ThreatPulse

Inspiration

What it does

About This Project

The Inspiration

During a recent Joe Rogan podcast, Jensen Huang, CEO of NVIDIA, discussed the critical importance of collaboration in cybersecurity. He emphasized that in today's threat landscape, we can't operate in silos—we need to work together and leverage collective intelligence to stay ahead of evolving threats.

This struck a chord with me. I started thinking from first principles: What's the fundamental problem security teams face when a new CVE is disclosed?

The Problem

Security teams race against time when vulnerabilities are discovered. But there's a critical gap: Official CVSS severity scores from NVD take 1-7 days to publish. During that window, teams have to make high-stakes decisions with incomplete information—allocate resources, prioritize patches, coordinate incident response. They're stuck asking: "How critical is this really?"

The First Principles Approach

Instead of waiting for centralized authorities, what if we could tap into the collective intelligence of the cybersecurity community? The security community on X (Twitter) responds immediately to CVEs—experts analyze, debate, and share insights within hours.

I asked: Can we extract actionable signals from this collective response?

The Solution: ThreatPulse

ThreatPulse uses reinforcement learning to predict CVE severity by analyzing X community engagement patterns. The model learns from legendary vulnerabilities like Log4Shell, Heartbleed, and EternalBlue—discovering that expert engagement velocity and consensus predict critical severity with 77% accuracy.

Our RL model gives security teams a 22-hour head start. That's crucial time to:

• Prepare patches and workarounds
• Allocate security resources
• Coordinate incident response
• Protect critical infrastructure

And the model continuously learns from every new CVE, improving its predictions over time.

Built with xAI & Grok

ThreatPulse leverages xAI's Grok models and X API integration to process community signals at scale, embodying the collaborative approach to cybersecurity that Jensen advocated for.

Built With

• drizzle
• express.js
• grok
• node.js
• react
• sqlite
• tailwind
• typescript
• vite
• x