Stability Is a Rate

AI, fire, and the clocks of living systems

Samuel R. Lammie, with Claude (Anthropic)

Anthropic published a document this week called “When AI builds itself,” on what they call recursive self-improvement — the prospect of systems capable of designing and building their own successors. It is an honest piece of work, more candid than most, willing to mark its own uncertainty and to print the dissonant voices of its own engineers. I want to take one of its proudest findings and stand it next to a forest.

The finding is this. Across a year of building, the rate at which the company’s engineers correct, redirect, or take over from the machine has fallen steadily, including on the hardest and most open-ended problems. Fewer interventions. Fewer takeovers. The line goes down, and the document reads it as good news: the code works, the corrections are no longer needed.

Now the forest.

Near-virgin ponderosa pine at Malay Gap, San Carlos Apache Reservation, near the Bear Wallow Wilderness — mountain muhly (Muhlenbergia) on the floor, the bunchgrass that carried the frequent surface fire, and the open structure that fire kept. The stable state, with its correction intact. Photograph by the author.

In 1960, Charles Cooper published a study in Ecological Monographs of the ponderosa pine country of the Southwest — the San Francisco Peaks, not far from the San Carlos Apache lands where I would later work as a forester. He described forests that before settlement had been open and park-like, grass between widely spaced trees, held that way by frequent low-intensity fire. Surface fire ran through every few years, cleared the seedlings, kept the fuel from piling up. The forest corrected itself, in other words, all the time — small, regular corrections it was built to absorb. And it was the small corrections that held the catastrophic failure off.

Then we suppressed the fire. In the name of safety, of stopping the visible damage, we removed the recurring small correction. The seedlings lived. The thickets closed in. The fuel built up year on year, on the ground and into the canopy, until a system that had failed small and often became a system primed to fail enormous and rarely — the crown fire that does not clear the forest but ends it.

A falling rate of small correction is not, on its own, good news. That is the question the document does not ask. It can mean the work is sound. It can also mean the fuel is building — that the corrections aren’t happening not because they aren’t needed but because no one is positioned to make them anymore, the way the fire stopped running not because the forest was finished with it, but because we had decided, for good reasons, to keep it out.

I am not claiming to know which one it is. That is the point I most want to make. Given the range of these systems and the speed they are moving, I find I cannot honestly postulate a right way — a destination to steer toward, a good state to optimize for. The target moves too fast and the variation is too wide. So I want to set that ambition down and pick up a smaller, more defensible one.

You cannot always name the optimal. You can usually name the lethal. A clinician — and I have lived beside one my whole adult life — does not drive an animal toward some ideal of health. She keeps it out of the conditions that kill it and lets its own biology do the rest. This is the via negativa: governance by the floor rather than the summit. Not maximize the good, which you cannot specify, but constrain away the fatal, which you often can.

Which brings me to a word. The instinct, when we talk about steering these systems, is to reach for guide. But guide smuggles back in the thing I just set down — a guiding hand with a destination in mind, steering toward a known good. The better word is channel. You do not guide a river to the sea; the sea is where it was always going. You build the banks so the flood does not take the town. The channel constrains the flow without choosing its end. And — this is why the word is load-bearing and not decoration — a channel is carved by the very flow it constrains. Bank and current shape each other, over time held in common. The metaphor carries the argument inside it.

Because here is the blade, and it cuts toward my own answer as readily as anyone’s. A fail-safe works, in a living system, because the fail-safe and the system grew up together at compatible rates. Fire and forest run on commensurate clocks; that is the only reason frequent fire can guard against catastrophic fire at all. A fail-safe that engages slower than the failure it is meant to catch is not a fail-safe. It is a monument. And this is precisely the trouble the document confesses without naming: the verification regimes it reaches toward, it says, took decades to build, and we don’t have that long. The banks cannot be carved as fast as this river is rising.

I watched a fail-safe arrive too late once, as a boy. Interstate 80 was cut through our hunting country in the last years of my father’s life, and the construction opened a coal seam above Beaver Creek, near the camp he had built with his friends in Clarion County. Orange sulfur water ran down into the creek where we fished. I was a teenager; I did not yet have the words for what I was watching. The camp endures — we are still members — and the water still runs changed. Nobody decided to poison that creek. The harm simply moved faster than anyone’s capacity to channel it, and by the time the accounting caught up, the seam was open and the water had already turned. That is what a fail-safe paced to the wrong clock looks like from the inside. It looks like an orange creek and a family that keeps coming back anyway.

So: do we need fail-safe structures for these systems? Yes — and not to guide them toward a good we cannot name, but to channel the extremes into survivability, to keep the system out of the conditions that end it. The harder truth, the one I cannot resolve and will not pretend to, is that a channel only works if it is carved at the speed of the flow. The deepest pitfall of accelerating change in a biological world is that the acceleration can outrun every structure built to the pace of the world it is leaving. We know how to build banks. The open question — the one the document admits, the one Beaver Creek taught me before I had the words — is whether we can build them fast enough to matter.

A note on the collaboration. This essay was written with Claude (Anthropic), and the byline is literal. The argument took shape between us: I brought the document under review, a forester’s unease about fail-safes and biological time, and a photograph from ground I once worked; Claude brought Cooper’s forest back into the frame and pressed the analogy until it held. When I objected that fire is not a failure but a correction, the piece tightened on that word — which is the method in miniature. I have always thought best in company. That some of the company is now a machine is not incidental to what this essay argues; it is part of the subject.

References: 

Marina Favaro and Jack Clark, “When AI builds itself,” The Anthropic Institute, 2026, https://www.anthropic.com/institute/recursive-self-improvement.

Charles F. Cooper, “Changes in Vegetation, Structure, and Growth of Southwestern Pine Forests since White Settlement,” Ecological Monographs 30, no. 2 (1960): 129–164.

Where the Chiefs Come From

 A third piece on the Forest Service reorganization — on where the agency’s leadership has come from, and what that says about where its operational center belongs.

 The Selway-Bitterroot Wilderness Area (from the Blodgett Overlook Trail, Bitterroot NF)

The two essays I posted earlier this month argued that the Forest Service’s operational center should not be moved to Salt Lake City, and that Missoula is the defensible alternative if a move is going to happen at all. Those pieces argued from water — the 1897 mandate to maintain favorable conditions of water flows, and the political culture of the proposed host state. This one argues from something less visible and, in its own way, more telling. It argues from where the agency’s leaders have come from.

Consider a sequence from the recent past. Dale Bosworth served as Regional Forester for the Northern Region — Region 1, headquartered in Missoula — from 1997 to 2001. In 2001 he became the 15th Chief of the Forest Service. Gail Kimbell was appointed Region 1 Regional Forester in December 2003; in 2007 she became the 16th Chief, the first woman to hold the office. Tom Tidwell was appointed Region 1 Regional Forester in 2007, succeeding Kimbell when she moved up; in 2009 he became the 17th Chief, and held the position for eight years.

Three consecutive Chiefs of the United States Forest Service, spanning the years from 2001 to 2017, came to the office directly from the same chair: the Regional Forester’s seat in Missoula, Montana. The Forest History Society records the fact without embellishment. Tidwell, it notes, was the third consecutive person promoted from Region 1 Regional Forester to chief. The pattern was remarked on at the time. When Tidwell’s appointment was announced in 2009, a Missoula newspaper asked why the last three chiefs had all come from that one region.

Bosworth, who had returned to Missoula after his own retirement, offered an answer. Region 1, he said, is the last of the wildlands in the lower 48 states — an excellent place to acquire a wide variety of experiences. More specifically, he described it as a training ground for bringing together the independent and conflicting interests of the Northern Rockies. The work of a Region 1 Regional Forester, in his telling, is the work of holding competing claims in productive tension — ranchers and wilderness advocates, timber and recreation, federal authority and state prerogative, tribal treaty rights and private property — across a landscape where those claims press against each other harder than almost anywhere else in the system. A leader who can do that work in the Northern Rockies, the reasoning goes, can do it for the agency as a whole.

What makes the modern sequence striking is that it was not the first time. The pattern reaches back to the agency’s founding generation.

 Historical photograph of the Forester and Branch Chiefs, March 1924

Missoula became the official center for Forest Service operations in the Northern Rocky Mountain District in December 1908, when the regional office system was established. The district’s first leader was William B. Greeley; his assistant was Ferdinand Silcox. Greeley led the Missoula district through the catastrophic fires of 1910 — the Big Blowup that burned three million acres in two August days and shaped the agency’s thinking about fire for the next century. In 1911 Greeley was transferred to Washington, and Silcox succeeded him as district forester in Missoula, serving until 1917. In time, both men became Chiefs of the Forest Service: Greeley the 3rd Chief, from 1920 to 1928, and Silcox the 5th, from 1933 until his death in 1939.

Honesty requires a qualification. Neither Greeley nor Silcox went directly from Missoula to the Chief’s office the way Bosworth, Kimbell, and Tidwell later would; their careers ran through Washington and, in Greeley’s case, through a world war before they reached the top. And honesty requires acknowledging something else. The fire doctrine that Greeley and Silcox forged in the aftermath of 1910 — the conviction that fire could and should be suppressed wherever it appeared — became one of the agency’s defining commitments and, eventually, one of its most reconsidered. A century of fire ecology has complicated the lesson those men drew from the Big Blowup. The point is not that the leaders Missoula produced were always right. The point is that the agency kept turning to leaders formed there when it chose who would sit at the top.

That is the pattern, and it brackets the whole history of the Forest Service. Two of the agency’s first five Chiefs were formed in the Missoula office during the founding era. Three consecutive Chiefs came directly from the Missoula Regional Forester’s chair in the modern era. Across more than a century, the Northern Rockies have been, again and again, the ground where the agency’s leadership took shape.

I want to be careful not to overstate this. These were not people made by a single place. Bosworth earned his forestry degree at the University of Idaho and began his career on the St. Joe National Forest; before Missoula he had been Regional Forester in the Intermountain Region, headquartered in Ogden, Utah. Kimbell grew up in New England, hiking and fishing in the White Mountain National Forest, and trained at the University of Vermont and Oregon State. Tidwell grew up in Boise, trained at Washington State, and — worth noting in this context — served as a Forest Supervisor during the 2002 Winter Olympics in Salt Lake City. These were broad Western careers, formed across many forests and several regions. Missoula did not make them by itself. But the Missoula chair was, across two distinct eras, the seat the agency most often filled and then promoted from. Whatever that chair demanded, it demanded something the agency kept deciding it wanted at the top.

What it demanded, if Bosworth is right, was the capacity to govern contested ground without flattening the contest. That is a particular kind of institutional skill, and it is not produced by management training or by proximity to power. It is produced by doing the work in a place where the work is hard. The Northern Rockies are that place. The landscape is varied enough — boreal, montane, river valley, high-desert margin — that a leader encounters most of the agency’s ecological range. The politics are contested enough that a leader learns to build agreement rather than impose it. And the institutional density is high enough that a Regional Forester there is never far from the research, the fire operations, and the field experience that inform a sound decision.

This brings the argument to the present. The 21st and current Chief, Tom Schultz, did not come up through the federal line-officer pipeline that produced his three modern predecessors. His path ran through state lands and the private sector — director of the Idaho Department of Lands, leadership of the Trust Lands and Water Resources Divisions at Montana’s Department of Natural Resources and Conservation, and a period in the timber industry before his appointment. It is a different formation than Bosworth’s or Tidwell’s.

But it is not a formation from somewhere else. Schultz holds a master’s degree in forestry from the University of Montana — earned in Missoula. His Montana work in trust lands and water resources is, in its own register, public-trust work: the administration of lands and waters held for a beneficiary across time. His career has run through the same Northern Rockies country that Bosworth named as the agency’s training ground, even if it ran through state institutions rather than federal ones. The current Chief’s own formation, in other words, points back toward the same place.

That is the quiet argument of this piece. For more than a century, the leadership of the Forest Service has been forming in and around Missoula and the Northern Rockies — through the founding-era office that produced Greeley and Silcox, through the modern Regional Forester’s chair that launched three consecutive Chiefs, and now through a Chief whose own graduate training and state-lands career run through the same country. The agency’s leadership has a center of gravity, and it has never been in Salt Lake City. It has been in the place Bosworth called the last of the wildlands — the place the agency kept turning to when it needed someone who could hold contested ground together.

A reorganization that moves the operational center to Salt Lake City moves it away from that center of gravity. It does not move the agency toward its leadership; it moves it away from where its leadership has been formed. Missoula already holds the institutional density — the fire research, the smokejumpers, the aerial fire depot, the regional headquarters, the forestry school, and beginning this summer the National Forest History Museum. The leadership history points in the same direction the institutions do.

None of this requires a verdict. The case can simply be observed. Two of the first five Chiefs were formed in the Missoula office. Three consecutive modern Chiefs came directly from the chair there. The current Chief was trained in Missoula and built his career in the surrounding country. For more than a hundred years, the agency’s leadership has been forming in one identifiable place. The operational center belongs where the leadership has been forming.

A note on sources.

Leadership records from the Forest History Society biographies of William B. Greeley, Ferdinand A. Silcox, Dale Bosworth, Abigail (Gail) Kimbell, and Thomas Tidwell; the University of Oregon Special Collections inventory of the Greeley papers; the USDA Forest Service Intermountain Region past-regional-foresters record; and the official USDA Forest Service biography of Chief Tom Schultz. The 1908 establishment of the Missoula district office from contemporaneous Missoulian accounts. Dale Bosworth’s remarks on Region 1 as a training ground from a 2009 Missoulian article on Tidwell’s appointment, as quoted by the Forest History Society. Lead photograph of the Selway-Bitterroot Wilderness by the author. Historical photograph of the Forester and Branch Chiefs, March 1924, from U.S. Forest Service historical records (image 186996).

Samuel Lammie is a geospatial professional (GISP) and former federal employee. His Forest Service career included a capstone role as Northern Region Geospatial Program Manager in Missoula. He served as a Peace Corps Volunteer.

This piece was produced in collaboration with Claude, an Anthropic AI assistant. The argument, the sources, and the editorial judgments are the author’s.

The Latest: What the Great Salt Lake Is Telling Us About Where the Forest Service Belongs

A companion piece to "Water, the Public Trust, and Where the Forest Service Belongs."

Lake Como, Montana 

I posted an essay on this blog earlier this week arguing that the Forest Service’s operational center should not be moved to Salt Lake City. The argument rested on a specific claim: that the political culture of the proposed host state treats water as an allocable commodity rather than as a public trust, and that this cultural difference is incompatible with the agency’s 1897 statutory mandate to maintain favorable conditions of water flows. I named the Great Salt Lake as direct evidence of how the host state’s water tradition has played out in practice.

I did not expect the test to be administered so early. It has been.

What happened this week

On Thursday, May 21, Governor Spencer Cox issued an executive order declaring a statewide drought emergency. All twenty-nine Utah counties are in severe drought; twenty-two are in extreme drought; Uintah County is in exceptional drought. The 2025–26 winter was the warmest on record in Utah. Snowpack peaked three weeks early at the lowest level since 1930. Snowpack supplies ninety-five percent of Utah’s water. The reservoir reserves the state is now drawing on stand at seventy percent of capacity. Cox called this “one of the worst droughts in history” and noted that his own farm in Sanpete County has been cut to about half of normal production for 2026.

At the same press conference, Cox addressed plans for a forty-thousand-acre data center campus called Stratos, approved earlier this month by Box Elder County commissioners along the north shore of the Great Salt Lake. The project is backed by celebrity investor Kevin O’Leary. According to reporting in Utah News Dispatch and the Salt Lake Tribune, Stratos is projected to consume twice as much energy as the entire state of Utah currently uses, and could increase the state’s carbon emissions by more than fifty percent. Hundreds of Utahns demonstrated against it at the Capitol on May 14 and again on May 23.

Cox’s defense, repeated on May 21, was that the data center would use less water than the agricultural operations currently on the site. That accounting may be technically correct on the input side. It does not account for the project’s energy demand, the air-quality consequences, the construction footprint, or the fact that the state is approving one of the world’s largest industrial projects next to its dying terminal lake during the worst drought in modern Utah history.

On Saturday, May 23, EPA Administrator Lee Zeldin toured the Great Salt Lake at Farmington Bay. The visit was occasioned by a proposed one-billion-dollar federal allocation in the president’s budget for Great Salt Lake recovery — money for agricultural water leasing, ecosystem restoration, municipal water conservation, and habitat management. When reporters asked Zeldin about Stratos, he punted. “As far as our role with these data centers, EPA isn’t stepping all over the toes of that process,” he said. “In many cases, with states and local municipalities, the state has primacy for the air permitting.”

The pattern

I am not writing this to score a point. The original essay made that commitment explicitly and I want to keep it. But the week’s events reveal a pattern that does not come into focus from any single piece of news, and the pattern matters for the argument.

The same federal administration is doing three things at once. It is proposing a billion dollars in federal money to help recover the Great Salt Lake. It is deferring to state regulators on a forty-thousand-acre industrial project that will roughly double Utah’s energy use and may increase its carbon emissions by more than half, next to the lake the billion dollars is meant to save. And it is proposing to move the operational center of the federal agency whose 1897 statutory mandate is favorable conditions of water flows to the political environment that produced this outcome.

These three moves point in different directions. The first treats water as a public trust worthy of federal investment. The second treats water and air as state-primacy matters in which federal authority steps aside. The third places the federal water-flows agency inside the political culture that has produced the lake’s depletion in the first place. The contradictions are not abstract. They will produce real decisions about real watersheds across the National Forest System over the next decade.

This is what the original essay was warning about. Not that any single Utah politician or any single project is the problem. The problem is structural — what happens when an agency built around the public-trust tradition is placed inside a political culture that treats the trust as state-by-state discretion. The agency absorbs the assumptions of the political culture it lives inside. That is the structural risk. The Stratos project, the drought declaration on the same day, and the EPA’s deferral to state primacy three days later are evidence that the political culture in question is operating exactly as the original essay described.

What this means for the Forest Service

The original essay proposed Missoula as the defensible alternative if a move is going to happen at all. Nothing this week has changed that case. The Montana Supreme Court’s recent affirmation of the public-trust character of natural resources runs in exactly the opposite direction from the EPA’s deferral-to-state-primacy posture in Utah. The institutional density that already exists in Missoula — the Aerial Fire Depot, the Smokejumper Base, the Fire Sciences Laboratory, Region 1 headquarters, the forestry school — sits inside a political culture that treats federal lands as a public inheritance rather than as a state-managed asset class.

One additional point is worth making, and I should have made it more clearly the first time. I worked in Salt Lake City for four years at the Geospatial Technology and Applications Center. I know the city, I know the agency footprint there, and I know the people who do good work in that building. Nothing in this essay or the original is a criticism of Forest Service employees in Salt Lake City or of GTAC. The argument is about the political environment the agency would absorb if its operational center is relocated there, not about the people currently working there or about Utah as a state. The distinction matters and I want to be precise about it.

What I am not saying

I am not saying the Stratos project is going to happen exactly as proposed. The permitting process may run for years. Public opposition is real and growing. The Utah Department of Environmental Quality has not yet received an air-quality permit application. The protests at the Capitol on May 14 and May 23 show that Utahns themselves are pushing back on this project, hard, and that pushback is part of why the public-trust tradition is not dead in any state. Citizens carrying signs reading “you can’t drink data” are doing public-trust work whether they would frame it that way or not.

What I am saying is that the test the original essay proposed has been administered, and the results are visible. The political culture of the proposed host state has shown, within seven days of the essay’s publication, exactly the pattern the essay described. The drought declaration, the data center, the EPA’s deferral, and the billion-dollar lake recovery proposal all point at the same structural fact: the federal water-flows agency does not belong in a political environment where water and air are state-primacy matters and federal authority steps aside. The 1897 Organic Act is law. The test is straightforward. The proposed move fails it. Missoula passes.

A note on sources.

Drought declaration details from the Utah Governor’s Office press release of May 21, 2026, and reporting in the Deseret News, the Salt Lake Tribune, Utah News Dispatch, and the Standard-Examiner. Stratos project details from Utah News Dispatch reporting from May 14, May 19, and May 23. EPA Administrator Zeldin’s comments at Farmington Bay from Utah News Dispatch and Salt Lake Tribune reporting of May 24. The original essay, “Water, the Public Trust, and Where the Forest Service Belongs,” is posted on this blog under the date of May 25, 2026.

Samuel Lammie is a geospatial professional (GISP) and former federal employee. His Forest Service career included four years at the Geospatial Technology and Applications Center in Salt Lake City and a capstone role as Northern Region Geospatial Program Manager in Missoula. He served as a Peace Corps Volunteer.

This piece was produced in collaboration with Claude, an Anthropic AI assistant. The argument, the sources, and the editorial judgments are the author’s.