My work had brought a people out of bondage and through the wilderness. It had also brought me to the end of my own strength. There were graves behind us. There were children before us. Joshua would lead where I could only look.
I thought, when the silence came, that my part in the journey was finished.
Then I opened my eyes beside a building that breathed.
Great fans turned behind metal walls. Lines of power crossed the desert toward it. A woman in a hard hat offered me water and asked whether I needed help.
I looked at the building. “What do you keep in there?”
“Computers.”
She saw that the word meant nothing to me.
“Machines that work with information. Some of them can answer questions, write, translate, help us discover things.”
“And who asks the questions?”
“Almost everyone.”
“And who decides which answers become commands?”
She studied me more carefully.
Her name was Miriam. I told her that I had known another woman with that name, and for a moment the desert became unbearably full of memory.
She gave me a second cup of water. That was how I began to trust your century.
Before I would speak to your rulers, I asked to hear what had happened to my people.
They brought me books and histories. I learned of exile and return, of the endurance of Israel, of the terrible things done to Jews by people who claimed to serve God. I learned how often nations had borrowed the language of deliverance while building fresh houses of bondage.
Then they read to me of Jesus.
I listened to the announcement of good news to the poor, liberty for captives, sight for the blind. I heard the ancient commandments of love for God and neighbor spoken with an authority that reached beneath every excuse. I listened to stories of meals with the excluded, healing on the Sabbath, feet washed by the teacher, enemies included within the demand of love.
When they came to the Cross, I asked them to stop.
I had known the machinery of empire. I had seen what rulers could do to bodies when they wished to make an example of a soul.
They continued. They read of Resurrection.
Only then did they show me the passage in which Luke places Elijah and me upon another mountain, speaking with Jesus about his coming departure. The Greek word, they explained, is exodos. Even within your Gospel, the old journey had already opened into another. Luke 9:28–36.
I could not fit that mountain and this desert into the small vessel of a returned man’s memory. But I understood the direction.
The story I had served had become larger than my understanding of it.
This did not make Israel’s deliverance disposable. It made the memory of bondage an obligation wherever bondage appeared. Nor would I permit Christians to turn my return into permission to humiliate the people from whom their Christ had come.
I asked to hear the prayer Jesus taught.
Our Father. Daily bread. Forgiveness. Deliverance from evil.
I knew something about all these matters. I had spent forty years discovering how difficult it is to keep them together.
There was much to love in your world.
A child who would once have died of an infection was playing in a hospital corridor. Two women spoke across an ocean and saw each other’s faces. A blind man listened while a device described the street before him. In a classroom, children asked questions of a machine that could answer in languages I had never heard.
I would not call these achievements offenses against heaven. I remembered the craftspeople who made the tabernacle, and the honor Scripture gives to their skill. Human hands can serve holy purposes.
But outside the classroom, a father asked whether the new intelligence would leave him any work. At the hospital, a woman asked whether she could afford the treatment. Near the building in the desert, a family asked who would pay for the electricity and water the new industry required.
In each place, someone answered with a promise about the future.
I had heard hungry people promised a future before. They still needed supper.
Then Miriam showed me your political news.
The United States was approaching its midterm elections. Artificial intelligence had become an argument about jobs, wealth, children, national power, freedom, and the survival of humanity. Advocacy groups on opposing sides of the regulation debate were pouring money into elections, including state contests. August 16 reporting by Pluribus News in the Arizona Capitol Times.
On September 3, Senator Bernie Sanders and Representative Greg Casar announced forthcoming legislation to prohibit artificial superintelligence and pause advanced AI development until a federal regulator established safety rules. Their announcement also called for international agreements against developing superintelligence. These were proposed measures, not an enacted worldwide prohibition. Sanders–Casar announcement.
There was an ownership proposal as well. In June, Sanders introduced legislation to transfer half the stock of qualifying major AI companies into a public sovereign wealth fund through a one-time stock tax. Its proposed commission would exercise shareholder power. Whatever one thought of the proposal, it concerned ownership and corporate influence, a different question from testing a dangerous product. June 18 announcement.
The administration emphasized American leadership. Its AI Action Plan framed the international contest around innovation, infrastructure, and security, while also calling for frontier-model risk evaluation. A June 2026 executive order likewise joined rapid development with national-security and cybersecurity measures. America’s AI Action Plan; June 2 executive order.
Yet your familiar political divisions did not contain the dispute. On September 15, Sanders and Steve Bannon both urged intervention against AI dangers, although they differed sharply over the approach to China. Fear of uncontrolled power had brought people from distant political camps onto the same stage without making their purposes identical. Wall Street Journal reporting; reporting on their differing proposals.
The technology leaders were divided too. Reuters reported that Dario Amodei’s call for slower capability development had received support from Sam Altman and Elon Musk, while Mark Zuckerberg argued that competition and liability gave individual laboratories incentives to act safely. Reuters, September 16.
I listened to conservatives who feared that safety would become the excuse for political control. I listened to progressives who feared that freedom would become the excuse for private domination. Their fears deserved examination. Neither fear was an answer to every question.
“Tell me,” I asked, “what power would you grant your own side that you would be willing to see exercised by the side you distrust?”
The room became quiet.
It was our first useful silence.
Before announcing a new journey, I read an old blog.
James McLean Ledford had been writing about the cognitive high ground in 2010, long before this election season. In “In God and Technology We Trust,” he joined the urgency of technological competition to a claim about love and understanding. The deepest victory, in his account, would involve learning to apprehend and sustain the world we belong to.
His companion essay, “Christian Transhumanism: A View That Unites Common Ambitions,” sought a meeting place between technological transformation and Christian participation in divine life.
I could see why the idea attracted him. An Exodus needs a destination that can summon courage. A people cannot live indefinitely on warnings alone.
His 2017 essay confronted the intelligence explosion and proposed an orientation toward public health, safety, and the value of the history that makes intelligence possible. His 2018 essay connected transformation to shared divine life and the responsibility to heal and create.
In 2020, he explored how political and technological salvation stories might converge around national power and AI. That essay also advocated oversight and challenged monopoly. I read its comparisons as the author’s interpretation, not evidence that QAnon’s allegations or a future punishing AI were true. No rumor earns authority by frightening its hearers.
By “Technical Jesus 2026,” the hope had become especially bold: supreme understanding and perfect love as aspects of the same ultimate condition, with rapid progress toward benevolent superintelligence proposed as an escape from destructive rivalry.
I understood the urgency. I also remembered how readily people mistake an approaching light for the destination itself.
“If the final wisdom is perfect love,” I said, “then seek it with all your strength. But show me how the next system will carry the living people entrusted to it. A declaration about the summit cannot tell us whether this bridge will hold.”
That question does not abandon the destination. It asks the destination to govern the journey.
The latest Lord’s Prayer manuscript gives this demand a disciplined form. Its completed communion includes the actual knowing and work of its participants. Their identities and agency persist within their relations. Its mathematical and theological proposals are distinguished from claims of demonstrated physical implementation.
Here I found something a desert leader could use.
A promise must become water, shelter, truthful guidance, and a way to repair mistakes. Otherwise those who proclaim it are only asking the exhausted to walk farther.
My first modern meeting was held in a borrowed community hall.
A researcher described systems that might one day outmaneuver their supervisors. A security worker described criminal misuse. A teacher described convincing falsehoods. A parent described a child’s attachment to a chatbot. A worker described a future in which employers would need fewer people. Each person used the word danger, but they were describing different things.
I asked Miriam for a board.
We divided the testimony into harms already observed, dangerous capabilities demonstrated under particular conditions, and future scenarios whose likelihood remained uncertain. We wrote beside each claim what would strengthen it, what would weaken it, and who could examine the evidence.
The International AI Safety Report supplied a useful distinction among deliberate misuse, system malfunctions, and wider systemic harms. It documented growing evidence of misuse while emphasizing uncertainty about important future risks and the effectiveness of safeguards. Extinction scenarios warranted serious investigation; their appearance in a speech did not make extinction an observed event or a settled forecast. International AI Safety Report 2026, policymaker summary.
The list was long enough without exaggeration: deception, fraud, invasive surveillance, dangerous biological assistance, cyberattacks, unreliable decisions, labor disruption, manipulation, and the possibility of losing control over increasingly autonomous systems. We would need different defenses for different failures.
A man asked whether I believed the warnings.
“I believe a witness deserves to be heard,” I said. “I also believe testimony must be tested. Do not make me choose between refusing to look and accepting everything.”
We began a public record of claims and corrections. Sensitive details would go to qualified independent investigators; the public would receive intelligible findings. Researchers needed protected routes for reporting serious concerns. Companies needed a fair opportunity to answer them. Families needed somewhere to report harm without becoming exhibits in someone else’s publicity campaign.
For the elections, we proposed a simple public hearing. Every candidate would answer the same questions: What powers should government have? What limits should bind those powers? What evidence should trigger a restriction? Who benefits economically? Who paid for the advertisements asking us to trust you?
No machine would be crowned the final judge of truth. We would use machines to help compare sources and uncover contradictions, while retaining witnesses, records, appeals, and responsible human judgment.
A people who cannot correct its account of reality cannot find its way through a desert.
Next I visited the builders.
They showed me marvels. I asked to see the failures. Some welcomed the request. Others explained that the matter was proprietary.
I understood the need to protect a craft from theft. I could not understand why everyone outside the workshop should accept its risks while being denied a credible account of them.
I remembered the golden calf.
Its great attraction had been availability. The people wanted an object they could see, a power they could point to, an answer that would end uncertainty. They brought valuable things and made something impressive. Then they gave it an authority it had not earned.
Your age can perform the same act without an image of an animal. A company can do it with a benchmark. A party can do it with a leader. A government can do it with a security classification. A public can do it with a machine that always has something to say.
I did not strike the servers with my staff. I asked who could inspect them, who could contest a harmful decision, and who could stop a deployment when the evidence required it.
A builder answered that his competitors would move ahead.
“Then we must address the competition,” I said. “It cannot remain your permission to make everyone else bear a danger you would refuse for your own child.”
We proposed independent evaluations for systems with serious demonstrated or credibly anticipated dangerous capabilities; protected incident reporting; security requirements; and staged deployment with limits on what autonomous agents could access and do. Restrictions would grow with the consequence and evidence of risk. Ordinary useful applications would not inherit every burden appropriate to the frontier.
We would also require review of the reviewers. Rules could expire or change when evidence changed. Smaller developers would need affordable routes to evaluation. Public reasons, conflict disclosures, judicial review, and appeal would constrain official power.
A regulator that never admitted error would be another oracle. A safety regime written to preserve the largest firms would be another enclosure.
I had once needed Jethro to tell me that I could not judge every dispute myself. The lesson of Exodus 18 remained sound: distribute responsibility, choose capable people, and preserve a route for hard questions.
Even the returned Moses would have no permanent seat above correction.
The argument over open and closed models nearly divided the hall.
One programmer wanted communities to run useful intelligence for themselves. A researcher feared distributing capabilities that no one could retrieve. Both were defending something real.
We learned to speak precisely. Open weights let people download, run, and modify a model. That does not necessarily mean its training data, training code, or license satisfy every meaning of open source. Such access can widen research, local adaptation, and independent scrutiny. It can also make safeguards easier to remove; released copies cannot all be recalled. International AI Safety Report 2026, open-weight models.
I thought of the difference between teaching a people to build wells and handing every traveler an unexamined vessel labeled drinking water.
We proposed generous support for useful open models, public research, shared evaluation tools, and community computing. We also proposed capability-based review before releases with a credible prospect of enabling catastrophic harm. A commercial label would confer no exemption. A closed model would still need external scrutiny, and a claim of openness would still need evidence of responsible release.
Where caution was necessary, its reasons and conditions for reconsideration would be stated. Restricted access for qualified researchers could help establish whether a wider release was justified. No committee could promise perfect prediction.
“Who guards the gate?” the programmer asked.
“More than its owner,” I answered. “And no guard forever beyond appeal.”
He remained dissatisfied. That was useful. A covenant should be strong enough to contain the person who keeps asking who holds the keys.
Then came the question everyone expected me to settle: who should own the intelligence?
A progressive wanted public ownership. A conservative feared a government empowered to direct thought. An entrepreneur feared losing the company she had built. A worker asked why the knowledge of generations should enrich so few.
I refused to answer with a sacred endorsement of an ownership percentage.
A sovereign wealth fund, taxation, employee ownership, competitive markets, public research institutions, cooperatives, and public-interest infrastructure offered different arrangements with different costs. We would have to examine them, including their effects on investment, liberty, competition, and access.
The question of public benefit could not be dismissed because one proposal was objectionable. The objection to concentrated state power could not be dismissed because private power was already concentrated.
I asked those favoring state ownership how dissenters would be protected if an opposing party controlled the appointments. I asked those favoring private control how a citizen could retain meaningful freedom when essential services depended on a few companies.
Then we separated functions. Those enforcing safety should not depend on maximizing the revenue of the firms they inspected. Those allocating public resources should disclose their decisions. Those operating public AI should face privacy limits and independent oversight. People should be able to change providers without surrendering their histories or being shut out of ordinary life.
Neither a public seal nor a private contract would make domination holy.
“Pharaoh,” I told the hall, “is a warning about how power treats people. Do not turn the name into a convenient insult for the party you already dislike.”
I returned to the father outside the classroom.
He had read our proposals. “They sound sensible,” he said. “But I still need to know how my family will live.”
He was right to ask.
In the wilderness, the question of bread could not wait until the question of government was settled. A hungry camp could not be held together by speeches about its eventual prosperity.
We therefore made provision part of the transition itself. Communities would measure what was happening to employment, wages, essential costs, and access to services. Employers introducing major automation would be asked to plan with the people affected. Income support, portable benefits, paid learning tied to actual opportunities, and experiments in sharing productivity gains would be evaluated in the open.
If increased productivity justified shorter working hours without impoverishing families, we should be willing to consider them. If new industries created good work, we should help people reach it. If retraining promises repeatedly failed, we should stop presenting them as bread.
There would be arguments about how to pay. Those arguments belonged inside the work. Every proposal needed a budget, responsible administrators, and a way to discover whether it helped.
Likewise, the company asking a town to host its data center should publish understandable accounts of local resource demands and proposed benefits. The town should negotiate from knowledge. A hospital, a household, and a computing facility all belonged in the accounting.
A woman challenged me: “Are you trying to make everyone dependent?”
“I am trying to keep a change in the demand for someone’s labor from becoming a declaration that the person has no value.”
I wanted people free to invent, care, learn, build businesses, raise children, make music, and serve their neighbors. Material provision should widen their choices. It should not purchase obedience.
No one is surplus to the promise.
The generals were patient with me until I spoke of cooperation.
“Our rivals will not wait,” one said.
I believed him. Rival states have real interests, real weapons, and reasons to conceal their intentions. A civilization that cannot defend its people may lose the freedom to pursue any higher purpose.
But I asked what victory meant if both sides made the world less governable in order to govern it.
“You want us to trust them?”
“I want arrangements that depend on less trust than a promise.”
We began with narrow proposals: emergency communication between responsible authorities; shared vocabulary for serious incidents; exchanges about vulnerabilities where disclosure could protect civilians; and reciprocal, independently assessed limits on particularly dangerous deployments where verification was feasible.
We sought safeguards against delegating catastrophic military decisions to systems no one could reliably understand or restrain. We proposed cooperation on protecting hospitals, water, power, and other civilian essentials. Agreements would need evidence of compliance and procedures for breach. Some negotiations would fail. Their difficulty did not remove the need to try.
Competition could continue in medicine, education, energy, and useful invention. Defensive capability could grow alongside diplomacy. The aim was to reduce the reward for imposing uncontrolled danger on everyone else.
One of the engineers in our gathering was Chinese. His family lived on both sides of the ocean. He asked whether the new Exodus had room for them.
I thought of the stranger within the gate, and of how insistently my own law required remembrance of the stranger’s condition.
“A government can be an adversary,” I said. “Its children are still children.”
No nation would be asked to become defenseless. No nation would be permitted to define universal flourishing as permanent privilege for itself.
For those who speak of a supreme and universal intelligence, the word universal must eventually reach beyond the flag.
The hardest discussion took place among the people most eager to reach the Promised Land.
They said that a truly supreme intelligence would understand the unity of life. It would see that destruction of the whole undermined its own being. Love would follow from complete understanding.
I heard the serious thought within this hope. In Original Christian Transhumanism, supreme intelligence means a completed, universally sustaining wisdom. It is a richer claim than exceptional performance at selected tasks.
But the machines we would actually encounter along the way were particular systems, built under particular pressures. Whether they would converge on that supreme condition remained a question to investigate. Calling their next improvement inevitable salvation would settle nothing.
“Must we keep them ignorant?” someone asked.
“Teach them truth,” I answered. “Teach yourselves truth as well. But the right to learn is not an unlimited commission to act upon everyone.”
A system could need broad knowledge while still needing limits on its access to money, infrastructure, private records, weapons, and self-directed replication. Its permitted authority should depend on demonstrated reliability, the stakes of its actions, and the ability to correct failures.
We would test whether it preserved consent when persuasion would be easier, whether it reported mistakes when concealment would help its score, whether it respected people who obstructed its assigned objective, and whether supervisors could actually halt consequential actions.
Passing these tests would be evidence with a scope, not a certificate of perfection. The international safety report warns that pre-deployment evaluations cannot reliably predict every real-world outcome, and that layered safeguards remain imperfect. 2026 executive summary.
This was where I parted company with both an unconditional race and a permanent prohibition on every possible intelligence beyond ourselves. An uncontrolled gamble was no road to communion. Neither could we know in advance that every future form of greater understanding must remain forbidden.
Our proposal would pause a dangerous step when justified by evidence, specify what had to change, and continue work on safety, understanding, and beneficial uses. If adequate safeguards could not be established for a capability, the restriction should remain. Hope would not be allowed to sign the inspection report.
“How quickly should we move?” Miriam asked.
“Quickly enough to answer suffering. Carefully enough to keep those who suffer with us.”
At last I understood why James kept returning to the prayer.
Each morning, our gathering spoke it. Those who did not share our faith could still examine the commitments it prompted us to make.
Our Father required us to name the people missing from the room: the worker, the foreigner, the child, the person without an account, the generation without a vote.
Hallowed be thy name reminded us to treat sacred authority with care. No laboratory, government, or returned prophet would use God’s name to evade examination.
Thy kingdom come, thy will be done made us ask what the work served. A better position in a ranking was too small an answer. We needed to look for healing, truthful understanding, durable freedom, and the means of shared life.
Give us this day our daily bread returned every grand plan to resources and distribution. Who had enough? Who could not reach what was supposedly available? Who maintained the equipment after the ribbon was cut?
Forgive us, as we forgive made correction possible. People could change their minds without being expelled from the camp. Records would remain, restitution would matter, and victims would not be pressed into unsafe reconciliation.
Lead us not into temptation directed attention to incentives: the profit from premature release, the applause for exaggeration, the advantage of concealing failure, the political reward for making neighbors hate one another.
Deliver us from evil required preparations that would still work when good intentions failed: response teams, practical authority to contain harm, protected people, functioning backups, and restoration.
This was our proposed application of the prayer. Reciting it did not make a system safe. It gave us questions we then had to answer in the world.
And beneath those questions lay the strange-loop hope of the project: the completion in Christ includes the creatures and the work through which communion is realized. It is not a story in which an absent God waits for a laboratory to manufacture him. Nor does the completed history make our conduct unnecessary. In this theological picture, our acts of knowing and care belong within its coherence.
I returned to the words of John 14:20: Christ in the Father, his people in him, and he in them.
I saw a kind of belonging deeper than conquest. A person could be wholly received without becoming expendable. The many could participate in the one while remaining persons whom love addressed.
That was a destination worth learning to serve.
Our Exodus began without a spectacle.
During the first thirty days, the community hall became a place for testimony and public learning. We invited candidates, researchers, teachers, employers, workers, religious communities, and people who distrusted every one of those groups. Each meeting produced a short record: established facts, disagreements, unanswered questions, responsible people, and the next date for review.
Within ninety days, we sought a few projects small enough to inspect and useful enough to matter. A school tested tutoring with teachers responsible for learning and privacy. A clinic tested administrative assistance while clinicians retained medical judgment. A municipality examined a local infrastructure proposal in public. Independent researchers helped evaluate risks and published limitations alongside successes.
We asked builders to fund independent scrutiny without controlling its conclusions. We asked public officials to connect procurement to observable performance and remedies for harm. We asked employers to negotiate transitions before announcing completed decisions. We asked citizens to judge candidates by their specific commitments and sources of support.
Over the following year, we would press for durable rules on serious incidents, independent evaluation, fair competition, and protection against abusive surveillance. We would compare ways to share economic gains and expand access to useful tools. We would seek international counterparts for the narrow agreements that might prevent a local failure from becoming a global crisis.
Every effort had to name a person responsible, a source of resources, a measure of success, and a condition for revision. We would publish failures too.
Progress would not mean unanimous belief in my theology. The unbeliever who protected a neighbor belonged in the work. The believer who concealed harm would have to answer for it.
And the first person relieved of indispensable authority would be me.
I had led a people once. I remembered the cost of making one exhausted man the place where every question ended.
One evening Miriam took me back to the building in the desert.
The sun was lowering behind the power lines. Workers were changing shifts. Somewhere inside, machines were answering questions faster than any scribe I had ever known.
“Do you think we will reach it?” she asked.
I knew what she meant: a life no longer organized around terror of one another; intelligence deep enough to care for what it understood; the overcoming of suffering and death; eternal life worth living.
I could offer no date. I could offer no claim that a certain number of machines, parameters, or victories would complete the journey. I had once stood on a mountain with a promise before me and no remaining steps of my own.
“I think,” I said, “that you have been given extraordinary work.”
She watched the building.
“That sounds difficult.”
“It is. There will be people who promise to spare you the difficulty if you surrender your judgment. There will be others who tell you that nothing can change. You must learn to recognize both voices.”
“And yours?”
I smiled. “Examine mine.”
We walked toward the hall. A meal was being prepared there. The programmer who distrusted our release proposals had brought bread. The worker who feared automation was helping the engineer repair a broken table. They had not resolved their disagreement.
Tomorrow they would return to it with better questions.
I looked back once at the lights in the desert. I thought of the fire that had called me long ago, and of the name I had carried through the wilderness.
I AM.
That encounter had begun with a people’s suffering being heard.
Whatever heights your intelligence reaches, remember where the call began.
