Hacker News Reader

Comment by floor2 | Hacker News Reader

Maybe this is an upside to the looming AI-taking-jobs issue.

I bet a ton of these bureaucratic approve/deny/request-more-information workflows in current EMR systems will be replaced by AI in a few years, putting the people out of work.

Once AI eats their jobs, less resistance to changing the system.

View on HN

Unemployed people: "Oh wow, now I see why we should shutdown the people eating machine"

Trillionaires with killbot armies: "No, I'm not seeing what you're seeing at all"

Hopefully these jobs go away sooner than later, they are not exactally fufilling work & there are much better jobs out there.

I am now buidling an aprove deny API, going to call it dPANEL for deterministicPanel. Also chatGPT seems ungodly enthusiastic to help me build it...

Pricing Strategy for Health Insurance Approve || Deny API Pricing Structure Base Subscription Fee: Annual Subscription: $120,000 per year, covering up to 50,000 API calls. Volume-Based Pricing: Beyond the initial 50,000 calls: Tier 1: 50,001 to 100,000 calls at $1.50 per call. Tier 2: 100,001 to 200,000 calls at $1.25 per call. Tier 3: Over 200,001 calls at $1.00 per call. Multi-Year Discounts: 2-Year Contract: 10% discount on the annual subscription fee. 3-Year Contract or more: 15% discount on the annual subscription fee. Additional Services: Custom Integration: $30,000 one-time fee. Premium Support Package: $25,000 per year for dedicated support and quarterly performance reviews. Early Termination Benefits: Early termination within the first year incurs a 50% fee of the remaining contract value. Termination in the second year or later incurs a 25% fee of the remaining contract value. Cost-Saving Benefits Labor Cost Reduction: Automation of approval and denial processes reduces the need for manual labor. Increased Efficiency: Faster processing times improve operational efficiency and customer satisfaction. Error Reduction: Minimized human errors reduce costs related to claim reprocessing and disputes. Scalability: Efficiently manages varying loads without significant staffing changes. Regulatory Compliance: Helps ensure decisions are consistent and compliant, reducing potential fines. Data Insights: Offers valuable analytics that can lead to better risk management and operational adjustments.

Most of it doesn't even require AI. The majority of those jobs can be eliminated just by implementing existing interoperability standards between payer and provider organizers, and then writing some simple deterministic rules. I previously worked on projects to do just that. There is a tremendous amount of waste (and associated jobs) that can be eliminated without buying a single GPU.

They could also be replaced with even simpler code:

    bool should_approve_claim(int claim_id)
    {
        return false;
    }

Ha ha, but even single payer or socialized systems have similar mechanisms for preventing payments for treatments that don't meet medical necessity criteria. Blindly approving all claims would only serve to make the system more expensive and overloaded than it already is.

>I bet a ton of these bureaucratic approve/deny/request-more-information workflows in current EMR systems will be replaced by AI in a few years, putting the people out of work.

The EU AI act prevents exactly this scenario.

Reply on news.ycombinator.com