Dan Rapp

We all hear the talk about AI research scientists at well-funded SF Bay Area startups receiving new hire equity awards worth tens of millions of dollars. Now, with 75% of the Forbes AI 50 list using Pave, we have the data to prove it, and remember, 10% of people are paid above the 90th percentile. Starting in mid-March, Market Data Pro is getting its biggest upgrade yet, and you'll be able to see data like the image I shared today. Here's what's new: → Pay mix benchmarks for commission-based roles (e.g. account execs) → Target bonus % and bonus eligibility benchmarks → Base salary and total cash data for employees with and without bonuses → Multi-select filters; you can now set custom filter ranges and blend headcount, invested capital, revenue, and valuation demographics in a single view → Peer groups; benchmark against a custom list of your closest competitors Underneath all of it: we've enhanced our entire calculation engine. Results are now generated on-the-fly, there's more data across every role, and we've strengthened our data dominance rules so no single company can skew the market. The result? We can go deeper into benchmarks that were previously nearly impossible to find — with confidence in the accuracy of every number. We're walking through all of it live tomorrow. Join us. Registration link in the comments. 👇 #AITalent #Compenation #CompensationStrategy #TotalRewards

60

4 Comments

Foundational piece by Mark Scianna explaining in actionable terms the OG FDE model that still powers Palantir Technologies’ growth today : https://lnkd.in/guyRt-nq The central tension : what if you can’t generate >$1M ACV per FDE? Are you stuck with impossible economics or weak product feedback loops? At Palantir for Builders, we’re exploring this frontier : 1. Working with companies that have builders with agency : teams with the speed-to-value mindset inside more fluid organizations that enable faster expansion. 2. Contributing to product from our cross-portfolio vantage point : pattern recognition across customers that accelerates the core platform 3. Deploying a symbiotic team of FDE-sourced customer success management and engineering to our customers 4. Building a community of customer builders : accelerating them with engineering guidance, customer success services, and events such as Buildcamps 5. Leveraging a fast-growing FDE level partner network to operationalize use cases with low product lift The question isn’t whether FDE can work on smaller deals. It’s whether we can systemically capture the same product insights and compounding leverage that made the model transformative at large enterprise scale.

4

In August, I started with a stubborn question: Why do we keep asking people to hand over identity just to prove one attribute? That question turned into weeks of iteration, a ton of guidance from professors. Marin Smillov & David M. Valladares, Ph.D. and just this recent weekend it became real: My presentation, “Proof Without Identity: Minimum-Disclosure, Minimum-Metadata Age Assurance for Digital Life,” was accepted to the 2026 Richard Macksey National Undergraduate Research Symposium at John Hopkins University (March 19–21). The core idea: Instead of “trust Big Tech to manage more sensitive data,” build frameworks that minimize what’s collected in the first place. Reveal only what’s necessary (ex: “18 years or younger”). Resist correlation across sessions. If you have experience in privacy/security patents, digital identity / age assurance, or presenting technical work to a humanities-first audience, I’d genuinely value any guidance especially on: - What to share publicly before filing a patent - How to frame the core contribution - How to stress-test the idea with the right questions. I’ll be heads-down preparing between now and March. I’ll share a full recap after the symposium with what worked, what didn’t, and what I learned presenting this to a humanities-first audience. Question for the security + privacy folks: what’s the hardest tradeoff in real-world age/eligibility assurance—usability, abuse resistance, or policy constraints? #Cybersecurity #Privacy #DigitalIdentity #Research #LearningInPublic

20

11 Comments

If you were watching GTC this week, you could feel the shift. Not just in scale - though the numbers are staggering - but in direction. Jensen Huang is now talking about $1 trillion in AI system demand, driven by agentic systems that don’t just generate… they act. ( https://lnkd.in/gyTrRwCS ) The conversation has clearly moved beyond copilots to systems that can operate in the real world. And that introduces a different kind of question - one that isn’t getting nearly enough attention: What does it take to actually trust these systems? Not just the model, but the full chain of trust - from silicon and firmware, through software, supply chain, updates, and into runtime behavior in the field. That’s why this conversation next week at RSAC matters. (https://lnkd.in/g6B4W7JD) At RSAC Conference 2026 - security’s preeminent global event, March 23-27 in San Francisco - we’re hosting the launch of Trust on the Edge with Bill Lattin and David Sequino, moderated by Evan Kirstel. (https://lnkd.in/g4k4SE48) What they bring is something the industry is about to realize it needs. This is a perspective shaped by decades securing systems across aerospace and defense, national infrastructure, medical devices, automotive and transportation, industrial and OT environments, semiconductors, and global telecom networks. Systems where failure isn’t a breach notification - it’s grounded aircraft, disrupted supply chains, compromised safety, or exposed critical infrastructure. Systems that have to remain trusted not for quarters, but for decades. That discipline - hardware-anchored trust, cryptographic enforcement, lifecycle governance - has largely lived at the edge of the industry. But as AI becomes agentic, and increasingly physical, it’s moving to the center. For C-suite leaders, this is where the conversation is going. Not just what AI can do, but how trust is established, enforced, and proven across systems that now have real-world consequences. If that’s on your mind, this will be a worthwhile discussion. 📍 RSAC 2026 - Booth #561 (South Hall) 📅 Wednesday, March 25 at 4pm 🍸 Happy Hour to follow - So stay for a libation! A timely conversation on what it takes to secure what’s coming next. Available now on Amazon - https://lnkd.in/ga-5eBdn #RSAC2026 #Cybersecurity #AI #DigitalTrust #PhysicalAI #TrustLifecycleManagement

14

Jensen Huang just made agent governance an enterprise boardroom problem. At GTC 2026, NVIDIA unveiled the Agent Toolkit — OpenShell, NemoClaw, and 17 partners building on a shared agentic foundation. GPUs made AI powerful. Agents make it autonomous. Governance is what makes it deployable. But runtime sandboxing is necessary, not sufficient. Futurum's analysts said it plainly — OpenShell is "a necessary component of agent trust, not a complete solution." The real challenge isn't monitoring agents after the fact. It's enforcing intent before execution — in real time, non-bypassably. At algedonic.ai, every agent action is wrapped with policy, telemetry, and feedback loops inspired by biological algedonic signals — the pain/pleasure mechanisms living systems use to self-regulate. The future isn't more agents. It's governed autonomy at scale. #AgenticAI #AIGovernance #GTC2026 #EnterpriseAI #Algedonic https://lnkd.in/gNCS_vuT

5

One of the things I've run into working with LLMs is the importance of clearly specifying and clarifying intentions. AI is great at pumping out a bunch of code but that code needs to be directed effectively. I've tried several agentic workflows and kept running into the same problem: lots of execution cycles but nothing steering the WHAT and WHY the way a Product Manager would. I've been looking for places to remove myself from development loops once solid plans are in place. So I extended the compound-engineering-plugin from Every Inc. to add a product-owner agent that sits inside the workflow, holds agents accountable to product requirements, and pushes back when things drift. Spending more time up front on planning and goals pays off. But it pays off even more when there's something in the workflow enforcing that alignment after you step away. Incidentally teams of humans operate the same way. Open sourced it here if you're curious: https://lnkd.in/g6cvc3Eu More technical and usage specifics and discussion on X: https://lnkd.in/gUW3nBFN

9

A study from Rice University, published in PRX Quantum Oct. 1, 2025, has found that energy transfers more quickly between molecular sites when it starts in an entangled, delocalized quantum state instead of from a single site. The discovery could lead to the development of more efficient light-harvesting materials that enhance the conversion of energy from light into other forms of energy. https://lnkd.in/ejMZu7_c

3