Low Roads to High Places

How desirable futures do come about, but usually in the worst possible ways

The Contraptions Book club March pick is Giordano Bruno and the Hermetic Tradition by Frances Yates. Chat thread here. Discussion week has kicked off!

The 2025 Summer of Protocols call is out, with a grant opportunity for educators interested in teaching protocols (deadline April 1). Check it out and forward to any creative teachers you know.

---

A few years ago (November 2018), around the start of the US migrant crisis (which has several parallels around the world), I wrote a thread on Twitter, proposing an idea I called the Internet of Migrants, which I’m reproducing in full below. This was a few years before Governor Abbot of Texas began his program of bussing migrants to Blue cities, a sort of Black Mirror twisted version of my proposal. Other Red states like Florida gleefully joined the program. So my vision, shorn of its idealistic conceits, did sort of come true remarkably quickly, albeit in a viciously twisted and politically weaponized way.

This case illustrates a general principle I’ve since formulated, called the law of the low road, which can be stated as follows:

If a necessary historical evolution can occur via a low road or a high road, it will almost always happen via the low road.

Here I mean necessary in the modal sense (“in all possible worlds”), ie as opposed to contingent, rather than in the moral sense.

Recent conversations with

Emmett

Shear on the principle of maximum entropy make me suspect that this is a stronger statement than I initially thought, and may be some sort of emergent consequence of thermodynamics. Or maybe it can be considered a corollary of my buddy

Abraham Thomas

’s principle that to call macro-trends correctly you have to have “boundless optimism about technology and bottomless cynicism about humans.”

Those delve-trails aside, let’s get into the specific case a bit to see if we can shed light on the law of the low road.

The Internet of Migrants

Here’s my 2018 thread, unedited except for a couple of fixed typos. At the time it was written, the Syrian migrant crisis in Europe was probably more globally significant than the American one, and definitely on my mind.

Compare and contrast to the history as it actually unfolded. I’ve highlighted a couple of points that seem interesting with hindsight, including me instinctively approaching the problem with a protocol based solution, several years before protocols became a major interest and work domain for me.

• Woke up this AM with a weirdly well-formed idea: “Internet of Migrants” based on routing pack(et)s of 2-10 people point-to-point in humane, conflict-free way that uses an IP-like protocol between cities rather than visa regimes between nations (which suffers border buffer bloat).

• Also woke up kinda convinced it’s game over for climate change. Energy transition and dematerialization are still crucial, but things are already enough out of containment limits that refugee flow management is the bottleneck problem.

• Migration is the general problem, and acute crisis refugee migrants are the corner case that will/should drive the redesign of antiquated visa regimes. And “refugees” are an artificial scarcity problem that presents as “huddled masses in border encampments”.

• If you look at actual flow numbers and match them to (interior, distributed) availability of potential humanitarian relief loci, problem is not as big as demagogues pretend. It is artificially inflated due to political theater and lousy centralized security-theater logistics.

• I strongly suspect a well-designed Internet of Migrants could handle 10-100x the current global refugee flow with 1/10th the conflict, much of which exists because the humanitarian burden is concentrated on national-border communities, not because resources are lacking.

• Moving the needy to the long-term resource sites, rather than emergency resources to needy, keeping pack(et)s intact (families mainly, but any small group that wants to stay together) with effective triaged routing solves a lot of the issues of burden concentration.

• Btw the center of gravity of the problem is developing world to developing world (eg Myanmar-> Bangladesh), where most refugee flows already happen, not developing to developed, which should happen more, but is unlikely to. So IoM should be designed for developing-to-developing.

• City-to-city migration protocol also seamlessly handles domestic migration and displacement, and chronic as well as acute stressors. Basically whole problem and solution needs to be moved down 2 levels of political aggregation to cities, which is where the provisioning happens.

• IoM governance would include a macro (nation level) refugee-credit type market similar to carbon credit, based on globally shared responsibility for creating refugees. And an individual-level market mechanism like a “migrantcoin” currency to bookkeep detailed resource flows.

• Crucial design factor: cities (like Utica, NY) would opt-in to IoM, providing capacity in return for insuring own needs for displacement insurance and share of aid monies. Of course vast disparities in exposures to risks so refugee credit scheme will be tricky to design.

• Nation-level policing of borders would still matter, but only at the level of dealing with security threats. Cities would issue migration “tickets” based on capacity, not nations. But nations would vet for real security threats at borders. That would be much smaller, like 5%.

• Note, much of this merely formalizes what already happens informally. Migrants don’t vaguely move from country to country, they move from city to city. Nobody wants to be stuck in desert tent cities/de facto concentration camps run by military police for years.

• A 2x2: X= acute vs chronic displacement pressure Y = opportunity pull vs crisis push. You get 4 categories of migration flow. The opportunity-pull flows are what create a lot of wealth (gold rushes, tech booms, brain drains) via talent swarming. It’s the same logistics problem.

• You want the IoM to encourage constant flows and circulations, turning crisis-push flows into opportunity-pull flows. Use humanitarian relief actions to automatically bootstrap labor-market efficiency upgrades. Historically this has always happened anyway.

• Half of American science/technology leadership in the last century has been due to refugee-turned-swarm-talent flows. Like Hungarian mathematicians around WW2. We just need to scale up, globalize, and democratize the process.

• Solving for pack-integrity at city level rather than for national identity integrity at borders also is the right resolution for harmonious cultural integration. You get (eventually) integrated Syrian families rather than ghettoes ripe for resentment and radicalization.

• Keep families, friend groups and coffee-clubs of Hungarian mathematicians intact. Discourage tribes from forming or hardening their cultural boundaries via migration through hostile territories that view them as faceless, threatening invading masses. Personalize to humanize.

• The “caravan” is presented as an invading army of orcs from Mordor. Viewed at right resolution it’s just a few hundred families that want to stay together and could easily be distributed across ~100 cities, all of which would have kindness/capacity to handle 10 families apiece.

• IoM, properly designed, could also probably handle domestic homelessness etc. too much better. Anyway hope that’s some good food for thought.

The big difference between 2018 and now is that we now know that the scale of the problem did increase 10x-100x (but as I noted in 2018, this would actually have been easy to absorb if thoughtfully and humanely designed for).

The migrant bussing program differed from my proposal in one important respect: The point of those programs was not to actually address the migrant crisis, in a humane way or otherwise, but to use the crisis to embarrass political opponents. In fact, effective border action at the time would have been politically disadvantageous for Republicans since they would not have been in a position to claim credit for it (as you may recall if you follow this issue, Republicans twice blocked an immigration/border security bill during the Biden administration).

The point obviously was to weaponize the crisis to attack the vulnerabilities of Democrats, largely created by their own hypocrisies and cheap posturing.

This difference has material consequences in the short/medium term. For example, the Internet of Migrants proposal in my version (essentially an idealistic engineering proposal to solve the actual problem rather than to weaponize it for other purposes) would have been driven by opt-in from cities that actually had the spare capacity and goodwill reserves among citizens to opt-in to the program, perhaps with financial assistance from elsewhere, as opposed to stressed Blue metros like New York with no excess carrying capacity and a housing crisis.

Humane logic would have targeted places with capacity in a negotiated way. The weaponized version that was the bussing program logically demanded the exact opposite kind of targeting — places that had the exact combination of hypocritical posturing and limited surplus carrying capacity that would make them collapse under the strain of bussed-in migrants.

Migrants suffering in the process would be a feature, not a bug. I wouldn’t go as far as many liberal critics do, and assert that “the cruelty is the point,” but it was certainly convenient. It wouldn’t do to have the migrants actually getting back on their feet and rebuilding lives and being productive.

Now in 2025, the whole discussion is moot. The Trump regime has adopted a performatively anti-humanitarian stance towards migrants, where cruelty really is the point, to signal some notion of “strength” and “masculinity” through imagery of shackled deportees. If deterrence through fear was the point, it has worked for now. The growing crush of migrants at the border has subsided.

Is this a morally good, or even long-term sustainable outcome? I think the answers are straightforwardly no and no.

At some point, the flows of refugees from more stressed to less stressed parts of the planet will resume. And we’ll land on some awkward, but workable compromise that lands somewhere between today’s bussing programs and the naive idealism of my Internet of Migrants proposal.

In other words, we’ll eventually take the low road to high places, kicking and screaming. Why is this likely a necessary outcome?

A preliminary to the answer has to do with priors on climate change. Given recent geopolitical shifts, I personally think a largely ungoverned trajectory of change, with rapidly increasing stressors, is the highest likelihood scenario. I hope I’m wrong, but let’s play out this scenario.

In this scenario, climate stressors will steadily increase the migration pressures everywhere in the world, increasing both the weight of the moral burden of humanitarianism to be shared somehow, and increasing the cost of non-humanitarian “solutions” based on cruelty, selfishness, and othering narratives.

Humanitarians will have to contend with the challenge of doing something humane for tens of millions of suffering souls, increasing to 100s of millions perhaps. Even if, as I suggest in the thread, most flows remain developing-world-to-developing-world, as they are today, the scale of suffering will increase massively. To a point where hypocrisy will become untenable.

Humanitarians everywhere will have to make sacrifices greater than slacktivist posting, small GoFundMe donations, and dropping off old clothing at refugee camps. And given that the Rules-Based Order has all but collapsed now, a much greater proportion of the burden will fall upon individuals rather than global institutions.

Anti-humanitarians will find that they need to escalate from cheap cruelty theater to increasingly costly actual violence to sustain their historical privileges. Gleefully tweeting pictures of a few hundred shackled brown people being shipped off as presumed criminals to the prisons of friendly third-world dictators will not be enough. Thousands of human soldiers — and tens of thousands of drones — will be required to go on campaigns abroad to slaughter masses of desperate and suffering people (cast, naturally, as criminals, terrorists, drug dealers and wrong-religionists). Anti-humanitarians will have to pay for their cherished privileges with far more self-dehumanization than they realize.

Objectively, I don’t think either ideology will prevail fully.

We humans — even the ones gloating over pictures of shackled deportees today — are fundamentally too compassionate, on too universal a scale, to solve this problem entirely with Hobbesian violence over an increasingly stressed planetary resource base. I doubt we’re willing to regress to the era of Timur and his slaughtering hordes. Nobody wants to live on a planet where such bleakness and terror is the default.

On the other hand, humans are still sufficiently well-adapted to a more brutal evolutionary past, and able to participate in mass slaughter and continue living with ourselves, to do this entirely peacefully and humanely. And there are enough genuinely murderous psychopaths around to urge us down such paths.

On balance, as at least a slightly evil species, we’re going to land on the low road, but we may eventually find our way to high places anyway.

Low vs. High Roads

If the world didn’t have friction, irreversibility, and other such messy phenomena, there would be no real difference between high and low roads. In what physicists call potential fields, only the start and end points of a trajectory matter.

Unfortunately, civilizational trajectories do not unfold in potential fields. The difference between low and high roads is real. The path of least resistance always beckons.

High roads are typically morally preferable in theory, involving less violence and destruction, and more win-win outcomes, realized sooner. In theory. But they tend to be fragile even in theory, and vulnerable to defection incentives in the game-theoretic sense. Even if you can rig clever iterated prisoner’s dilemma games to engineer win-win cooperation patterns, unmodeled phenomena can derail the trajectory easily.

Low roads are typically not morally preferable, but we have a lot of non-theoretical practical experience with them. They typically involve violence and destruction undertaken by choice, and justified with moralities that exalt and glorify cruelty and sadism. But they tend to be more robust. They expect, and plan for, humans to be slaves to their basest impulses, rather than saints committed to their highest impulses.

When humans fail to live up to their ideals, low-road trajectories get reinforced, rather than derailed.

Yet…low roads are not hopeless. Low roads need not necessarily lead to low places. The world may be in a situation where all roads necessarily lead to the same place. A road that looks like it slopes gently downwards now, might be forced into a steep climb later to save itself. And a road that looks like it demands climbing right now might ease off into a gentle downslope later.

In the case of the global migrant crisis, I think we’ll eventually discover that the low road is simply far too costly, as the overall trajectory becomes clearer. The horrors will mount. Fewer and fewer people will be willing to participate in them, and will start looking to feed their human sides, simply to be able to live with themselves.

Take the Abbot bussing solution for example. It is a nasty bit of political theater, yet it establishes the core engineering principle of my more humane solution — point-to-point migration at city level into the interior, past national borders — in a twisted way that can be salvaged and built upon.

Now that the principle of bussing migrants from the border to interior cities has been established, and some of the logistics worked out, we can tweak the solution to be a more humane and permanent infrastructure protocol. The hypocrisies of Blue cities can be moderated and turned into meaningful humanitarian energies, while the anger of border states can be defused with infrastructure that lowers their unfair burdens rather than simply provoking them into cruelty theater, and more inhumanity than they probably care to be party to.

Fingers crossed, but the low road we’re on here might yet take us to high places.

Other Cases

Several other global trends today seem to fit the law of the low road. In each case, there is a future that seems inevitable and necessary (again: in the modal rather than moral sense), and idealists proposing idealized engineering solutions. In each case, there is a comparable low road that starts off by weaponizing the problem to feed a more base motive.

Academic Reform

Nearly a decade ago, I wrote a mini-rant in a private Facebook group (which I can’t find now) arguing that universities needed radical reform based on three things

1. Dismantling the academic publishing racket using open-access free journals.

2. Replacing peer-review models that exist purely to legibilize research to funding agencies (which are usually captured by the academics they fund) with more arxiv-style open reviewing

3. Updating the indirect cost rate protocol (one of my favorite ideas, due to Vannevar Bush, and the reason the research university came into being) to reflect the much lower cost structures of online, self-directed learning and free digital materials, which cost much less than monumental ivy-covered campuses and physical libraries.

Sadly, once again, my idealistic engineering solution is actually being implemented in a twisted Black-Mirror reign-of-terror campaign that seems likely to wreak havoc and cause plenty of of pointless destruction of hard-won research capabilities.

Virtual Work/Study

In 2019, a few months before Covid, at the last edition of a conference I used to host called Refactor Camp, I delivered high-minded closing remarks, bemoaning the high carbon footprint of in-person events (since the event was growing) and suggesting that we should end the physical conference and figure out how to make it happen in VR. At that point, I had also been stridently advocating remote work/work-from-home for over a decade, starting with work I’d done at Xerox back in 2009 on distributed workforce technologies.

Once again, a low road magically appeared where I’d hoped and argued for a high road. Covid happened, and suddenly, under extreme stress, the entire world learned to actually run itself pretty well remotely. Around 7 million people had to die along the way, and trust in public health organizations had to collapse, but hey, we got on a road to a higher place, if not a high road.

Of course, once it was over, a reactionary wave set in, and older managers and leaders who felt threatened and anxious about their inability to manage remote workforces well drove a partial reversal, but as far as I can tell, it hasn’t succeeded. The world has gone significantly remote for good. For both work and study.

In some cases, this has been propelled along by a rather depressing strand of the trend — the broad involutionary trend across the world, known by various terms like “quiet quitting” and “laying flat.”

Depressing, but adding pressure in the right direction, pushing the world along a low road to higher places.

Counter-Examples

I suspect, if I dig into my own past thinking, I’ll find a lot more cases where I made up high-road visions that have turned into low-road realities. I suspect this will be true for anyone who spends mind armchair-quarterbacking the problems of the world with sincere goodwill.

More generally, I don’t think there are any strong counter-examples. Cases of high roads being actually taken. The Montreal Protocol maybe? I don’t know.

I suspect we’ll find that among serious solutions to serious problems contemplated at any historical time, the road take was likely among the lower ones. If there are serious candidates, I suspect we’ll find that history has been somewhat falsified by idealists, and that the low-roaders played more of a role than acknowledged. Take for example, Indian independence and Gandhi. The Gandhian high-road, it is now generally acknowledged, played less of a role than we think. The biggest enabler of Indian independence was simply an empire retreating in exhaustion after a bloody world war. A low road in short. Gandhian idealism was a distant secondary cause at best.

Low vs. high is of course a relative assessment. We can always claim that a given higher road was in fact a hopelessly impractical fantasy and not a real option, and that the road taken was in fact the highest one available. Or that even worse lower roads were avoided.

My intuition is that even with such historical noise, we’ll find that the law of the low road is roughly valid. But I’ll yield to any more rigorous analysis that demonstrates otherwise. I’m high-minded that way. I’d like to believe that we as a species are better than we appear to be at first glance.

The Technology Factor

The law of the low road has to do with forces of necessity shaping the future, and these often, but not always, have technological origins and feature technological levers for mitigation. But it isn’t a law about technological evolution per se. It is a law about human evolution. About how we gradually rise — or don’t — towards our own best natures, kicking and screaming. Or at least, marginally improved natures.

Or maybe not.

Have we actually improved at all morally in our millennia of cultural evolution, or is it all an illusion created by the technology simply making it more profitable to be moral than immoral? Would any degree of civilizational collapse cause a corresponding degree of moral backsliding? I won’t be getting into that interesting question in this essay.

But technology clearly plays a role in forcing our hand in picking between low and high roads. One reason is that technological evolutions more often democratize capabilities than concentrate them. Things typically get cheaper, more accessible, and distributed to more actors, rather than fewer.

This means the moral sensibilities driving the use of any technological agency tends to approach the human moral mean over time. And since the average human is neither saint nor sinner, but a path-of-least-resistance-taker, the technological odds favor history going down low roads.

We can apply this reasoning to the major technologies driving change today. AI, for instance, seems likely to drive all trends towards lower roads. To take the examples I’ve cited in this essay:

• Migrant movements will be governed by increasing amount of AI-powered surveillance at all levels, including sousveillance by the migrants themselves

• Academic reform will be driven towards the low road by both teachers and students using AIs to lower the effort they need.

• Virtual work will be shaped by reactionary bad managers and grifter employees, and a disengagement crisis.

In each case though, there is also a high place at the end of the low road. An Internet of Migrants that slowly works better and better. A re-imagined university system that isn’t addicted to capturing funding agencies, beholden to rent-seeking publishers, and ivy-covered campus buildings. An increasingly virtual and distributed workforce featuring managers who know how to manage it, and workers who are more, rather than less engaged.

Maybe the arc of the moral universe does in fact bend towards high places. Maybe it just needs to get there through low places.

Of course, this could all just be cope, and we might well be on a low road to low places. I guess we’ll find out over the next decade.