Here's a summary of the themes from the Hacker News discussion, with direct quotes:
Focus on Verification and Deterministic Checks for Safety
A primary concern raised by users is the company's approach to ensuring accuracy in data entry, especially given the high-stakes nature of healthcare. The company emphasizes their methods for verifying that information has been correctly entered, particularly when dealing with systems that lack APIs.
- "we build in verification and deterministic checks, so customers can confirm it was done right" (candiddevmike)
- "How do you deterministically verify that information was correctly entered?" (ljm)
- "It depends on the source and destination. The trickiest case is when we're using browser agents for data entry. We can use the fact that we focus on repetitive tasks to our advantage - we know what sections of UI we need to check, and for what data. We can verify correctness by re-extracting data from the destination UI (via the DOM or OCR) and checking: That all expected fields were entered correctly. That no unexpected or extraneous data was added." (arcb)
- "When we have access to a direct data source (like an API or file store), verification is simpler — we can do deterministic checks and directly confirm the values." (arcb)
Navigating the Healthcare Regulatory Landscape (FDA, HIPAA)
The discussion touches on the complex regulatory environment of healthcare, specifically concerning the FDA and HIPAA compliance. Users question the company's approach to medical device exemptions and how it handles sensitive patient data.
- "I think you folks are really pushing the limits of the FDA's medical device exemptions for administrative work here. Are you working with the FDA on any kind of AI medical device certification?" (candiddevmike)
- "how are you guys handling HIPAA compliance for AI Agents? how is it that data in motion is secure?" (beingalikhan)
- "Great question. In the web agent case, we solely use HTTPS, and only between resources we either directly control (our servers), or whitelisted customer websites where we connect on HTTPS. An HTTP connection would fail the call stack, as would visiting a non-whitelisted link. A lot of our work happens away from the browser (in APIs and data stores), where we encrypt at rest and in motion." (arcb)
Integration with Existing Healthcare Systems (EHRs) and Technical Implementation Details
A significant part of the conversation revolves around how the AI agents interact with existing Electronic Health Record (EHR) systems and other healthcare software. The company clarifies its integration methods and the reasoning behind certain technical choices.
- "What EHRs do you integrate with?" (candiddevmike)
- "Amongst EHRs we currently work with Athena, though we do a lot of work on isolated file stores that our customers create for us." (arcb)
- "If you have an API, why are you using browser agents?" (candiddevmike)
- "We don't use browser agents if an when we have an API - we prefer the strongest data types we can access. It comes down to what our customers can work with. Some of them are fairly technical (have an IT team), and some aren't (have a legacy portal and operate on spreadsheets / paper)." (arcb)
- "I work in with Hospital software and we have APIs for referrals (from general practitioner). Thank god." (Mashimo)
- "What we currently are looking at is scheduling of staff, because somehow that involves different software (dr. vs nurse), and the staff builds a spreadsheet, and then enter it into other software. Totally whack how much time and effort they spend on that." (Mashimo)
Target Clinic Types and Clinician Adoption of AI
Users inquire about the specific types of clinics that are a good fit for the company's services and what drives the adoption of AI solutions by clinicians. The company identifies specialty groups with repeatable workflows as ideal candidates and notes a growing openness to AI among healthcare professionals.
- "You mentioned your early customer in obesity medicine. Are there specific kinds of clinics that are an especially good fit for you? Are you seeing any patterns in the kinds of clinics that are relatively eager to adopt an AI product like yours?" (noleary)
- "From our side, we’re currently focused on specialty and multi-specialty groups. For example, obesity medicine, cardiology, pathology centers, radiology clinics... These groups tend to have repeatable workflows and a lot of operational toil. That makes them a good fit for automation." (arcb)
- "Compared to 3 or 4 years ago, clinicians are much more open to AI. They've heard of ChatGPT or ambient scribes, and they often come to us with specific ideas of what they want AI to solve." (arcb)
- "Said that, we also hear a lot of of requests from groups that we have to turn down. Sometimes we can't guarantee success, or the product just isn’t ready for that use case. As an example, a ton of clinical interfaces only work on desktops, which we'd like to support but don't yet." (arcb)
AI Agent Technology Stack and Development Approach
The discussion delves into the technical infrastructure and development choices, including the use of specific libraries, LLMs, and data processing frameworks. The company highlights its preference for open-source solutions and its ongoing efforts to improve agent capabilities.
- "Is it just
<pre>validates :email, format: :email</pre>In Rails?" (ljm) - "This is a pretty cool idea and implementation. Any more details on the tech stack you guys are using (besides
browser-use)?" (MK_Dev) - "We have a fork of browser-use that lets us hand hold web agents since we know our tasks are repetitive. We can cache expected paths and fire alerts if we go off the rails. We'd love to contribute it back at some point, mainly a question of bandwidth." (arcb)
- "We're evaluating Cua (https://www.ycombinator.com/companies/cua) to containerize our agents; am a fan so far. We're also putting Computer Use agents from (OAI and Anthropic) to the test. Many legacy ERPs don't run in the browser and we have to meet them there. I think we're a few months away from things working reliably and efficiently." (arcb)
- "We're evaluating several of the top models (both open and closed) for browser navigation (claude's winning atm) and PDF extraction. Since we're performing repetitive tasks, the goal is make our workflows RL-able. Being able to rely on OSS models will help a lot here." (arcb)
- "We're building our own data sets and evaluations for many of the subtasks. We're using openai's evals (https://github.com/openai/evals) as a framework to guide our own tooling." (arcb)
- "I prefer open source when possible. Everything's new and early, and many things require source changes that others might not be able to prioritize." (arcb)
- "what made you fork browser-use? what were the missing bits? your use case sounds similar to what they're trying with their new workflow-use repo (I am not affiliated with them, just curious)" (tjsk)
- "It's a great repo! We had issues with iframes and framesets (which are old DOM tags) we had to write custom code for. Some DOMs need annotation to provide meaning to an LLM (for example, a button is clearly an "add demographics" button to the human eye, but is ambiguous in the DOM (ul contains li...). Some bottlenecks in navigation required manual attention. We keep those to a minimum. I think the future is being able to progress from highly deterministic JS code, to more agentic LLM-driven decisions. One does need to be able to control this for performance, cost, and accuracy. And yes we have some overlap with workflow-use's direction, but I hope that more such OSS methods gain popularity! It'd simply mean we can go after higher value and more complex clinical tasks!" (arcb)
- "Did you consider working around those using the vision models vs DOM parsing? Was cost/latency the bottleneck? Seems like the agentic future you describe would need more vision based parsing" (tjsk)
- "I believe we will at some point. All question of the right need coming up. Text OCR has gotten really good, and if you think of it from a UI perspective, the only real contract is that a screen will show text that's representative of the information entered. The DOM is useful but is a changeable contract!" (arcb)
- "Edit - one thing I'd love to find a good solution for is reliably extracting handwriting from PDF documents. Clinicians have to do this a ton to keep the trains running on time, and being able to digitize that knowledge on the go will be huge." (arcb)
The 'Invisible' AI and Empowering Clinicians by Reducing Toil
The company emphasizes a philosophy of being "invisible and yet effective" for their customers, aiming to reduce the burden of administrative tasks so clinicians can focus on patient care. This aligns with a broader sentiment about the potential of AI to alleviate physician burnout and improve healthcare throughput.
- "However, one of the big ones is that clinicians don't want more screens; they're already overloaded. So we're succeeding if we're invisible and yet effective for our customers." (arcb)
- "Another is that once you free up clinician time, they will quickly find higher-leverage tasks. It shows how overloaded the system is, and that there's pent-up demand to make it better." (arcb)
- "What could they be doing if they didn't have to do high-volume data entry. In one of our customers' cases, they instantly started helping doctors with responding to patients directly." (arcb)
- "From my family of medical professionals I can now go back and tell them: just wait, the future looks bright. This is one of the most grueling, labor intensive, boring, error prone, hate my job areas possible. It makes perfect sense for agents to perform these tasks, but as with everything else there will be a labor impact." (sibeliuss)
- "On the labor impact front, what we're seeing is that there is so much pent up demand for care, that any time we free up staff, they enable more throughput or more depth on casework. I hope we create more potential for humans to improve care because of our work." (arcb)
- "This will certainly be the case, and the promise of this tech. Thanks for taking on this space." (sibeliuss)