Ron Peled

@OpenSecurityAI has officially launched 🚀 We are thrilled to share one of the most remarkable releases from the Sola Security team yet. Now, anyone can evaluate their agents across diverse environments, identify performance gaps, and continuously refine them using the groundbreaking research that our AI research and Security teams released (and more will be in the near future). The inspiration for this project came a few months ago after our AI research team connected with a frontier lab. They wanted to leverage our benchmark and the Sola Security Brain to sharpen their model's accuracy on security-related questions. That conversation sparked a realization: we needed to build something far more impactful for the broader community. As agentic technology sees rapid adoption worldwide, it became clear that the industry lacked an unbiased way to test agents and measure their problem-solving capabilities across different domains. Establishing this standard is a foundational step toward engineering autonomous defense solutions. We have an incredible lineup of new features dropping soon, but we are equally eager to hear your ideas so we can elevate the community together. I’m calling all the security vendors to test and share the results in the leaderboard to help our community be better. Let the benchmarking begin!

A MUST READ ARTICLE Jack’s post details Block’s shift from traditional hierarchy (long an information-routing system built around human limits like “span of control”) to an AI-powered “intelligence” organization. A continuously updated company world model aggregates all internal artifacts (logs, decisions, code, progress), while a rich customer-signal model drives proactive decisions. Middle management is largely replaced by this system, leaving flatter roles: individual contributors building capabilities, DRIs owning outcomes, and player-coaches developing people. Security teams stand to benefit enormously from the same model. SOCs, incident response, and compliance functions are classic hierarchies where alerts, context, and escalations crawl through layers, creating delays that attackers exploit. An AI “security world model” could ingest endpoint data, threat intel, business context, and artifacts in real time, automatically correlating, prioritizing, and even composing initial responses turning security from reactive gatekeeping into proactive intelligence. Security ICs stay deep in the work (detection engineering, threat hunting), DRIs own specific risks or outcomes, and the intelligence layer surfaces only the truly novel or high-stakes calls to humans at the “edge.” The result: security moves at business speed instead of bureaucratic speed, with the same compounding advantage Block is chasing, provided the security team also hardens the AI system itself against attack and retains human judgment for ethical or existential decisions.

Small update, big headache removal. With over 7,000 workspaces and thousands of canvases being created every day, we heard from many customers about a critical need: the ability to fine-tune and update specific areas within an AI-generated dashboard. To meet this demand, @SolaSecurity developed "Focus Mode." This feature empowers builders to work on precise sections, ensuring the final result is even better and perfectly aligned with their vision. We are committed to moving fast and continuously improving for our customers. Can't wait to update what's coming in the coming weeks.

Last week, I shared a memo with the company about our trajectory. Hi team, We are on the verge of an industry shift. And the window to define it is right NOW. In the near future, every security team will operate through an AI agent as their primary interface with their entire tech stack. Need to find a stale Admin account? Map a misconfiguration in AWS? Automate compliance / audit evidence collection? The way security practitioners work is fundamentally changing, and it's changing fast. Every major technology shift produces a handful of defining companies. When cloud replaced on-premise, Databricks became a $100B+ business simply by making data accessible. The AI shift in security is bigger. It addresses a talent and complexity gap that humans alone can no longer bridge. For the first time in history, delivering elite, custom-tailored security posture isn't constrained by how many senior engineers you can hire. That's the market Sola is building for. We're not building another chat box. The chat box is 5% of the solution. The real work, our moat is: 1. The Security Intelligence Layer: A unified model of identities, assets, permissions, and posture across the entire stack, turning thousands of fragmented security data points into a shared, high-quality source of truth. 2. The Reasoning Engine: A layered system combining the edge of advanced technology such as graph computation, machine learning, and AI/Agents reasoning to evaluate risk across the environment and answer complex questions in minutes. 3. The Adaptive Security Platform: Every organization can shape Sola around its own, unique needs - turning answers into dashboards, workflows, micro-apps, and AI agents that reflect how they operate. We're not building yet another general LLM or a black-box platform with 'savior' pricing - we're building the Security Application Layer, the place where the real value accrues. We've made security self-serve, affordable, and collaborative. From a relatively simple AI assistant, Sola has evolved into a full AI-native Security agent and we're winning against bloated incumbents because we built the best end-to-end product, not the most expensive one. The future of cybersecurity isn't about how deep your pockets are, or how much code you can write. It's about getting the job done. To make that happen, we've assembled a team (all of you!) of top security practitioners, cracked engineers, high-agency AI researchers and Product managers and top GTM who’ve been in the trenches and are done with the old way of doing things. The opportunity to redefine cybersecurity is in front of us. It's ours to take.

The latest announcement by @AnthropicAI about Claude code for security vulnerabilities in code (nitter.cf/claudeai/status/202490…) marks a pivotal moment for our entire industry. This is the change we've been waiting for, especially as we see other industries going through similar shifts. The way we address security challenges is evolving, and this transformation will extend to other fields as well. This aligns with what we've believed in at @SolaSecurity from day one, and it motivates us to continue working and contributing more to our industry. Join the change at Sola.security.