How to Make a Coding Agent a General Purpose Agent - Harrison Chase

Why It Matters

The lack of secure, production‑ready infrastructure limits enterprise adoption of AI agents, keeping them in proof‑of‑concept stages. Solving these infrastructure challenges will unlock scalable, trustworthy agent products across industries.

Key Takeaways

• Infrastructure bottlenecks hinder agent scalability.
• Auth and permissions critical for production agents.
• Tool runtimes enable secure agent actions.
• LangChain and Arcade drive open‑source agent tooling.
• Enterprise adoption hinges on trustworthy agent execution.

Pulse Analysis

The excitement around large language model‑driven agents has often eclipsed a hard truth: scaling these systems requires more than clever prompting. While research breakthroughs improve reasoning and language understanding, real‑world deployments stumble on the plumbing that connects an agent to external services. Issues such as secure credential storage, reliable API orchestration, and consistent permission enforcement become the decisive factors that separate a laboratory demo from a production‑grade solution. This infrastructural gap explains why many high‑profile pilots stall before reaching broader market adoption.

In their keynote, Chase and Partee zeroed in on three core components—authentication, tool runtimes, and permissions—that form the backbone of trustworthy agent behavior. Robust auth mechanisms ensure agents can act on behalf of users without exposing sensitive data, while flexible runtimes allow dynamic execution of third‑party tools under controlled conditions. Permission frameworks then govern what actions an agent may perform, preventing overreach and regulatory breaches. By championing open‑source projects like LangChain and Arcade’s runtime libraries, they illustrate a path toward standardized, community‑vetted solutions that lower the barrier for enterprises to integrate agents safely into existing workflows.

The broader implication for the AI industry is clear: vendors that invest in hardened infrastructure will capture the next wave of commercial demand. Enterprises across finance, healthcare, and logistics are eager for agents that can automate complex tasks, but they require guarantees around security, compliance, and reliability. As open‑source ecosystems mature, they will provide the modular building blocks necessary for rapid, trustworthy agent deployment, turning today’s demos into tomorrow’s revenue‑generating products.

Original Description

Harrison Chase (LangChain) and Sam Partee (Arcade) Keynote at the Coding Agents Conference at the Computer History Museum, March 3rd, 2026.

Abstract //

Agents aren’t scaling because they’re smart—they’re stuck on infrastructure, and Harrison Chase with Sam Partee argue the real unlock is boring stuff like auth, tool runtimes, and permissions—because until agents can securely act as you, they’re just demos, not products.

Bio //

Harrison Chase

Harrison is the CEO and co-founder of LangChain, a company formed around the open source Python/Typescript packages that aim to make it easy to develop Language Model applications. Prior to starting LangChain, he led the ML team at Robust Intelligence (an MLOps company focused on testing and validation of machine learning models), led the entity linking team at Kensho (a fintech startup), and studied stats and CS at Harvard.

Sam Partee

Sam Partee is the CTO and Co-Founder of Arcade AI. Previously a Principal Engineer leading the Applied AI team at Redis, Sam led the effort in creating the ecosystem around Redis as a vector database. He is a contributor to multiple OSS projects including Langchain, DeterminedAI, LlamaIndex and Chapel amongst others. While at Cray/HPE he created the SmartSim AI framework which is now used at national labs around the country to integrate HPC simulations like climate models with AI.