I've known about Palantir for almost a decade now and to this day I cannot really describe what the company actually does. I just know that when I was in school, it was supposed to be a very prestigious company for super smart people that had tough interviews. Then many years later it became a meme stock for a while. Now its being hyped for doing AI although it doesn't seem to be doing anything similar to OpenAI or llms or developing self driving cars. I tried looking it up and just got a bunch of very vague unhelpful descriptions like they do something for the government. Can somebody give me a detailed and easy to understand explanation of what exactly Palantir does? Like what is their service/product exactly?
Stuff, things
Process spy data
Alex Karp does Pump & dump. It is the real answer. It’s not a tech company but a tech consulting company which sends its engineer to build solution for clients.
Ah I see. Like people always refer to them as an AI company. But like do they really specialize in building AI software or something?
They are kind of a special case of govt contractors. They get long term contracts to build special software for clients and deploy a small number of engineers to build the software. One of their softwares was called Gotham, you can look it up. Some of the software they developed for DHS was used for background checks, the kind that red flags people for national security.
Yeah I read about Gotham and Metropolis but didn't find any clear explanation of just what exactly they do. Do you know?
Beep boop drone strike Syria
Check YouTube
Not working for palantir, so employees correct me if I‘m wrong. They have 3 products: Gotham: Modular system that integrates disparate data sources for intel / military purposes into one ontology. Modules include map (Gaia), presentations, chat, graph search, satellite imagery (meta constellation) and much more on top of a RBAC system. Gotham is very flexible. Foundry: Modular system that also integrates disparate data sources into one ontology, but rather for commercial use-cases. Heart of it is a Spark / flink / lightweight runtime that processes data, captures lineage and writes them to the ontology. On top of that you can then use modules to create dashboards, what if analyses, AI models, no-code applications and much more. Apollo: CI/CD system that helps you to deploy software in complex setups. This is basically productionizing the experience that PLTR gathered while deploying Gotham to remote air-gapped regions. Let me know if you want to know more.
Metropolis was an early product for financial analysis, which eventually failed.
This is a pretty good explanation: https://blog.palantir.com/connecting-ai-to-decisions-with-the-palantir-ontology-c73f7b0a1a72
I honestly disagree. This just explains the concept of the ontology and how LLMs benefit from it on a very abstract level
Yes but it also talks about north and south of the ontology, all of which Palantir does. And there’s a more concrete example at the end of the blog post.
Big data analytics