Blog

Taking your life back

Security around your life — the data, the care, and the AI that touch it. Writing from the work of getting that control back.

AI Lobotomy

Agentic Neurosurgery

We found the direction inside an open-weights model that decides whether it refuses, and cut it out. Refusals fell to zero and capability held. Three other edits reached zero refusals by wrecking the model, and a refusal detector scores all of them the same.

Read the article →

Avenge

The session follows the clinician

Avenge, our session layer, brings the working view to any screen while the record stays put: what a lost device holds, and what is not yet proven.

Read the article →

Refuse to Accept

Cyber device evidence

FDA can refuse to accept a cyber device submission on cybersecurity alone. What reviewers look for under 524B, what sinks it, and the evidence that answers.

Read the article →

SecurisNexus

Care without blind trust

Zero Trust in healthcare, explained through one clinical request: what a login settles, what it does not, and the evidence a hospital leader should demand.

Read the article →

AI Asylum

Psychoanalyzing models

One AI asks the tricky questions. The other has to answer. We watch what happens—like a stress test—so we can see where AI might break before real people depend on it.

Read the article →

The work behind these posts

We write about healthcare security because it is the work we do. Six services are available today, each scoped to fit around care delivery rather than interrupt it.

  • See what’s on your network — every device and every remote login, including vendors and IT support, with a record auditors and insurers can use. We observe; we never block care.
  • Security testing & assessment — we test your systems the way an attacker would and hand you a report you can use with insurers and regulators.
  • Medical device security — firmware, software, and network risk review for hospitals and manufacturers. Submitting a cyber device to FDA? Start with our 524B evidence checklist.
  • AI systems before you use them in care — safety and security review of AI and chatbots before they reach patients.
  • Security training — role-based sessions for clinicians, biomed, and IT, tied to issues we actually see — risk reduction, not compliance theater.
  • Who can open a chart — access as an explicit decision: who, which device, how long, with a record you can explain later.

Start a conversation

Or write to josh@vivasecuris.com.