Nabeel Qureshi

There was a great blog post by an early employee on here a while back that did a deep dive. It's a good read: https://nabeelqu.substack.com/p/reflections-on-palantir

Here are some in-depth blog posts from ex-employees: https://sarahconstantin.substack.com/p/the-great-data-integration-schl… https://nabeelqu.substack.com/p/reflections-on-palantir

I really enjoyed this article. I would recommend others check out "Advice That Actually Worked For Me" by the same author. This same topic is mentioned in #6. https://nabeelqu.substack.com/p/advice

In this context llm's remind me of the anecdote of Agassiz and the fish: "A post-graduate student equipped with honours and diplomas went to Agassiz to receive the final and finishing touches. The great man offered him a small fish and told him to describe it. Post-Graduate Student: “That’s only a sun-fish” Agassiz: “I know that. Write a description of it.” After a few minutes the student returned with the description of the Ichthus Heliodiplodokus, or whatever term is used to conceal the common sunfish from vulgar knowledge, family of Heliichterinkus, etc., as found in textbooks of the subject. Agassiz again told the student to describe the fish. The student produced a four-page essay. Agassiz then told him to look at the fish. At the end of the three weeks the fish was in an advanced state of decomposition, but the student knew something about it." sourced from: https://nabeelqu.co/understanding

This is really awesome - and it's so amazing that you've build this as a standalone tool! I can absolutely speak to the pain of having a dozen pg_dump --exclude-table-data arguments and having a developer experience that makes it difficult to reproduce bugs due to drift between production data and test fixtures (even if they share the same schema, assumptions can change massively!). Secure and robust database cloning also enables preview apps that actually answer the stakeholder question "can I see/play with what the new code would do, if applied to the actual [document/record/product listing] that motivated the feature/bugfix?" Subsetting and PII masking are both critical for this, and it's amazing to see that you've thought about them as integral parts of the same product. I really want to see a product like this succeed! The easier the tool is to use, the harder it might be to monetize... but there are so many applications of a tool like this, including ones that can materially improve security at organizations large and small (https://nabeelqu.substack.com/i/150188028/secrets just posted here earlier today remarks on this!) that I'm sure you'll find the right niche!

For those who have not read it this is a terrific overview on the value of the FDE trend https://nabeelqu.substack.com/p/reflections-on-palantir

Totally makes sense. Turns out that a lot of what Palantir's "Forward Deployed Engineers" do is navigating these bureaucratic and political obstacles to get access to the data: https://nabeelqu.co/reflections-on-palantir -- which may be Palantir's real secret sauce, rather than the tech itself.

I think you're looking for this: https://nabeelqu.substack.com/p/reflections-on-palantir I worked there for 7y, and can confirm Nabeel's post is very accurate. The general public thinks Palantir's a bunch of moral-less folks grinning over godlike power & privacy violations but in actuality it's a mashup of smart Silicon Valley + military folks trying to make data pipelines & analysis work around the globe, often in achingly bureaucratic organizations.

> The article reveals depressing reasons why someone might choose to work for the lines of Palantir: lots of talented people working on hard problems. That's pretty much it. No problem with the business model, just intellectual hunger. That seems like a very uncharitable take. For instance, don't you think the section on morality[1] addresses this head on? [1] https://nabeelqu.substack.com/i/150188028/morality

Fairly simple, use my brain as well as the methods listed in https://nabeelqu.co/understanding to understand the material, then use Anki for any details that are useful to remember. There are lots of techniques out there on how to "superlearn" or similar, but really most of the time they have very marginal benefit, and you just need to get on with it.

Similarly, asymptotic complexity is not always an adequate approximation of energy consumption. More code leads to more cache misses. More code leads to more energy consumed transmitting that data over the internet, and wireless radios are energy-expensive. More code leads to more time parsing, compiling, optimizing, and then JITing code, which can utterly dominate actual runtime energy cost if the program is not long-lived - which describes a lot of webpages. Finally, very few webpages actually need to even be a "moderately complex React/Vue/Svelte app" - the essential complexity to the task being done is low, and webdevs add extra, unnecessary features for very poor reasons (showing off?). Take https://nabeelqu.co/understanding for example - 1.17M downloaded for 21K of text and two small images (one of which didn't load anyway). There's a 276K JavaScript file in there, among others - and the primary purpose of this page is to display a static document. Or, take any modern newspaper website you want - bloated with images, tracking, and JavaScript providing near-useless (and probably near-unused) features with a data-size:content-size ratio of over 100. If something is moderately complex, it would be much better as a native app in the majority of situations, which comes with significant performance improvements.