jdrez

prev @PalantirTech @Umich @monitors

[jdrez @waldo47]

After an incredible 2 years, today was my last day at Palantir.

In that time I built the engine for staffing 5-10% of US hospital beds and millions in revenue cycle workflows.

Heading into a new venture with one of the best partners I could ask for. More soon 👀

[jdrez @waldo47]

@OpenAIDevs I know that guy! @jxnlco

[jdrez @waldo47]

@trq212 @hive_echo Sickk. When can it call me when it hits an ambiguous decision instead of stopping (or AskUserTool).. Inbound calls from my pair programmer 😉

[jdrez @waldo47]

@restocc They should give a heads up so I could’ve put in pto

[jdrez @waldo47]

Compete with me on Agona 🤘
https://t.co/qYKHVzN3m7

[jdrez @waldo47]

Today I signed a pilot, launched an app, and ran pltr all hands.. There’s nothing Claude can’t do

[jdrez @waldo47]

This week my team closed an enterprise deal in an entirely pro-code infra based on Anthropic Agent SDK

Key takeaways
- Access to the problem is harder than the problems themselves (sometimes)
- Everyday you can do a lot more with a lot less
- Mapping units of human work to tasks is key

If you’re building cool agents I’d love to hear how you’re doing it

[jdrez @waldo47]

@restocc Good luck on the next chapter wishing you the best. Hopefully we get to work together soon

[jdrez @waldo47]

@restocc We’re frolicking rn

[jdrez @waldo47]

This and no mcp only commands +gpt 5 high

[jdrez @waldo47]

@ivanhzhao It feels like we’re stuck in the ‘rearview mirror’ phase partly because we’re rationing tokens—spending them on a cheap end-of-pipeline summary instead of allocating them throughout the workflow to maintain state, track evidence, and actually do the REAL work.

[jdrez @waldo47]

Just witnessed a friend who didn’t know how to navigate to the files in cursor because agent was so good

[jdrez @waldo47]

@m_chael @mercor_ai @whop @slashapp @BloomTrading @AbstractChain Whatever the sneaker -> PLTR homies launch once we vest

[jdrez @waldo47]

My team is hiring 2 NYC-based FDEs.

What we’re working on:
- Recovering $1.2B/year in revenue cycle leakage
- Building automation for document extraction and voice agents
- Designing a rules engine that encodes years of physician expertise into real-time healthcare decisions

Reach out if interested!

[jdrez @waldo47]

Interesting exercise. Walk around campus buildings and feed gpt collections of whiteboards to learn different subjects

[jdrez @waldo47]

POV: you’re an inpatient medical necessity denial and this is the last thing you see 😈

[jdrez @waldo47]

POV: You’re the next generation of umich Palantir being explained InterQual automation

[jdrez @waldo47]

@fromwpm Yes!

[jdrez @waldo47]

How does your company celebrate happy hour?

[jdrez @waldo47]

@__Tkat__ 🅱️addie