Before Vinay Prasad, the UCSF oncologist, served as head of CBER at FDA, he was famously outspoken—in blogs, vlogs, podcasts, and peer-reviewed articles.
During his time at FDA, those informal channels went quiet.
Now that he is out of FDA, he appears to be back at full volume. His YouTube channel, dormant for a year, has posted nine new vlogs in a month, most running 20 to 30 minutes.
https://www.youtube.com/@vprasadmdmph/videos
https://www.drvinayprasad.com/
How did life in Washington change him? I took his eleven new articles and his last eleven from Spring 2025 before his hiatus, and asked for an AI comparison. Embarrassingly, Claude counts "13" new articles; that's not central so I've left as-is.
SUMMARY
The two analyses below ask the same question: what changed when Vinay Prasad returned to public writing after his turbulent year at FDA’s CBER? The first, from Claude, is more quantitative and biographical, tracking article counts, length, citations, profanity, topic mix, and the conspicuous silence around his FDA controversies. The second, from ChatGPT, is more interpretive, arguing that Prasad’s center of gravity shifted from culture-war combat toward clinical epistemology: weak evidence, observational bias, overdiagnosis, medical AI, journals, and institutional incentives. Together, they both conclude that Prasad is out of power, but not out of influence.
##
Claude Sonnet 4.6
(Claude counts 13 new pieces, I had couned "11," this is not central, and I've left as-is.)
##
Same Prasad, Different Beat
A quantitative read on Vinay Prasad's Substack, before and after he ran FDA's CBER
When Steve Bannon goes dark, his audience knows the drill: he comes back narrating every grievance, every betrayal, every vindication. Vinay Prasad's return to Substack after running FDA's Center for Biologics Evaluation and Research follows a different script. His tenure was not quiet. He resigned once, abruptly, on July 29, 2025, after barely three months, amid a backlash from rare-disease advocates and a Trump-aligned influencer over a dispute involving a Duchenne muscular dystrophy gene therapy — then was reinstated two weeks later after FDA Commissioner Marty Makary intervened. He came back more combative, pushed out at least seven agency leaders, and by his second and final exit — announced March 6, 2026, effective at the end of April — was operating under an EEO investigation into complaints that he retaliated against subordinates and berated staff. Makary's public framing was that this was simply the tidy conclusion of a planned one-year sabbatical from Prasad's UCSF faculty post. None of the turbulence shows up in his first thirteen posts back. What does show up is more interesting, and more useful, than a grudge match would have been.
This essay compares his last eleven Substack posts before joining CBER (March 29–May 4, 2025) against the first posts published after he left (May 12–June 16, 2026). A methodological note: the source document for the return period labels itself "11," but a numbering glitch in the original sequence means the true count of distinct, dated entries is thirteen — itself a small, fitting irony for a body of work this preoccupied with getting numbers right. I coded each post by hand for length, citation density, profanity, and primary topic (one topic per post, by dominant theme) directly from the text.
| Metric | Last 11, pre-CBER (Mar 29–May 4, 2025) | First posts, post-CBER (May 12–Jun 16, 2026)* |
|---|---|---|
| Posts analyzed | 11 | 13 [?] |
| Span / pace | 37 days, ~1 every 3.4 days | 35 days, ~1 every 2.7 days |
| Average length | 524 words | 690 words (+32%) |
| Longest / shortest post | 939 / 173 words | 1,380 / 281 words |
| Profanity ("f*ck," "sh*t") | 7 instances across 2 posts | 0 |
| External, non-Substack citations | 2 | 21 |
| Topic: culture-war & politics | 4 of 11 (36%) | 2 of 13 (15%) |
| Topic: media/journal criticism | 2 of 11 (18%) | 0 of 13 (0%) |
| Topic: evidence & regulatory science | 3 of 11 (27%) | 6 of 13 (46%) |
| Topic: health-tech/AI & industry | 0 of 11 (0%) | 2 of 13 (15%) |
| Topic: personal/clinical narrative | 2 of 11 (18%) | 3 of 13 (23%) |
| Paywalled ("Paid") posts | 1 of 11 | 2 of 13 |
| Explicit subscribe asks | 2 | 7 |
*The source labeled this batch as "11"; but a duplicated index number and one unnumbered post mean the true count is 13.
The headline numbers
He came back wordier, footnoted, and quieter. Posts are 32 percent longer on average, and his longest return post — a 1,380-word account of an alleged surgical error at OHSU — runs nearly 50 percent past his pre-CBER ceiling.
The citation profile changed even more than the length: the spring 2025 batch carried exactly two links to anything outside his own Substack, while the return batch carries twenty-one, spanning NEJM, JAMA Network, Annals of Oncology, half a dozen PubMed and PMC links, the American Journal of Medicine, three YouTube videos, a SoundCloud podcast, three cross-posts to his Sensible Medicine collaboration, and — in one case — a Reddit thread where someone posted the legal filings underlying a malpractice story.
Profanity disappears entirely: seven instances of "f*ck" and "sh*t" across two spring posts, zero across the return batch. He is also working the subscription mechanics harder, with explicit asks more than tripling, consistent with someone rebuilding a paid audience after roughly ten months of enforced silence — his own Substack's "About" page stated plainly that the site would go inactive during his federal service, and it did.
THe Topic Shift
In 2026, reduce the culture wars and media criticism. Up with evidence and health tech.
Audition tape: the last eleven before CBER
Read in sequence, the spring 2025 posts behave like an extended cover letter for the job he was about to take. Nine of the eleven touch partisan politics or media criticism in some direct way; only a movie review (of Wim Wenders' Perfect Days) stays mostly clear of it, and even that detours into a riff calling news-anxiety "a mental illness."
The throughline is a defense of the incoming administration's direction at HHS and FDA: he cheers the prospect of pulling COVID-19 vaccination off the childhood schedule ("Good!!!"), argues for placebo-controlled trials before further boosters, and defends NIH and DEI funding cuts at length. Two posts are devoted to predicting that Harvard will capitulate to the Trump administration's funding pressure, written six days apart, both with the same prediction restated more confidently the second time. A media-literacy listicle doubles as an enemies list, naming STAT's Helen Branswell, former CDC director Mandy Cohen, and epidemiologist Michael Osterholm individually as examples of "biased and weak" coverage, while defending Stanford's Jay Bhattacharya from a quote he calls misconstrued. Profanity shows up exactly where you'd expect — a riff on seed oils and food additives, and a swipe at university DEI policy.
The closing entry is the most pointed. His literal last post before joining the government, dated March 29, 2025, is a lengthy takedown of Peter Marks — the outgoing CBER director whose chair he was about to take — accusing him of approving Sarepta's Duchenne gene therapy against the advice of FDA scientists and, in Prasad's own framing, having "indirectly killed more boys with DMD than he has saved."
Whether or not the timing was deliberate, the essay reads, in hindsight, like the closing argument of a confirmation hearing nobody was asking for.
The professor comes home, 2026
The return batch reads almost like a different writer, not because the voice changed but because the subject matter did. Four of the thirteen posts are dense methodology pieces in the register that made him academically famous a decade before any of this: a post-publication critique of a JAMA Internal Medicine observational study on COVID boosters, flagging that the all-cause mortality curves separate too fast to reflect anything but an unmatched control group; a parsing of why a new Annals of Oncology paper on GLP-1 drugs and cancer risk conflicts with longer-term randomized data; a structural piece comparing multiple myeloma and chronic myeloid leukemia that walks through lead-time bias, diagnostic-criteria creep, and the difference between survival from diagnosis and population-level mortality; and a four-point appraisal of the Daraxonrasib pancreatic-cancer trial presented at ASCO, flagging a 12-point differential dropout rate between arms as a quiet violation of randomization. None of this is partisan. It is the exact toolkit — surrogate endpoints, medical reversal, observational-versus-randomized evidence — that built his reputation as an oncology biostatistician before COVID politics ever entered the picture, and it is back with real intensity.
Two new posts read like biotech market notes rather than culture-war columns: one arguing that EMR control, not accuracy, will determine which medical AI tool wins between OpenEvidence, ChatGPT, Gemini, and Epic; the other dismissing a freshly published AI-comparison paper as obsolete the moment it printed, because the underlying models had already moved on.
And the single most viral-shaped piece of the batch is not political at all: an account, sourced to legal filings reposted on Reddit, of a 13-year-old's mitral valve allegedly sewn in upside-down at Oregon Health & Science University. He uses it to relitigate diagnostic groupthink and palliative-care overreach, and drops the only line in all thirteen posts that acknowledges he was ever gone — a passing reference to "my departure for government service," offered with no further elaboration.
Politics without the profanity
Politics hasn't vanished, it's just been demoted from headliner to opener. The closest analog to the spring's Harvard pieces is a single post on a DOJ finding that UC Davis and Yale violated the Supreme Court's affirmative-action ruling in medical-school admissions — same genre, same approving citation of a Trump administration enforcement action, just one entry instead of four. A second piece praises Jeff Bezos's framing of government waste. A third frames a newly published alcohol-health study as commissioned under Biden and "buried" under Trump, then spends most of its length on methodology before concluding that "reasonable scientists can disagree." That softer landing — methodological argument first, partisan applause line dialed down — is the pattern across the board. Even the one post that revisits a Biden-era figure by name, a retrospective on Paxlovid that blames former COVID czar Ashish Jha for overselling the drug outside its tested population, resolves into a discussion of placebo design and immortal-time bias rather than a culture-war jab.
One asymmetry holds steady across both eras: he does not criticize a Republican or Trump-administration figure by name in either window.
He praises DOJ civil-rights chief Harmeet Dhillon, Bezos, and Mehmet Oz; he criticizes Branswell, Cohen, Osterholm, Harvard's Alan Garber, and Jha. Coverage of Prasad himself shows the same split-screen quality he likes to write about: The Free Press runs a standing tag promoting him as essential reading, while a rival health-policy Substack calls him an "anti-vax authoritarian lackey." Neither tells you much about the underlying work, but both are a useful reminder that the media-bias complaint he's built a career on cuts in every direction, including at him.
The profanity hasn't disappeared so much as changed address. Around the same window covered here, a Sensible Medicine post had him calling hospital administrators "abject pieces of shit" — a register nowhere on display in the thirteen main-Substack posts analyzed above, which suggests a deliberate split between the flagship platform and the rougher-edged collaboration.
The conspicuous silence
This is where the Bannon comparison breaks down most informatively. Bannon's entire post-exile brand is narration: who screwed him, why he was right, what comes next. But Prasad's first thirteen posts contain one oblique sentence about having left government and nothing else — no account of the July 2025 ouster, no mention of the EEO probe, no rebuttal to the UCSF faculty petition that circulated in April 2026 opposing his rehire over "incredibly offensive public statements," and no reference to a Cancer Letter story that broke May 8, 2026 — four days before his return post went up — reporting that his CV inaccurately listed him as a 2016–2017 member of the President's Cancer Panel.
The Cancer Letter's editor, reflecting on the whole arc, called bringing an outsider critic into a regulatory chief-scientist role "sticking a finger in the eye"; a colleague on the same outlet noted that by the time he left, he had managed to unite "pharma, Congress, and patients" in opposition. None of that institutional memory makes it into the writing. Whether that reflects a strategic choice, a non-disparagement understanding, or simple disinterest in narrating his own martyrdom, the effect is a clean reset: he picked up the contrarian-biostatistician voice almost exactly where he set it down in spring 2025, skipping over everything that happened in between.
Why this should matter to you
The practical takeaway for this audience is not that Prasad got quieter — by length and frequency he is more prolific than before. It's that the share of his output devoted to oncology trial design, surrogate endpoints, and observational-versus-randomized evidence has nearly doubled, with heavier citation to primary literature than at any point in the pre-CBER sample. That is the register in which he has historically had outsized influence on how results get discussed at ASCO and AACR and in MolDX-adjacent circles, and he is now producing more of it, more often, than before he had the title.
The culture-war Prasad is still there — the UC Davis post alone confirms he hasn't lost the instinct — but it now reads as a side project rather than the main act. Combined with the YouTube series, the podcast, and the Sensible Medicine cross-posts, the pattern looks less like a man easing back into blogging and more like someone treating the platform itself, rather than the title he just gave up, as the durable asset.
##
Chat GPT 5
##
Vinay Prasad Returns: Same Voice, Different Instrument
Before Vinay Prasad went into FDA, his public persona was already unusually well defined: oncologist, methods critic, anti-groupthink polemicist, scourge of weak observational studies, antagonist of institutional media, and persistent critic of COVID-era medical consensus. In spring 2025, immediately before his FDA period, he was writing with the cadence of a man attacking the ecosystem from outside: universities, science reporters, FDA leadership, COVID vaccine policy, DEI, Harvard, and the broader biomedical establishment.
Now, in his return to public life, the striking fact is not simply that Prasad is “back.” It is that the material is both recognizably Prasad and meaningfully different. The voice has not softened exactly. But the center of gravity has shifted. The pre-FDA Prasad was a cultural combatant with medical credentials. The post-FDA Prasad reads more like a former regulator-clinician-methodologist who is trying to reoccupy the public square before anyone else defines what his FDA interlude meant.
The new essays are longer, more clinical, more technical, and less purely media-war-driven. They are still barbed. They still contain the Prasad signatures: fast-moving analogies, contempt for herd thinking, moralized language around weak evidence, irritation at journalists, and affection for randomized trials. But compared with spring 2025, the 2026 pieces are less about “the media is lying” and more about “the system is epistemically broken.”
That distinction matters for biotech, medtech, diagnostics, genomics, and policy audiences. Prasad may no longer hold power at FDA, but his essays still function as a weather vane for a particular wing of medical skepticism: anti-surrogate endpoint, anti-observational overclaiming, anti-regulatory complacency, anti-academic-progressive consensus, and deeply suspicious of industry-media-academic feedback loops.
The basic corpus
I compared two 11-article sets. The spring 2025 set runs from March 29 to May 4, 2025. The 2026 return set runs from May 25 to June 16, 2026.
| Feature | Spring 2025 set | 2026 return set |
|---|---|---|
| Number of articles reviewed | 11 | 11 |
| Total words, approximate | 5,700–5,800 | 8,800–9,000 |
| Mean length | ~520 words | ~800 words |
| Median length | ~520 words | ~815 words |
| Shortest piece | ~180 words | ~300 words |
| Longest piece | ~925 words | ~1,530 words |
| Dominant mode | polemic / media criticism | technical critique / clinical polemic |
| Main targets | media, universities, COVID policy, FDA/Peter Marks | weak studies, medical AI, academic medicine, overdiagnosis, clinical groupthink |
| Political temperature | higher and more explicit | still present, but more embedded in technical or institutional critique |
| Medical-technical density | moderate | substantially higher |
The most obvious change is length. The newer essays are roughly 50% longer on average. That matters because it changes what the pieces can do. The 2025 Prasad often launches one sharp claim, supports it with a few examples, and exits. The 2026 Prasad more often builds a case: an observational study is flawed because curves separate too early; a surgical disaster reveals groupthink; AI benchmarks are obsolete because model cycles outrun academic publishing; the “healthy person” is disappearing under a regime of screening, biomarkers, and pre-disease labels.
The change is not “more careful” in a bland sense. He is still often sweeping, categorical, and rhetorically aggressive. But the newer pieces spend more time inside medical reasoning.
Topic shift: from culture-war medicine to clinical epistemology
Spring 2025 Prasad is heavily occupied by three topics: COVID vaccination policy, universities/DEI/Harvard, and alleged media dishonesty. There are exceptions. One piece is a personal essay on Wim Wenders’ Perfect Days. Another is a food essay. But the main public-policy energy sits in the collision of Trump-era institutional reform, COVID memory, and distrust of mainstream medical journalism.
The 2025 titles tell the story:
| Spring 2025 theme | Examples |
|---|---|
| Media criticism | “A user’s guide to the media coverage of health and science”; “3 ways the media coverage of health care is dishonest”; Peter Marks departure essay |
| COVID vaccine policy | COVID booster clinical trials; Kennedy possibly pulling COVID shots for kids |
| Universities / DEI / Harvard | Harvard compliance; Democrat presidents and universities; science funding and diversity grants |
| Personal/cultural essays | Perfect Days; food and cooking |
| Science-funding skepticism | “Is more science better than less science?” |
By contrast, the 2026 return corpus has more clinical-methodological range:
| 2026 return theme | Examples |
|---|---|
| Methodologic critique | JAMA IM COVID booster paper; GLP-1 observational cancer study; Paxlovid/Tamiflu comparison |
| Clinical-system critique | OHSU mitral valve case; “healthy people” and overdiagnosis |
| AI and medical information markets | Gemini/OpenEvidence/ChatGPT paper; “Which AI will medical doctors use?” |
| Oncology | CML vs myeloma progress; daraxonrasib in pancreatic cancer |
| Politics/institutions | UC Davis admissions and DOJ; alcohol study framed through Biden/Trump |
| Medicalization | The endangered “healthy person” |
This is the central point: Prasad’s post-FDA writing is more directly relevant to biotech and medtech readers than his spring 2025 writing. The 2025 pieces were relevant because they revealed the worldview of a potential regulator: skeptical of COVID vaccines for low-risk groups, hostile to Peter Marks’ regulatory legacy, angry at media framing, and sympathetic to Trump-era pressure on universities. The 2026 pieces are relevant because they reveal the analytic template he still applies to products, evidence, and institutions.
The newer Prasad is asking: Is this study biased? Is the endpoint meaningful? Is the claimed benefit clinically real? Are academics dressing up a weak finding as a social good? Are hospitals, journals, AI vendors, and pharma advertisers aligned against the practicing doctor or patient? For anyone in diagnostics, genomic medicine, AI, oncology, drug development, or coverage policy, that is more than personality journalism. It is a map of objections likely to recur in FDA, CMS, payer, and guideline disputes.
The rhetoric: still sharp, but less purely media-centered
Prasad’s spring 2025 rhetoric is often openly political. The villains are “the media,” “universities,” “left wing science journals,” Harvard, Biden-aligned institutions, and Peter Marks. His media essay literally offers a user’s guide for reading allegedly biased health reporting. The Peter Marks essay frames FDA leadership through the language of media deception. The Harvard essays present federal pressure on universities not as authoritarianism but as democratic correction.
The 2026 rhetoric remains combative, but the villain set is broader and more professionalized. The targets are not only journalists or Democrats. They are observational researchers, JAMA IM editors, Nature Medicine timelines, OHSU groupthink, palliative care outsourcing, OpenEvidence’s advertising model, UpToDate’s slowness, physicians who over-order scans, and the medical profession’s habit of turning healthy people into patients.
That creates a different tone. The spring 2025 essays often feel like a man at war with a political-media class. The 2026 essays often feel like a man at war with medical epistemology itself.
A crude but revealing metric: in the 2025 set, “media” and journalist-related terms dominate. In the 2026 set, the language shifts toward “trial,” “observational,” “randomized,” “control,” “bias,” “mortality,” “study,” “paper,” and “data.” Political names remain, but they are less dominant. Technical vocabulary rises.
Controversialism: fewer grenades, more scalpels
There is a difference between controversy by insult and controversy by implication.
In spring 2025, Prasad often chooses maximum-friction phrasing. Universities are “lost.” Media coverage is “dishonest.” The COVID vaccine schedule for children is treated as corrupt or irrational. The Peter Marks essay is personally and institutionally severe. The Harvard pieces actively defend aggressive Trump pressure on universities.
In 2026, the language is still provocative, but the controversy often comes from where the argument leads. If a JAMA IM COVID booster study is invalidated by healthy vaccinee bias, then journals, reporters, and policy advocates are all implicated. If a medical AI paper is obsolete upon publication, then the academic publishing model is structurally unfit for frontier AI. If the OHSU case is framed as groupthink, palliative care outsourcing, and hospital PR culture, then a malpractice story becomes an indictment of modern medicine. If “healthy people” are endangered, then screening, prevention, biomarker medicine, and large parts of medtech business logic are all placed under suspicion.
That is arguably more interesting than the 2025 mode. The new pieces are less likely to be read only as red-team political commentary. They are more likely to irritate multiple constituencies simultaneously: journals, hospitals, AI vendors, pharma advertisers, guideline writers, oncologists, screening companies, and academic medical centers.
Political slant: less volume, same direction
The political slant has not disappeared. It remains recognizably anti-progressive-academic, anti-DEI, anti-Biden-era public health, and more sympathetic to Trump/RFK-style institutional disruption than mainstream academic medicine would be. The UC Davis essay is the clearest example: Prasad frames race-conscious admissions workarounds as open defiance of the Supreme Court and suggests DOJ pressure may spread. The alcohol essay frames a Biden-commissioned, Trump-buried study as an example of how media narratives are built around science. The AI and overdiagnosis essays are less partisan but still anti-establishment.
Compared with spring 2025, however, the politics are less all-consuming. In 2025, Trump, Biden, Harvard, DEI, Peter Marks, RFK/Kennedy, and media bias are central. In 2026, politics often enters as context rather than the whole engine. The deeper engine is institutional distrust.
That makes the 2026 corpus more durable. A biotech executive may tune out a Harvard-DEI polemic, but will pay attention to how Prasad dissects observational GLP-1 cancer data, medical AI incentives, oncology survival graphs, and the medicalization of healthy populations.
Media bias: from “the media lies” to “the media cannot appraise science”
This is one of the most important shifts.
In 2025, Prasad’s media critique is ideological and adversarial. The media is biased, dishonest, propagandistic, anti-Trump, and prone to misrepresenting figures such as Jay Bhattacharya or Peter Marks. His recommended response is to distrust the reporter, check sources, examine quoted experts, and assume institutional framing is suspect.
In 2026, his media criticism becomes more epistemic. The media problem is not only that journalists are biased. It is that they cannot critically appraise science quickly enough or deeply enough. In the JAMA IM COVID booster piece, the problem is that reporters propagate a flawed paper because they lack in-house methodological skill. In the alcohol essay, the media turns a study into a Biden-versus-Trump morality play. In the AI benchmark essay, the academic-media cycle is too slow to matter.
This is more sophisticated and more relevant to industry. The complaint is not merely “liberal reporters are unfair.” It is: the information system that converts preprints, journals, press releases, quotes, and headlines into conventional wisdom is structurally unreliable. That argument, whether one agrees or not, is exactly the kind of argument that shapes investor narratives, payer narratives, regulatory narratives, and public trust.
What biotech and medtech should notice
Three points matter most.
First, Prasad’s post-FDA writing is intensely hostile to weak evidence dressed as certainty. Observational studies are repeatedly treated as guilty until proven otherwise. Target-trial emulation does not impress him unless the design solves confounding. Surrogate logic is suspect. Survival-over-time graphs can mislead. He is still the same evidence hawk who has long criticized oncology approvals, but in 2026 the critique is broader: AI, prevention, screening, overdiagnosis, alcohol, GLP-1s, antivirals, and hospital systems all get pulled into the same framework.
Second, he is increasingly focused on incentives. Journals are slow and sometimes captured. AI platforms may be advertiser-influenced. Hospitals may prefer PR narratives to hard self-examination. Physicians may overuse tests. Academic medicine may use social goals to justify weak evidence. That incentive-based critique is powerful because it travels across domains. A diagnostic company selling early detection, an AI company selling clinical decision support, a genomics company claiming utility from biomarkers, or a pharma company promoting observational real-world evidence could all find themselves inside the Prasad blast radius.
Third, Prasad’s rhetoric remains a double-edged sword. His clarity and aggression make him readable. His analogies stick. His impatience with cant is part of the appeal. But the same style can overrun nuance. A Prasad essay often compresses uncertainty into verdict. That makes the writing compelling, but it can make some judgments feel more prosecutorial than adjudicative.
The “Bannon out of power” analogy
The Steve Bannon analogy is useful, not because Prasad resembles Bannon ideologically or professionally, but because both show how a figure can remain influential after leaving formal power. The important question is not, “Does he still control the agency?” He does not. The question is, “Does he still shape the mental model of people who distrust the agency?”
For a slice of physicians, investors, policy skeptics, anti-establishment conservatives, heterodox academics, and evidence purists, Prasad remains a narrator. He tells them which studies are fake, which institutions are captured, which headlines are propaganda, which claims require RCTs, and which sacred cows deserve slaughter. FDA service may even strengthen that role. He can now write not only as UCSF oncologist and public critic, but as a former senior FDA official who briefly saw the machinery from inside.
That is why the return matters. It is not merely content production. It is reentry into narrative power.
Bottom line
The spring 2025 Prasad was louder politically. The 2026 Prasad is more interesting professionally.
Before FDA, he was writing as an outsider attacking the media-academic-public-health complex. After FDA, he is writing as a returned insider-outsider attacking a broader medical knowledge system: journals, observational studies, AI benchmarks, hospital narratives, overdiagnosis, academic admissions, and clinical overreach.
For biotech and medtech audiences, the key takeaway is not whether one likes Prasad. The key takeaway is that his critique has become more directly aimed at the evidentiary architecture on which companies depend. If your product rests on observational inference, surrogate outcomes, screening expansion, AI benchmarking, biomarker enthusiasm, guideline capture, or journal-mediated hype, the post-FDA Prasad is a preview of the attack.
He is out of power. But he is not out of the conversation. And in some ways, the conversation is exactly where he is most effective.
##
##
