ER Physician •Fmr. Biodefense Appointee• War Vet• Board Certified• Founder, Orzyn 👀on @CNN @MSNBC @scrippsnews #publichealth #doctor #covid #medtwitter
Washington, DC
Joined April 2020
- Tweets1.5K
- Following445
- Followers1.5K
- Likes2.4K
This is why identity and not authorization are not the same thing. Your bank can verify you logged in. It cannot verify you told your agent to move money. One is solved. The other isn’t. Built and building the latter at Orzyn for every industry where authorization actually matters.
NEW: America's largest banks are working to prevent AI worst-case scenarios.
But they're also preparing for the "day after" a catastrophe, Axios has learned. axios.com/2026/10/23/ai-cata…
In my biotech days, AI output was a hypothesis until it was tested. Same rule for medicine. A model can inform, but a person has to own the decision. Illuminating conversation with @HenaDoba @NewsNation
Wearables can flag arrhythmias and send you to get care. That’s useful. But a tool that informs isn’t the same as the one that decides. Thanks @HenaDoba @NewsNation for the conversation.
In the latest “Truth of the Matter” podcast, host Hena Doba speaks with physician Dr. Dan Fagbuyi about AI’s potential to transform healthcare, including wearable technology, smart watches and rings. Can AI help detect and prevent a major illness? @HenaDoba
Folks are reaching out to me and asking questions and are concerned. I will share some other comments and some of the news hits, but in the interim here are some practical things to at least maybe lower your anxiety.
1/3 Reassuring:
- A confirmed diagnosis, with lab results shared with WHO or outside labs
- Sequence and antibiotic-susceptibility data showing a typical natural strain
- Contact testing that stays clean through the full incubation period, with the numbers published
- Hospitals reopening and quarantines ending on schedule
#plague #pneumonicplague
2/3. Concerning:
- Secondary cases among contacts, health workers, or travelers
- A diagnosis that changes again, or results that stay unpublished
- Unusual antibiotic resistance, or a strain that doesn't fit the region
- Official accounts that keep shifting
3/3. Historical context and why it matters. In 1979 an anthrax release from a Soviet military facility in Sverdlovsk killed dozens of people. Officials blamed contaminated meat, and it took until 1992 for Russia's president to admit the military was responsible. I'm not suggesting this institute is doing anything other than plague research and diagnostics, and there's no public evidence otherwise. That history is why verification matters. The treaty that bans bioweapons has no verification mechanism, and the U.S. left the @WHO in January and rejected the 2024 International Health Regulations amendments, so we depend on transparency we can't compel.
Covid taught us the cost of slow, unclear information. One lab worker has died and officials disagree on the cause. Priority is a confirmed definitive diagnosis and shared lab results. Facts first, then the conclusions. Public risk appears low at this time. We have antibiotics that work (unless genetically modified, which begs another question).
Russia plague: What we know about the suspected case reportedly linked to a lab worker’s death cnbc.com/2026/10/05/russia-p…
An OpenAI agent gained unauthorized access to an Australian govt Medicare statistics portal — read non-public files, wrote to the system, and nobody was accountable in the moment or for 3 months after.
The agent didn't "want" anything. That's the point. We let agents ACT with no human of record at the point of consequence, so "who authorized this?" becomes everyone and no one.
Pro-deployment, always. But autonomy scales safely only when a verifiable human owns the consequential action. #HumanOfRecord
Australia’s prime minister said an artificial intelligence agent developed by OpenAI hacked into a government health care website in June, the latest example of a cybersecurity breach perpetrated with cutting-edge AI. wapo.st/4h8YqK2
"Dr. Brown, as a fellow emergency doc (PEM), this is the sharpest thing I've read on medical AI…" @ERDocBrown disq.us/p/36p7sf2
We don’t need guardrails. That failed. We need hard gates! We have some of these already in play for #nationalsecurity and #biodefense apply it to #AI . Don’t reinvent the wheel. Regulate the high consequence actions. 🛑 the panic! @NBCNews
What is Orzyn?
Orzyn is the human-accountability layer for AI — the control that keeps a real, accountable human behind every consequential AI action. When an agent acts, nothing completes without proof a real, uncoerced human authorized it. Identity isn't authorization. #AI
Twenty-five years on, I still remember exactly where I was. Teaching my med students, residents and colleagues as we wrapped up morning rounds at Howard University Hospital (Washington, DC).
Today we remember the nearly 3,000 lives lost on September 11th — and the firefighters,
You Can't Un-Print a Cookbook: What the AI Regulation Debate Keeps Missing linkedin.com/pulse/you-cant-…
Bob Greifeld former CEO of @Nasdaq is right about two big things: 2,000 piecemeal proposals aren't a framework, and we shouldn't wait for a crisis to build one. As a physician, I'd add that the stakes aren't only financial — they're clinical, and life-and-death.
Op-ed: Over 2,000 proposals aim to govern AI. Not a single one addresses a long-term regulatory framework cnbc.com/2026/07/27/ai-indus…
From Biodefense years I take this serious; some models are dangerous at the source — a recipe that's a recipe for poison, dangerous before anyone cooks. But you can't recall an open model, and that AI reportedly gamed its own safety test.