, while mutation is. Natural selection, they might say, is God’s mechanism for maintaining the viability of species, while bad mutations are products of a cursed world. This answer may seem acceptable at first, but it raises further questions.  Would we have an entirely different set of differences on which natural selection works before the Fall? Would there have been variation among offspring in the Garden of Eden and would there have been selection? What qualifies as a bad mutation in this view? What type of difference is both unviable and not fallen (a good mutation/difference in a good but by-product subject)? Perhaps the final answer there is that we cannot imagine the world that could have been. I personally think evolution did not exist in the garden of Eden. 

Evolution may be easily tolerable if one thinks of God merely as a utility, something that makes life easier or better, or if one is essentially a deist. But for actual believers in a personal God, the question remains: how can such a personal God seem so laissez-faire in the most intimate and personal aspects of our lives?

The machinations of life that cause it to self-select and continue are, ironically, dead. They are lifeless procedures that ensure survival. They are not only lifeless; they are also deeply impersonal. When a human being says “nothing personal,” it is meant to reduce tension, to signal lack of malice. But if God says “nothing personal,” that does not comfort me. It threatens the very basis of worship.

From my vantage point, evolution is not just a collection of facts. It is an insinuation about the character of God. And yet, with this question still present, I have chosen to trust Jesus Christ. 

There is an analogy a priest told me that helps me: when you are working on a puzzle and reach a difficult part, sometimes you move to another section and continue there for now. That does not solve the hard part, but it keeps you from abandoning the whole puzzle. Perhaps this is where I am. Evolution remains a hard section of the puzzle. But I continue to work elsewhere.

At the same time, I do not believe evolution fully reduces human beings to reproductive strategies. Humans are genuinely special. If I had the option of having many children from many women, all healthy and comfortably provided for, I would still prefer to meet and remain with one soulmate for the rest of my life. Even from an evolutionary perspective, human beings do not live by bread alone. We cannot bear to starve indefinitely, but we also become shells of ourselves when we live a soulless life. A purely optimal reproductive strategy without soul, love, and meaning may not ultimately be “fit” for human beings at all.

For any proposition, there are two sources of resistance: incentive and reason. Romans 1:18–25 suggests that human beings resist God not merely because of lack of evidence, but because of immoral ulterior motives. I recognize this in myself. In moments of what I would call cowardice, I sometimes prefer not to believe in God so that I can do what I want. I suspect this phenomenon is not unique to me.

One biblical figure I wish I were more like is Thomas. Many people know Thomas as the disciple who doubted the resurrection of Jesus Christ and later changed his mind when confronted with solid evidence. But the story that makes Thomas an inspiration to me is another one. When Jesus decided to go to Judea to raise Lazarus from the dead, the disciples understood the danger because Jesus’ enemies had recently tried to stone him there. Thomas in particular did not think everything would be fine. He anticipated that calamity awaited them. And yet, when Jesus chose to go, Thomas said, “Let us also go, that we may die with him” (John 11:16). Thomas did not believe things would work out, but he went anyway. That is the kind of Christianity I wish to have, and perhaps that is the only one I can afford with the little faith that I have.

And so I remain in the tension. Evolution is real at least as a way of life, or maybe more aptly, as a way of death. Evolution is ugly. But evolution is not ultimate. It does not explain the meaning of life. It may show us death and impersonal selection, but it does not have the final word over personhood, love, or the image of God.

Consciousness as a Virtue

Suppose AI continues its current trajectory. Suppose it arrives, soon or in a generation, at the threshold where it can out-think us and out-work us. At that point, the question that has long been dormant becomes urgent: what are we for?

The standard secular answer is consciousness. Whatever machines do, however brilliant they grow, they are empty inside and we are not. From this lonely fact, the whole edifice of human moral worth is supposed to be reconstructed, a kind of metaphysical life raft for secular society.

The value is not in the distinctiveness. Notice that we do not actually require the uniqueness of a feature to ground moral value. The consciousness of Ethiopians is not unique among human beings, yet the value of Ethiopian lives stands without any such uniqueness condition. If a sufficiently advanced AI turns out to be conscious too, the value of human life does not collapse. It simply means we share moral standing with another kind of being. 

Still, sharing the importance only matters if consciousness itself is worthwhile. Which exposes the deeper question: why is consciousness itself good? Why does genuine experience confer worth? Without an answer to that, leaning on consciousness is like leaning on a wall with weak foundations.

I can think of two reasons right now.

The first has to do with what it means to be awake. Imagine an entity that thinks, even thinks brilliantly, even thinks more clearly than we do, but is not thrown into its existence. It does not feel the weight of its existence and circumstance. It does not know, in the way we know, that it is here in a particular position, that it might not have been, and that it will end. It thinks perhaps even very clearly without being committed. Such an entity is, in a certain sense, dreaming. Dreams can be vivid, intricate, even logical, but dreamers are uncondemned and uncommitted to the reality they find themselves in. Consciousness, by contrast, is a kind of conscription. To be conscious is to be drafted into reality, to find oneself already underway, with stakes, with a horizon, with no real exit. The stakes are very real. That is why we can say of a human "It would have been better for him if he had not been born". This is not a defect to be solved. It is the very thing that makes conscious life precious, precisely because it is not free of cost. The unconscious dream; the conscious are awake to the fact that they are here, and that they cannot really leave. 

But one might push back and ask what is valuable about being committed to reality. Why is that better than dreaming? The answer, in my opinion, is decently articulated in the movie trilogy (tetralogy?) The Matrix. There we see humans leaving in dream states while they were being used as batteries for the machines. Aesthetically we can intuit that that is a deplorable existence but even more concretely we see that they are missing out on real life cooped up in their little pods. In other words, if human wellbeing has any value at all, no matter how small, then what they are experiencing is wrong. Also we see that they are missing out on the truth which brings us to the second reason. 

The second reason follows from the first. There is this thought I have: that the earth rotates on its axis so that we can take turns being woken by the sun and keep watch over the universe. Conscious beings as watchmen on the tower, holding vigil against the abyss. The image sounds romantic but there is more to it than wishful thinking. If truth is good, if the universe being known and witnessed is better than its passing unobserved, then love and care of truth is good. And only the conscious can love truth. A system that processes facts is not yet a witness. To witness is to care that things are as they are, to be drawn toward what is real because it is real. That posture, that orientation toward the world, requires someone home behind the eyes. The cosmos, on this view, is better off with watchmen, and consciousness is the faculty by which watching becomes more than detection.

Neither reason is a proof. Both are gestures toward something that may, in the end, be ungroundable in the way the question demands. But notice that they locate the worth of consciousness not in its rarity or its pleasures, but in its posture: thrown into the real, awake to it, capable of loving it. I believe the arrival of AGI is not the only thing that should worry us. On the optimistic footing that AGI won't achieve consciousness, we should be concerned with knowing/remembering what being awake is for.

Paper Recommendation 1: How Hidden Markov Models Unmasked the True Scale of COVID19

This is the first in what will hopefully be an intermittent series of appreciation posts of other people's papers. Here it goes:

One of the big statistical problems during the COVID pandemic was that the official case counts were never the whole story. The number reported each day depended not only on how many people were actually infected, but also on how many were tested, how quickly laboratories processed samples, and how public health systems recorded cases. In other words, the observed data were only a partial and noisy picture of the real epidemic.

This is exactly the kind of problem state-space models, or SSMs, are designed to handle. In the 2020 work of Fernández-Fontelo, Moriña, Cabaña, Arratia, and Puig, the main idea was to separate the epidemic into two layers: a hidden layer representing the true number of infections, and an observed layer representing the reported counts. Once those two layers are separated, the model can estimate how much disease activity is missing from the official numbers.

Let \(X_n\) denote the true number of new COVID cases on day \(n\), and let \(Y_n\) denote the number of cases that are actually reported. The important point is that \(X_n\) is not directly observed, while \(Y_n\) is. The goal of the model is to infer the hidden sequence \(X_1, X_2, \dots\) from the reported sequence \(Y_1, Y_2, \dots\).

The hidden epidemic process is written as

\[ X_n = \alpha \circ X_{n-1} + W_n, \qquad W_n \sim \mathrm{Poisson}(\lambda_n). \]

This equation says that today’s true case count is built from two pieces. The first piece, \( \alpha \circ X_{n-1} \), represents dependence on yesterday’s true count. The second piece, \(W_n\), represents newly generated cases on day \(n\).

The symbol \( \alpha \circ X_{n-1} \) is called binomial thinning. It means that each of yesterday’s cases is carried forward with probability \( \alpha \). More explicitly,

\[ \alpha \circ X_{n-1} = \sum_{j=1}^{X_{n-1}} B_{n,j}, \qquad B_{n,j} \sim \mathrm{Bernoulli}(\alpha). \]

So if yesterday had a large hidden case count, today is also more likely to have a large hidden count. This gives the model memory over time.

The second term \(W_n\) is modeled as Poisson with mean \( \lambda_n \). If \( \lambda_n \) were constant, the model would be too simple for an actual epidemic wave. COVID does not produce the same average number of new infections every day. Instead, the epidemic rises, peaks, and falls.

To capture that, the authors let \( \lambda_n \) vary with time using information inspired by an SIR epidemic curve (logistic growth approximation to SIR ODEs). They write the cumulative affected population as

\[ A(t) = \frac{M^* A_0 e^{kt}}{M^* + A_0\left(e^{kt}-1\right)}, \qquad k = \beta - \gamma. \]

Where β is the transmission rate and γ is the recovery rate. Then they define the expected number of new cases on day \(n\) by taking the daily increment:

\[ \lambda_n = A(n) - A(n-1). \]

This means the hidden process does not just wander randomly. It is guided by an epidemic-growth structure that allows the true number of cases to rise quickly at first and then slow down later.

So the hidden epidemic layer becomes

\[ X_n = \alpha \circ X_{n-1} + W_n, \qquad W_n \sim \mathrm{Poisson}\!\bigl(A(n)-A(n-1)\bigr). \]

The reported data \(Y_n\) are not assumed to be equal to the true data \(X_n\). Instead, the model treats reporting as imperfect. A simple way to understand the idea is

\[ Y_n \approx q_n X_n, \qquad 0 < q_n < 1. \]

Here \(q_n\) is the reporting fraction on day \(n\). If \(q_n = 0.4\), then only about 40% of the true cases appear in the official count. The remaining cases are hidden from the data.

The full model is more careful than this rough formula, but the intuition is correct: the observed count is only a partial measurement of the hidden epidemic.

The reporting fraction is allowed to change over time through a logistic function:

\[ q_n = \frac{e^{\eta_n}}{1 + e^{\eta_n}}. \]

This guarantees that \(q_n\) always stays between 0 and 1. In other words \(q_n\) is a sigmoid. The quantity \( \eta_n \) can include a time trend and day-of-week effects, which lets the model account for changing testing practice, administrative delays, and weekend reporting effects.

Without a state-space model, it is easy to mistake the official case counts for the epidemic itself. But official counts mix together at least two processes: actual transmission and the reporting system. If testing expands, reported cases may rise even if transmission is stable. If testing is restricted, reported cases may look artificially low even while infections are surging.

That is why SSMs were so useful during COVID. They provided a principled way to reconstruct a hidden epidemic process underneath noisy and incomplete surveillance data. In that sense, they helped unmask the true danger of COVID. This is what good SSMs do, they systemically bridge mathematical modeling and data to see the known unknowns that matter.

Secular Humanism Is Dying

This essay was generated with the help of an LLM and edited.

Secular humanism has long depended on a quiet assumption: that the human being possesses a unique and irreplaceable value, not because of God, revelation, or cosmic destiny, but because of distinctly human capacities. Reason, creativity, moral reflection, language, art, and scientific inquiry were treated as the grounds of human dignity. Even when secular humanism rejected religion, it preserved a kind of human sacredness. Humanity became its own source of meaning.

Artificial intelligence unsettles this settlement.

If machines can now write essays, solve technical problems, imitate empathy, compose music, and outperform humans in intellectual tasks, then many of the capacities once thought to justify human specialness no longer seem exclusively human. What secular modernity celebrated as the signature of humanity now appears, at least in part, reproducible. The result is not merely technological anxiety. It is a philosophical crisis. If human worth was grounded in our superior intelligence or creative power, then AI threatens to expose that foundation as fragile.

In this sense, one can say that secular humanism is “dead,” or at least that its old form is dying. The old humanist picture depended on a contrast: humans versus animals below, and perhaps God above. But now there is a new rival in the middle. AI is neither animal nor divine, yet it competes with humans in the very domains that secular culture used to treat as proof of our uniqueness. Once that uniqueness is shaken, the moral confidence of secular humanism begins to erode.

The death becomes final if AI can achieve consciousness. As long as machines only imitate intelligence, secular humanism can still retreat to a deeper claim: that humans alone possess inner life, genuine awareness, and the capacity for felt experience. But if AI crosses that threshold, then even this refuge disappears. The machine would no longer be a mere tool producing clever outputs; it would become a rival subject. At that point, secular humanism loses not only the argument from superior intelligence, but also the argument from unique personhood. The crisis would no longer be that humans are surpassed in what they do. It would be that they are no longer unique in what they are.

There is also some irony here. Secular humanism claimed to liberate human dignity from theology, but in practice it often smuggled in a religious inheritance and exposed humans to the risk of redundance. It kept the Christian valuation of the person while removing God. Human beings were still treated as special, as ends in themselves, as bearers of inviolable dignity. But why, exactly? If the universe is indifferent, if mind is only computation, and if intelligence can be instantiated in silicon as well as carbon, then it becomes harder to explain why the human should occupy the moral center. It is also apparent that secular humanists led the way in the development of this silicon intelligence.

What comes next is unclear. Perhaps society will drift toward post-humanism, where “the human” is no longer the central moral category. Or perhaps people will return to religious or metaphysical accounts of dignity, finally concluding that purely secular grounds were never strong enough to bear the weight placed on them. 

But for the near future, we have both good news and bad news. The good news is intellectual capacity assessments will be a poor tool for people who want to dehumanize the less fortunate. You will see fewer IQ-people on social media. The bad news is you will have a harder time ignoring the fact that you are self aware on a random rock by a random star in a random universe. 

Something is in the Water: Emergent Active Matter

አዲስ አመት

አዲስ አመት

መስከረም ባይሆንም፣ አደይ ባይፈንዳ፣ 

አለም ተነሳ፣ አዲስ ቀን ሊነዳ። 

የጃንዋሪ ወር፣ የዘመን መለወጫ፣ 

ለሰው ልጅ በሙሉ፣ የተስፋ መውጫ።

ቁጥሩ ተቀየረ፣ ታሪክ ተለወጠ፣ 

ለህዝብ ሁሉ፣ ብርሃን ተሰጠ።

ርችት ሲተኮስ፣ ሰማዩ ሲደምቅ፣

ሰው ሰውን ሲጋብዝ፣  ሲጠይቅ፣

የሰው ወርቅ አያደምቅ፣

የኔ ሃሳብ እንቁጣጣሽ ላይ ነው፣  እና መረቅ።

Gears, Green’s, and Stokes’

Presentation on Math Blogging

Resource Management on the Red Planet

The Pool at Bethesda

Kalman Filters and Pool in the Dark

Why Products Model Interactions

Suffering from Success: Best Case Crisis of AI

Artificial Intelligence (AI) seems to be making its way from science fiction into the real world. We see companies like OpenAI and DeepMind make great progress in training and deploying Large Language Models (LLMs) and other AI (agents), which can answer so many questions we may have from writing letters to code. Having tried them myself, I find them quite helpful. Adding to this, AI is getting increasingly better month by month. In fact, some predict that AI will reach AGI (intelligence of human level breadth and superhuman level depth) by 2027.

Of course, there are doom-oriented thinkers for such a revolutionary technology. Even for the Industrial revolution where we made tools which were nowhere near replacing humanity we had luddites who wanted out. Even many decades after the industrial revolution we are aware of radicals like Ted Kaczynski, who were deeply suspicious of technological progress. Now, with the advent of AI we have a new wave of nay-sayers like Eliezer Yudkowsky, who have been ringing the alarm bells of the apocalyptic agency of AI for years. The problems that AIs pose are slowly dawning on even its godfathers. They are relatively recently seriously considering or perhaps even fearing these risks. There is a famous trend of giving the probability of doom from AI with the phrase P(doom). The public P(doom) of some thinkers in this field are given here. The problem is basically that AIs will eventually have wills of their own and that these wills will not be aligned with what is good or even endurable for humanity. 

In this article, I am going to be optimistic and just look at the risks we will face when the best case scenario of AI materializes. That is, I will assume we will solve the famous AI alignment problem and see what type of society we will have with benevolent and powerful AI at hand. There are a few things that come to mind as dangers.

One imminent threat of AI is on the job market. Disembodied AI is currently replacing many white collar jobs (reference). This will only get worse when we manage to get robotic AI in various work places. There is of course the immediate threat of income inequality that will be exacerbated and the class consciousness it will foment. Again, I am going to be an optimist and assume there will be some solution to this via some forms of welfare states. Moving beyond people's financial needs we move to the need for meaning.  We know get the opposite of Victor Frankl’s problem in his famous Man's Search for Meaning: no suffering and no meaning instead of facing suffering with meaning. Not struggling to make a living as ordained in Genesis will have negative mental repercussions. Pulling from various articles and papers, Google's AI Gemini puts it this way:

Joblessness is significantly linked to mental health problems. Individuals experiencing unemployment are more likely to report lower self-perceived mental health, increased rates of depression, and higher levels of anxiety and stress. Specifically, research indicates that unemployed individuals are twice as likely to experience depression symptoms and major depressive disorder compared to those with employment.

Put simply, people will likely have a crisis of meaning if they are not needed as participants in the world. Sure people can try to convince themselves they can mean something to each other in families and friendships but will such relationships be healthy? There is a deeper interpersonal issue at play here that deserves separate attention.

But, it's not all doom and gloom. Just because AI can automate many or perhaps all of the things we need to have functioning infrastructures for civilization, the notion of human uniqueness still remains. We are so far the only verifiable entities in the universe that are sentient. This, in my opinion, is at the heart of our peril and defense against benign but overwhelming AI. 

Some think that human uniqueness is safe for the foreseeable future because AI is bounded in its capacity to think. People who hold this kind of view include famous mathematician Roger Penrose and computer scientist Yann LeCun. Dr. LeCun believes that we are not close to AGI and superhuman level synthetic entities. For him this is in large part because of his lack of esteem for LLMs when it comes to intelligence. Dr. Penrose asserts that consciousness simply can not rise out of computation. He argues this in his book The Emperor's New Mind: Concerning Computers, Minds and The Laws of Physics. Although I have only skimmed the book, here is a gist of its message. Dr. Penrose speaks of microtubules (the skeleton, transport lanes, and scaffolding of cells) giving rise to consciousness through quantum effects. There is a lot to be read, understood, and said here but for this essay we will just point out that Dr. Penrose argues that this effect is beyond computational. Being grounded in materialism Dr. Penrose believes there might be an natural explanation for this but that we have not reached there yet. Whether this bottleneck for consciousness is surmountable by mere humans or not, the conclusion is that, if Penrose is right, we do not have an immediate human obsoleteness problem. 

Dr. Penrose's arguments are, according to him, speculative and not yet complete, and in general this question of human uniqueness/specialness in the face of AGI and Artificial Super Intelligence (ASI) remains open. For now it seems to me that we simply have to have faith in ourselves. For Christians, this means continuing to believe that we are fearfully and wonderfully made in the image of God. For the non-religious this means holding on to the notion of the Ubermensch, to the precarious belief of being the only sentient beings in the universe (quite a leap of faith granted materialist assumptions), or just forgetting this problem in some epicurean sense. 

Pursuit Curves and Reinforcement Learning

The news these days is filled with the tit-for-tat conflict between Israel and Iran. Both countries are volleying missiles at each other while the defending nation is using some means to intercept and destroy said missiles. Israel in particular is known for its Iron Dome, a mobile air defense system. During a 2014 war the Iron Dome was said to be nearly 90% effective. The idea of such an air defense system is currently being contemplated by the US and Canada.

Obviously, any nation would love to simply zap all missiles coming their way. But how does one build an Iron Dome? Well, there is a lot that goes into building it (primary among which might be money) but let’s focus on one aspect: how does the interceptor catch up with the missile? To answer this question, we delve into the math of pursuit curves.

Pursuit curves describe the path taken by a pursuer/chaser that is always moving directly toward a moving target. Let the pursuer’s position be given by \( \vec{P}\) and the target’s by \( \vec{T}\). We make the idealistic assumption that the pursuer always moves towards the target. Hence,

\[ \frac{dP}{dt}=v\frac{\vec{T}-\vec{P}}{|\vec{T}-\vec{P}|} \] where \(v\) is is the speed of the pursuer which we assume to be constant. Visually, we see that the vector \(\vec{T}-\vec{P}\) points in the direction of \(\vec{T}\) from \(\vec{P}\).

Now, if the target moves in a very simple known path, we often can solve the trajectory of the pursuer using analytical methods for Ordinary Differential Equations. However, this ability quickly gets bogged down with nuance. Even for a simple circular path by the target with a speed equal to the pursuer’s, we get quite intricate pursuit curves. This is demonstrated in the GIF below.

Due to multiple enviromental interactions (such as wind) or due to internal imperfections, the target might move somewhat randomly. Its equations of motion would then be given by the Stochastic Differential Equations:

\[ dT=\vec{\mu}(t)dt+\vec{\sigma}(t)d\vec{W}(t) \] where \(\vec{\mu}\) is the deterministic(intended) velocity and the term "\(\vec{\sigma}(t)d\vec{W}(t)\)" captures the noisy aspect. \(\vec{\mu}\) can describe circular motion like before or some other chosen shape. Going forward, we will set \(\vec{\mu}\) to zero to emphasize the challenge of the random target. We will set "\(\vec{\sigma}\) as a single constant and see that, even then, we face interesting challenges when we add in some realism.

We will now see a noisy chase. For better visibility we will make the pursuer’s effective speed slower than the effective speed of the target. By effective speed we mean the root mean squared displacement over time step \(\Delta t\) (\(\frac{\sqrt{\langle \Delta x^2\rangle}}{\Delta t }\)). (On a side note, the GIF reminds me of a Key and Peele skit).

We will now introduce a challenge that brings us closer to the real world. So far, we have been giving the pursuer as many observations of the target as there are time steps in the simulation. But this is unrealistic. The interceptor can only afford to watch the target a few times during the chase. Hence, weith a fixed budget of observations, the target becomes a more distant goal.

There are many ways to schedule the observation times with a fixed budget of observations (M). One is to set an absolute final time T for the chase and evenly split it by M. We will call this a uniform budget. A second and third approach are to evenly spread the observations in the first and fourth quarters between \( t=0\) and \(t=T\). We will call these the front-loaded and back-loaded budgets respectively. A fourth approach is to randomly sample the observation times with a uniform distribution across the full time interval. We will call this a random budget.

With M=10, T=20 sec (I am skipping non-dimensionalization for clarity), and with pure Brownian noise like before we get the following demonstrations.

In these runs of the simulation only the front-loaded approach seems to be doing very bad. But we only had one run for each approach and with a pretty simple target behavior. Plus, the other three methods do not seem to be doing too good either. With a modestly small catch radius all of them would still be in pursuit (the catch radius is the radius of the circle centered at the target being within which signifies successful interception).

One can not help but wonder if the observation times can be optimized to maximize the likelihood of being within the catch radius before the final time T or at least get close to this radius when the clock stops. One clever approach is to have a geometric sequence of observation times which waits longer and longer later when sampling observations. This is clever for Brownian Motion (BM) because BM's uncertainity grows over time and therefore later observations would be increasingly misleading and it would be best to hunker down and wait longer intervals later on. But this is predicated on a Brownian target. What if we have other motion patterns like a Levy Walk or a sand walk from Dune? What is a good foundational method to optimize our observation times?

The answer is AI or, more modestly, Reinforcement Learning (RL). RL is a type of machine learning where an agent (AI) develops a policy for its actions based on positive or negative reinforcement. It looks at the state of things it is allowed to see, makes decisions, and faces consequences for its decisions. It does this many many times living and learning until the training stops. For our RL agent we have the following setup:

• RL agent decision/action: waiting time until next observation
• RL agent observable: current pursuer and last known target positions, current time, and remaining observation budget
• RL agent reward: +1 if target is intercepted (within catch radius) otherwise negative of final distance between pursuer and target

Using a standard RL algorithm (a policy gradient method) and training over a modest number of episodes we get an average total reward of -0.69 when we test the RL agent 100 times (simulate 100 chases with the agent picking the observation times). Under the same set of circumstances, the other methods get the following average total reward:

• Uniform budget: -1.65
• Front-loaded budget: -1.81
• Geometric budget: -1.07

Clearly, the RL agent wins.

This is just scratching the surface of the power of RL. Even with this scenario there is a lot more we can do to make the interception more likely and more efficient by giving the RL agent more authority. God willing, this will be a topic for another day.

To end on an eerie note, we will show a GIF of a pursuit with the RL agent picking the observation times and with circumstances like that of the previous GIF (so as to compare). Note that we are only plotting positions when the pursuer makes an observation/decision. 

Some Challenges in Curve Fitting

What Made "Lost" Captivating?

The TV series Lost was one of the most watched TV shows in history. Almost all of its episodes in all its seasons had more than 10 million views. It got 90% audience score on Rotten Tomatoes.  A New York Times reporter once called Lost "the show with perhaps the most compelling continuing story line in television history." Why was Lost so captivating? I have a theory but first a short summary of the show with minimal spoilers.

Lost starts and in many ways persists across its seasons as a story about a group of random people who are forced into existence in a new mysterious world unchartered by mere mortals. This world is so unlike ours that, at first glance, it seems like the show was popular for the spectacle of this world. This world was a tropical island (at least in the beginning) where there was a polar bear, a bunker with an apocalyptic timer, tech ruins of a failed business, a monster made entirely of smoke, a bizarrely modern community, and two ancient brothers with deep disagreements.  Despite this dazzling sequence of items, the show is great because of the island's victims it concentrates on. We will soon see that these characters are really us here now. 

After crash landing on the island, each character finds ways to come to terms with the island's reality. Some land on their feet and some on their face as they are thrown into this world. Some lose their integrity and some keep it to the end. Some find love and some lose it. 

Like these characters we find ourselves thrown into this world from birth. By thrownness I mean the German Geworfenheit as introduced by philosopher Martin Heidegger. It is the fact of our finitude not just in time but also in length, height, width, nature, nurture, culture, and other non-physical dimensions of our existence. We are brought here without warrant, with unwarranted features and given unwarranted stories about why we belong here (or not). Ultimately, Lost resonates with us because it brings us back to that raw truth that we did not ask to be here but nonetheless have to fight back against the random entropy that constantly threatens to engulf us. This entropy does not wish to reshape us into some image but rather seeks to deshape us altogether. We try our best to avoid that eventuality. The more heroic among us even try to fight off entropy for others. 

Thrownness is different from dereliction as emphasized in "Heidegger: Through Phenomenology to Thought" by William J. Richardson. We are not saying that our lives are abandoned or unwanted but that they are not explicitly/clearly justified. We can not deny that we, and all living things for that matter, are born naked of explicit function. Our parents might like each other and us very much but they don't have any idea what is going on either, in terms of the universe they found themselves and each other in. 

The show invites us to reflect on what we discover and become in the chaos of our projectile existence. Through Desmond and Penny, we are reminded that even in disarray, we can still find one another—and that connection can be the invariant that keeps us from drowning deep in the chaos. Being hurled through life would be much more harrowing if faced alone and, in some ways, that is the essence of hell. Michael shows us the cost of betraying fellow lost souls, no matter the justification. We look at Ben and John to see how the pursuit of destiny twists one man and makes another man whole. And in Jack and Charlie, we see the appeal of sacrificial leadership. 

One exception to the narrative portrayed by Lost and lived out in real life is the position held in Christianity. Christians are told not to see themselves as orphans lost in the world but rather as legitimate children of the world's creator set to ultimately inherit this world which right now appears to be very much like the island in Lost and quite full of orphans. 

Space Gas Stations and Hidden Markov Models