DevSide 3: Extracting Table Data + AIP Orchestrators

In this episode, I explain how to extract structured data from our SEC filings using LLMs. I'm also showing how to use AIP Orchestrators to deal with token limits and rate limiting!
AIP Orchestrators: www.palantir.com/docs/foundry...
Join the discussion: / discord

[@gielwinkelmolen7045]

I'm wondering what percentage of viewers of these video's are actually Pltr investors who don't understand anything of these technical video's. I bet it's high. I am sure as hell one of them haha. But I am learning and trying to wrap my head around this. Thanks Codestrap for everything you do 🙏

[@crackyflipside]

Love these videos dude, I'm working on a project to study an old weightlifting program that reads off of PDFs. Studying this video and taking notes haha

[@rajkarri2991]

I totally get that on separate LLM calls stuff after burning some dollars on my self service AIP account. Thanks for the tip 🙂

[@berkshiretube]

I admire your interest to harvest SEC filings with Foundry

[@NeillRyan]

Hey Codestrap, I was also wondering about the technical jargon in a lot of PLTR videos and marketing, it's not just your content but official PLTR content as well. I click on an exciting looking thumb nail and after a minute or so my eyes have glazed over and I really can't keep up with most of whats being said. I appreciate it's a bloody complex piece of kit so there's definitely some geeking out to be done. While they've definitely gotten better at it, I still feel that PLTR are really missing a trick when releasing what should be "sexy content", it always seems to descend into complex techo-babble which about 1% of people can understand.. I think there needs to be a strong marketing push towards making simplistic content based on clear results (which is basically the only thing new customers are interested in) - features, benefits, cost savings, efficiency improvements should all be front and centre. - Hopefully you can get into Karp's ear for me...🤣 Peace brother

[@Teddy-mj2zx]

Code. Thanks for these demos! I'm assuming that if AIP is this good with unstructured data (text?) then if it were leveraged against databases to correlate data from multiple sources it would help save time and effort for software developers using python scripts to stitch data together?

[@codestrap8031]

For sure. I'll be pushing up another video in 10 minutes on how to publish reusable Python libs. Makes it easy to reuse all those scripts across projects.

[@midwestcannabis]

Party🎉🎉