As of last week, any biotech company can pay $12,500 to have an experimental drug reviewed by a private board in Montana, and if approved, sell it directly to consumers at any price it likes, after testing that may have involved as few as 20 healthy volunteers (Jessica Hamzelou, MIT Technology Review, July 30, 2026). This matters because it recreates something specific and historical: a market of active compounds sold direct to the public, with honest disclosure and no required proof of safety or benefit, which is exactly the American drug regime between 1906 and 1938, the one we replaced after it killed people at scale. The caveat worth holding onto: the federal Right to Try Act has produced 21 eligible products in seven years while the FDA's expanded access program clears roughly 2,000 submissions a year at better than a 99 percent approval rate, so the thing this law removes was never the actual bottleneck.

Jessica Hamzelou, senior biotech reporter at MIT Technology Review, which is editorially independent and owned by MIT. She has covered this story since the bill passed in May 2025, which shows. This is reporting, not research, so the usual questions about study design do not apply. The questions that do apply are about sourcing, framing, and whose numbers made it in.

No disclosed funding conflicts for the outlet or the reporter. The conflicts in this story belong to the people in it, and they are not incidental. They are the story.

• Infinita. The review board was announced by Infinita, the venture firm founded by Niklas Anzinger, who previously built Infinita City inside Próspera, the private charter city on Roatán, Honduras, where a clinic sells follistatin gene therapy for roughly $25,000. Anzinger helped draft the Montana bill. He now runs the entity that reviews applications under it. Board members are paid from the $12,500 fees.
• The board. Of the six listed members, at least four are embedded in the longevity industry the law serves. Felipe Sierra is chief scientist at Hevolution, the Saudi funded longevity nonprofit. Jamie Justice runs the $101 million XPRIZE Healthspan, which Hevolution co-funds. Jessica Flanigan holds the ethicist seat and is the author of Pharmaceutical Freedom, the definitive philosophical argument for exactly this policy. Matt Kaeberlein is the most scientifically credible name on the list and the one critics will say launders the rest.
• The bill's origin. Drafted with help from the Alliance for Longevity Initiatives, a 501(c)(4) advocacy group, and from unnamed biotech companies. Representative Kathy Love noted during floor debate that the proponents testifying were primarily biotech industry, not private citizens.

Straight beat reporting on a regulatory milestone. Hamzelou tracked the Montana Department of Public Health and Human Services rules from proposal through adoption, got the announcement of the first review board, and interviewed people on both sides: Anzinger and Stephen Martin building the thing, board member Kaeberlein, two biotech CEOs deciding whether to apply, and three critics (Aaron Kesselheim at Harvard, Chris Robertson at Boston University, and the FDA, which declined). She asked the FDA for comment and printed the non answer, which was the right call.

The machinery exists now. Twenty five new rules, ARM 37.106.3301 through 3325, took effect July 25, 2026, after a comment period the department answered in 123 numbered responses, most of them declining the change requested, citing rural access. A private board calling itself the Montana ETRB is open for business under Montana Governance Services Inc., which sits under the Infinita umbrella. Two applications are in. One is WinSanTor, whose CEO says he will sell his phase 2 peripheral neuropathy drug at cost and hopes to collect data that speeds his actual approval. The other treats hearing loss.

The two facts that carry the piece are pricing and eligibility. Under federal expanded access, a company may charge only the cost of making, shipping, and monitoring the drug, and must justify that to the FDA. In Montana it can charge whatever the market bears. And Montana access is not limited to the seriously ill. A person interested in living longer qualifies on the same terms as a person with glioblastoma.

The most telling quote is not from a critic. Thomas Joudinaud, CEO of Ceres Brain Therapeutics, called Montana's system "very interesting and very pragmatic" and "suitable for our drug," then declined to apply because he is afraid of what the FDA will do to him later. Martin asked the agency for assurance. The agency restated the statute. Robertson's warning is the sharpest line in the article: do not bet on anything the FDA says today applying when the rubber hits the road later.

• It found the seam. The pricing contrast between expanded access cost recovery and Montana's open pricing is the single most important structural fact about this law, and Hamzelou got it in cleanly.
• Real sourcing on both sides. Kaeberlein's middle ground argument is presented as a serious position rather than a foil, which it deserves to be. It is the strongest thing the pro side has.
• It reported the reluctance. A biotech CEO who likes the law and will not use it tells you more about the law's actual prospects than any critic quote.
• It is close to alone. As of this writing I could not find dedicated feature coverage of the July 2026 rules from STAT, the Post, the Times, Nature, Science, or KFF. The most consequential drug regulation story in the country is being covered by one tech magazine and the longevity trade press. That is itself a finding.

• The comparison that would have ended the argument is missing. The article never gives the utilization numbers. Federal right to try has covered 12 products from 2018 to 2022, 4 in 2023, and 5 in 2024, with zero contributing clinical outcomes data to any FDA determination. Expanded access, meanwhile, took in 2,022 submissions at CDER in fiscal 2023 and allowed 2,010 to proceed, 99.4 percent, usually in under 48 hours and immediately for emergencies. A crowdfunding study by Snyder, Bateman House, and Turner found 21 of 26 expanded access campaigns reported getting the drug, versus 1 of 29 right to try campaigns. Kesselheim gets to say the FDA is not a bottleneck. The reader never gets to see it.
• One number is stated loosely. The article says roughly 17 percent of drugs are found inadequately safe during phase 3. The underlying paper (Hwang, Kesselheim et al., JAMA Internal Medicine, 2016) found that of 344 late stage failures, 59 were safety failures. That is 17 percent of the failures, or 9.2 percent of the 640 drugs entering pivotal trials. The precise version is still alarming. Roughly one in eleven drugs that had already cleared phase 1 and phase 2 turned out to be unacceptably unsafe. Montana sells after phase 1 alone.
• A sourcing error. The piece credits state senator Ken Bogner with the 2015 original law. Bogner was not elected to the Montana Senate until 2018 and took office in January 2019. He sponsored the 2023 expansion (SB 422) and the 2025 treatment center law (SB 535).
• A price figure that drifted. The article cites $218,872 as the median price of rare disease drugs. The underlying figure is an average annual cost per patient for drugs carrying an orphan indication, against $12,798 for non orphan drugs. Still a useful benchmark for what Montana pricing could look like. Not a median, and not restricted to recent launches.
• Board size discrepancy. The article describes a five member board. The ETRB's own site lists six, adding Danielle Ruiz, a Montana nurse practitioner and Methuselah Foundation fellow.
• The opposition it missed. The most interesting critics of SB 535 were not public health officials. The Montana Medical Freedom Alliance, a health freedom group, opposed the bill as too vague to be safe or ethical, and objected to liability protections they compared to vaccine manufacturer shields. The medical freedom movement split over the most expansive medical freedom law in the country. That wrecks the clean partisan read, which is why it is worth having.

The paragraphs above lean on the MIT summary. This section leans on the primary source: the adopted Montana rules themselves (ARM 37.106.3301 through 3325, effective July 25, 2026), which create a new license class called an Experimental Treatment Center and the department's answers to 62 public comments. There are two kinds of clinic, split only by how long a patient stays: outpatient (under 24 hours) and inpatient (24 hours or more). Both can perform surgery, administer anesthesia, and operate inside an existing hospital. What follows is the oversight each is required to carry, and, in the third column, the seam a bad actor could work.

Here is the part of your read that inverts the usual comfort. Patent medicines were dangerous precisely because they contained real drugs. Mrs. Winslow's Soothing Syrup carried something like 65 milligrams of morphine per ounce and was used to quiet infants. The AMA's 1911 catalog of these products had a section titled, without irony, "Baby Killers." The genuinely inert nostrums were the harmless ones.

So the observation that these Montana compounds are probably pharmacologically active is the risk statement, not the mitigating one. Anything that clears preclinical work and enters phase 1 is bioactive by construction. The molecule doing something was never in question. Whether anyone knows what it does at scale is the entire question, and after phase 1 the honest answer is no.

Phase 1 asks whether a healthy volunteer tolerates a dose. The FDA puts typical enrollment at 20 to 100 people. It is not designed to measure whether the drug works, and it is not designed to catch anything uncommon. The arithmetic here is unforgiving, and it has a name.

Hanley and Lippman Hand, in JAMA in 1983, gave us the rule of three. Observe zero bad events in a group of n people, and the true rate could still be as high as roughly three in n. Run a clean phase 1 in 20 volunteers and you have not ruled out a serious harm rate of 15 percent. A drug that badly injures one person in ten will still show a clean sheet in a group of ten about a third of the time. Absence of a signal in a small group is not evidence of safety. It is evidence of a small group.

Then the ceiling case, to show that this is not just a small numbers problem. Vioxx was studied in VIGOR, a trial of 8,076 patients. The cardiac signal was already in that data, and got read as the comparison drug being protective rather than Vioxx being harmful. It took five more years and something like 20 million American users to settle the question. The Lancet estimate was 88,000 to 140,000 excess serious coronary events, roughly 44 percent of them fatal. Eight thousand patients in a controlled trial was not enough resolution. Phase 1 offers twenty in an uncontrolled sale.

And the general pattern, aimed straight at Montana. Onakpoya, Heneghan, and Aronson catalogued 462 medicines withdrawn for safety between 1953 and 2013. The evidence that finally pulled them was anecdotal case reports in 72 percent of cases. The median lag from the first reported harm to withdrawal was six years, and it did not shorten across six decades of trying. If the signal usually arrives as scattered case reports and takes six years to act on inside a system built to catch it, Montana has neither the system nor, given who collects the fees, the incentive to look.

Which is why the disclosure does less than it appears to. You can disclose that a drug is experimental. You cannot disclose a risk whose size nobody has measured. Informed consent moves liability from the seller to the buyer. It does not move information that does not yet exist. A signature collected in that state of ignorance is not a transfer of understanding. It is a receipt.

This is also the honest answer to Flanigan, whose argument deserves better than a wave at paternalism. A randomized trial is not primarily a protection. It is an instrument, and its job is to make causation visible against the noise of hope, self limiting illness, and disease that varies on its own. Distributed individual purchase destroys that instrument's resolution. Not for the buyer, who was going to get their outcome regardless, but for everyone downstream who needed to know whether the thing worked. That is a collective good no individual can buy at any price, and it is the piece of the case that autonomy does not answer.

Your larger point is that this represents a drift toward do it yourself healthcare, of a piece with the "do your own research" reflex. It does, and the trust data backs the direction more sharply than I expected. In KFF's January 2025 polling, 85 percent of Americans trusted their own doctor for health information, against 53 percent for the FDA. Republicans trust their own doctor at 84 percent, essentially the national number, while the partisan gap on the CDC runs 46 points. Annenberg found the same fracture inside the agencies: 67 percent confidence in career scientists, 43 percent in agency leadership. The relationship survived. The institution did not.

But the do it yourself frame has a flaw built into its slogan, and it is worth naming plainly. Research is not reading. Research means running a comparison, and a single person cannot run a control arm on themselves. What individual experience produces instead is testimony, and testimony has a known and ancient failure mode. Samuel Hopkins Adams described it in Collier's in 1905, hunting these same patent medicines.

The modern taxonomy (Hartman, 2009) just formalizes what Adams saw: it separates the improvement a treatment did not cause, which is regression to the mean and placebo and the illness resolving on its own, from the improvement that was never real, which is post hoc reasoning and confirmation bias. The do your own research crowd is not lazy. It is doing the wrong operation with real effort. Gathering, when the work is in aggregating. And aggregation is exactly the collective function that Montana dissolves and sells back one patient at a time.

One thing from last week survives intact and is worth keeping. If the story were simply the individual reclaiming power from the government, business should have left the picture. It did not. Montana built for profit clinics, unlimited pricing, a $12,500 gate fee, and a 2 percent profit set aside that functions as an indulgence. The same structural relationship, a company selling an unproven product to a frightened person for money, has not changed. It has only been recast, from the thing the little guy needs protection from into the ally helping the little guy escape the state. The Montana Medical Freedom Alliance saw this and opposed the bill as industry capture wearing freedom's clothes. The people most fluent in medical freedom rhetoric rejected the most expansive medical freedom law in the country.

The timing is almost too neat. On July 29, one day before this story ran, Katelyn Jetelina argued that it is time to stop blaming Anthony Fauci for the pandemic. Her case is that concentrating six years of national anger on one man is the mechanism by which the institutions that actually made the decisions, the school boards and employers and legislatures, avoid ever being reviewed. Her proposed remedy is a 9/11 style commission. Systematic, aggregated, adversarial accounting.

Read the two pieces together and they are the same event from opposite ends. In one, a country replaces its trust in institutions with trust in its own doctor. In the other, it replaces its anger at institutions with anger at one official. Adore the physician, revile the agency. Blame the man, spare the system. Both moves swap an institution for a person, and both let the institution slip the only kind of evaluation that could actually fix anything. Jetelina is asking for exactly the thing Montana abolishes: judgment pooled and reviewed, rather than distributed and sold.

The other side, sorted by how seriously to take it.

• Samuel Hopkins Adams, "The Great American Fraud," Collier's Weekly, 1905. Full text at Gutenberg. The do your own research problem, diagnosed before the FDA existed.
• FDA history of the 1906, 1938, and 1962 laws. fda.gov. The Radithor detail is in the 80 Years of the FDCA exhibit.
• Lewis Grossman, Choose Your Medicine (Oxford, 2021). Two hundred years of American therapeutic choice claims. The corrective to any "this is new" instinct.

• Hanley and Lippman Hand. "If nothing goes wrong, is everything all right?" JAMA 1983;249(13):1743-5. The rule of three, at the source.
• Onakpoya, Heneghan, Aronson. "Post marketing withdrawal of 462 medicinal products because of adverse drug reactions." BMC Medicine 2016;14:10. DOI.
• Graham et al. Rofecoxib and serious coronary heart disease. Lancet 2005;365:475-481. DOI. The Vioxx toll.
• Hartman. "Why do ineffective treatments seem helpful?" Chiropractic & Manual Therapies 2009;17:10. DOI. The modern taxonomy of why uncontrolled experience misleads.

• CRS Report R45414, "Expanded Access and Right to Try." The fastest way to understand how the two pathways differ.
• FDA's Right to Try annual reporting summary. The utilization numbers, from the primary source. Short, and devastating.
• Snyder, Bateman House, Turner. "Is right to try being tried?" Regen Med 2020. DOI. 21 of 26 versus 1 of 29.

• The adopted rules, MAR 2026-427.2, effective July 25, 2026. If you read one thing, read this.
• SB 535 as enrolled. The eligibility language shifted from "considered all" other approved options to "evaluated" them.
• montanaetrb.org. Read the disclaimers.

The FDA convened an advisory panel in July 2026 on easing restrictions on compounded peptides, and the agency's own scientists publicly warned against it. Same deregulatory pressure on unproven enhancement products, different direction, same month. Montana is not an isolated event. It is the state level version of something happening federally.