How to Build Superintelligence Inside Your Company
TL;DR
Y Combinator has transformed into an AI-native organization by building a shared 'organizational brain'—a centralized database and registry of 350+ internal tools that allows non-technical teams to encode workflows in English rather than code, moving beyond single-player coding agents to true organizational superintelligence.
🧠 The Foundation: Unified Context Layer 2 insights
Centralized data is the superpower
YC runs entirely on its own software with one Postgres database containing every critical table—companies, founders, financial transactions, and CRM notes—allowing agents to answer arbitrary complex questions that previously required hours of SQL writing.
Denormalize data for agent consumption
Organizations should consolidate data from disparate SaaS tools into a single schema or 'big table' format optimized for agent retrieval, similar to Google's Bigtable, enabling agents to 'see around corners' and interpret user intent across organizational knowledge.
🛠️ Infrastructure: The Tool Registry 3 insights
Build a shared tool registry
YC created an internal registry now containing over 350 tools that turns agents into useful workplace assistants, allowing domain experts to encode their own workflows as English prompts rather than requiring engineers to build deterministic software.
Start with database access
The breakthrough moment was giving agents read-only SQL access to the production database combined with model file reading capabilities, proving that security concerns often hamper capability more than actual risk.
The resolver pattern
Tool registries function as 'resolvers' (similar to OpenClaude's Skillify), where agents discover and call available capabilities via markdown entry points, creating a dynamic interface between human intent and system actions.
👥 From Single-Player to Multiplayer 2 insights
Solve the multiplayer harness problem
Current popular agents (Claude Code, OpenClaude, Hermes) are designed for individual use; the unsolved challenge is enabling organizational-level collaboration where teams share context and capabilities rather than operating in isolated silos.
Remove the engineer bottleneck
Traditional workflows require finance teams to explain processes to engineers who then code deterministic solutions; agent infrastructure allows domain experts to directly control software tools, eliminating back-and-forth delays and increasing the complexity of questions teams can ask.
⚙️ Organizational Best Practices 2 insights
Apply DRY and MECE principles
Maintain tool registries that are 'Don't Repeat Yourself' (DRY) and 'Mutually Exclusive, Collectively Exhaustive' (MECE), ensuring agents have clean, non-redundant capability sets that they can intuitively navigate without confusion.
Record all artifacts as building blocks
Treat AI as a foundational building layer rather than a copilot by recording organizational artifacts and workflows, creating a persistent 'shared organizational brain' that captures collective institutional knowledge and scales expertise across the company.
Bottom Line
Companies should stop using AI merely as a copilot and instead build a centralized infrastructure layer—combining a unified data context with a curated registry of internal tools—that allows every team member to encode their expertise directly into agent-accessible workflows.
More from Y Combinator
View all
Why Good Companies Go Bad (And How to Stop It)
Eric Ries exposes how standard corporate governance systematically destroys founder value, arguing that 'shareholder primacy' rules and expiring voting controls inevitably lead to mission drift, while offering frameworks for building enduring mission-controlled companies.
Personal AI Is the New Personal Computer
Y Combinator CEO Gary Tan details his return to software engineering after a 13-year hiatus, shipping hundreds of thousands of lines of code while running YC full-time by leveraging AI coding tools and developing "token maxing" methodologies that transform exhaustive research and development tasks into solo weekend projects.
How Razorpay Became India’s Largest Payments Company
Harshil Mathur recounts Razorpay's journey from a coding side project to India's largest payments platform, detailing their pivot from education to startups, the year-long regulatory wait that created competitive moats, and how surviving a bank crisis through radical customer transparency cemented their B2B trust foundation.
Beyond Bigger Models: Recursion As The Next Scaling Law In AI
Recursion at inference time—rather than simply scaling model size—may be the next breakthrough in AI reasoning. Recent research on Hierarchical Reasoning Models (HRM) and Tiny Recursive Models (TRM) demonstrates that recursive architectures using shared weights can solve complex reasoning benchmarks like Arc Prize with minimal parameters, outperforming massive traditional LLMs.