Such sad news. As a kid growing up in RI I used to love watching Computer Chronicles on our local PBS station each weekend. Stuart and Gary were the best. RIP to a legend
That's a great point. Sometimes we look for architecture or technology solutions for a problem that could be easily solved at the sales level by negotiating a PPA (Private Pricing Addendum) with AWS.
I can quickly see something like this turning in an AI arms race between insurance and the provider with each auto-approving/denying/disputing the other. All the while locking out smaller players because they can't afford the 3rd party disputotron.
I already have a solution to the downcoding practices of these health insurance carriers.
I recently created an application called EMpowerAI that uses AI to analyze clinical notes and assign appropriate billing codes based on medical complexity or documented time. It also can enhance the Assessment/Plan to justify higher billing codes if the note content supports it.
As a Cardiac Electrophysiologist, I optimized the application for cardiology and EP, though it is scalable to other specialties. I am looking for beta testers and would appreciate any feedback. Here is a link to the app:
You would have to leverage the law (if you have one) that involves the state resolving the dispute because otherwise the automated disputes would probably be dropped on the floor. The insurance company has the leverage because they're actually in possession of the money and the contract that gives them stupidly high discretion on how much to pay out.
Doing nothing but flipping the burden, doctors get paid whatever they invoice and insurance have to claw it back would make a lot of this stonewalling bullshit go away. But with an openly corrupt government paid by insurance it'll never happen.
For a workload of that size you would be able to negotiate private pricing with AWS or any cloud provider, not just CloudFlare. You can get a private pricing deal on S3 with as little as half a PB. Not saying that your overall expenses would be cheaper w/a CSP than DIY, but its not exactly an apples to apples comparison of taking full retail prices for the CSPs against eBayed equipment and free labor (minus the cost of the pizza).
egress costs are the crux for AWS and they didn't budge when we tried to negotiate that we them, it's just entirely unusable for AI training otherwise. I think the cloudflare private quote is pretty representative of the cheaper end of managed object-bucket storage.
obv as we took on this project the delta between our cluster and the next-best option got smaller, in part bc the ability to host it ourselves gives us negotiating leverage, but managed bucket products are fundamentally overspecced for simple pretraining dumps. glacier does a nice job fitting the needs of archival storage for a good cost, but there's nothing similar for ML needs atm.
Yes, the lottery has rules and governance. It's a much safer bet. Startups can decide to devalue their employees' shares. I'd venture that the odds are at least stated on the back of the ticket with the lottery. Employees of privately held startups are often sold a dream of future riches that rarely happens. Even where there is an exit, often even founding employees get taken for a ride.
I do the same thing at AWS, except use a 100, 200, 300, 400 level scale. I am just looking to see what the candidate thinks they are strongest on and weakest. If the candidate says they have never heard of AI/ML topics, there's no point in quizzing them on it. Not everyone can know everything, it wouldn't be fair to judge someone based on not being able to answer questions for a domain they don't claim to have any experience in. Or worse yet, waste valuable / limited interview minutes on fruitless questions.
reply