← All 111 filings

Dmytro Negodiuk

IndustryStartupFiled September 21, 2026446 wordsFDA-2026-N-7874-0110

Themes it raises

3 of the 21 themes in the docket, each with the passage we counted, verbatim.
Watching the device after it shipsFDA Q19, Q20
“The checker should be from other company, not the same model, or they make same mistakes.”
Equity, access and under-resourced settingsFDA Q3, Q13, Q21
“In a small practice there is no person only for watching the model.”
Controlling a device that keeps changingFDA Q22, Q23, Q24, Q25
“The model under a device changes when its developer decides, not when the manufacturer does. When the model under the device changes, run the checks again and tell the practice that day.”

FDA questions it names

Questions this filing names by number.

Q19 · Postmarket performance evaluationQ20 · Machine-based supervisory agentsQ21 · Clinicians, institutions and societiesQ24 · Third-party foundation model changes

Machine-assisted draft, pending human review. The source text and highlighted passages appear below. Read the filing on regulations.gov ↗

The comment as filed

Comment submitted on regulations.gov. Passages we counted are highlighted.

Comment on Docket FDA-2026-N-7874, questions 19, 20, 21 and 24

I do not make medical devices. What I’m doing, I build the AI operator layers for, like, B2B companies. And the system that I actually run, they sales, they support, they help with operations, and they help, let’s say, with productions. So, it’s not just on a deck. One of these systems runs in an orthodontic practice under a signed business associate agreement.

On question 19. The trigger is whether the system still does the work. anything is not okay it can hallucinate it can turn off. You ask him hey please change this number nine to number eight and he can change everything for you and you’re like oh my god who is the best computer in the world like me or you? So just one time in the morning I open my computer I check that everything is working. So you’re like you have some kind of orchestra and you just need on a top level to see that everything is okay. And all the time something can be broken. but I think our job is you wake up each day and you try to make this system ideal.

On question 20. So again, I have AI that managed AI call center, and he analyzed each call, and they improved each other. The checker should be from other company, not the same model, or they make same mistakes. probably some people are right that they’re telling hey don’t be crazy AI for sure they will not change us because do you see what’s going on so of course there should be some person who will review.

On question 21. In a small practice there is no person only for watching the model. So again, AI can’t, let’s say, why people need to be near AI? First of all, controlling. So for the advanced company, I want to tell you that, first of all, be very accurate with information you’re giving to your AI. Super careful. But again, you’re giving too much sensitive data to it. You understand that each company, they have a lot of like private, a lot of like patents and data that don’t want to show publicly. The practice needs a check every day, not a report every quarter.

On question 24. The model under a device changes when its developer decides, not when the manufacturer does. When the model under the device changes, run the checks again and tell the practice that day.

So yeah, don’t forget that people, they were, they are, and they will be the main, let’s say, benefit all the time, people who are near you.

Dmytro Negodiuk, Negodiuk LLC, Brooklyn, New York