This is how I feel about vibe coding. Any project I try that has any kind of complication has this immediate burst of progress. Things are amazing and it feels like a superpower. Then... as I add more complexity, things crash to a halt. The only projects that I think I can

Still facing these core three problems with LLMs: 1. Teaching the model new facts 2. Reliabling retrieving facts 3. Verifying correctness (i.e., explainability) Despite claims from today's leading AI labs, these foundational problems remain unsolved and continue to shape my AI

This one paper might kill the LLM agent hype. NVIDIA just published a blueprint for agentic AI powered by Small Language Models. And it makes a scary amount of sense. Here’s the full breakdown:

Its not the published rate card. it's the hidden tokens "thinking" generates. More reason SMLs like the ones we are building will win. youtu.be/mRWLQGMGY80?si…

we are so back

I'm going to keep repeating this until it becomes the first principle of AI adoption: It’s not about the last-mile shiny things like Agent visual tools. That’s the easy part. It’s about solving for model grounding, security, and cost. Full episode drops Thursday! CodeStrap's

You can build your own App Store/Marketplace with Foundry DevOps. 🚀 If you want to build an application and white-label it, you can easily package it, deploy it, and centrally manage the deployment while allowing customization for each one. Many IT groups use it to manage

“We want to empower you through malleable software to give you superpowers in procurement and allow you to spend more time using your expertise.” Agathe Verro shows Chad Wahlquist how Palantir is enabling human+AI teaming to drive ROI across procurement.

AI made this video. Zuck vs Sam -> fight

If you are a Foundry user or developer, be sure to check out FutureOps. I will be there as a speaker, and we are an official sponsor. linkedin.com/posts/futurops…

Yann LeCun: "If you are interested in human-level ai, don't work on LLMs"

This is how it's done. Elect leaders who know how to govern. Not politicians who only know how to lie for a living. Bravo to our heroes, our leaders, and our community.

CodeStrap's "Code and Connor" Episode 8 features our friends from i4C/ForgeSight out of Canada: Chris Atack, COO, and Brett Adams, Lead FDE. In this episode we discuss the fierce competition for Palantir-trained talent, US and Canadian politics around AI and our advantage over

Happy Friday, Palantir fam.

Andrew Ross Sorkin i’d argue you’re too easily moved by the choice words. “privilege” and “equality” are selected because they confuse the notion of fairness. another way to frame his statement “if you’re accustomed to succeeding due to hard work, ingenuity, risk/sacrifice, and, sure, some amount

Some nights are so perfect you need a double fearure. Oliva and Lonsdale Rare Leaf with a side of rum drinks. 🍻

Fortunes are made on the way down and collected on the way up. We've been building for the valley of despair for years. SMLs, cost, scale, governance. We are ready. LFG!

Handy feature in NodeJS: "These classes are used to associate state and propagate it throughout callbacks and promise chains. They allow storing data throughout the lifetime of a web request or any other asynchronous duration. It is similar to thread-local storage in other

While the rest of the world is waking up to domain-specific small models, we've been building for this future for years. Our goal: AI that builds AI. Today's progress highlights domain-specific solvers that support our engineers in making coded changes to our NX mono-repo using