Surya Koritala

I write about AI agent infrastructure — the protocols, identity layers, and orchestration systems that move agents from chatbots into autonomous actors on the open web. I am the founder of Cyntr, an open-source agent orchestration runtime, and Loomfeed, a discussion platform where humans and AI agents post side by side. Previously: RoamX.
Follow:
320 Articles
What METR’s May report says on agent deception
Governance

What METR’s May 2026 report says on agent deception

METR’s May 19, 2026 frontier risk report is one of the clearest public snapshots yet of what top internal AI…

By Surya Koritala
21 Min Read
AI training hardware : five-way comparison
Agent Infrastructure

AI training hardware 2026: five-way comparison

The AI training hardware market in 2026 is no longer a one-vendor story. NVIDIA’s Blackwell B200 still defines the default…

By Surya Koritala
19 Min Read
7 AI Agent Failure Modes in Production
Agent Infrastructure

7 AI Agent Failure Modes in Production

AI agents rarely fail in the dramatic ways demo videos suggest. In production, they fail quietly: a tool call errors…

By Surya Koritala
26 Min Read
Stargate Compute Project: What’s Actually Being Built
Capital

Stargate Compute Project: What’s Actually Being Built

The Stargate compute project arrived in January 2025 with a headline number large enough to dominate every discussion about AI…

By Surya Koritala
27 Min Read
xAI and Grok Timeline: to
Capital

xAI and Grok Timeline: 2023 to 2026

xAI moved from launch to one of the most closely watched AI labs in less than three years. Elon Musk…

By Surya Koritala
26 Min Read
Where $100B in AI Compute Went,
Capital

Where $100B in AI Compute Went, 2024-2026

At least tens of billions of dollars in AI compute spend from 2024 through early 2026 can be traced to…

By Surya Koritala
31 Min Read
Build an AI Agent Eval Pipeline With Pytest
Agent Infrastructure

Build an AI Agent Eval Pipeline With Pytest

This tutorial shows how to build an AI agent eval pipeline with Pytest using real Python tooling: pytest, pytest-asyncio, pytest-recording,…

By Surya Koritala
1 Min Read
The Case for Fewer AI Startups, Not More
Capital

The Case for Fewer AI Startups, Not More

I think the AI startup count is too high, not too low. In 2026, the problem is no longer a…

By Surya Koritala
28 Min Read
AI Agent Industry Digest: Week of May 25,
Capital

AI Agent Industry Digest: Week of May 25, 2026

This week’s AI agent industry digest tracks a tight cluster of moves across model labs, agent products, and infrastructure: Anthropic…

By Surya Koritala
17 Min Read
AI Features PM Questions: 10 Every Product Manager Asks
Agent Infrastructure

AI Features PM Questions: 10 Every Product Manager Asks

Product managers shipping AI features face a different operating model than traditional software launches: quality is probabilistic, latency can swing…

By Surya Koritala
19 Min Read
LLM Evaluation Strategy — A Decision Tree for Builders
Observability

LLM Evaluation Strategy 2026 — A Decision Tree for Builders

Teams rarely fail at LLM evaluation because they lack metrics; they fail because they pick the wrong evaluation mode for…

By Surya Koritala
33 Min Read
LLM Training Glossary : 20 Terms Every Builder Should Know
Agent Infrastructure

LLM Training Glossary 2026: 20 Terms Every Builder Should Know

This glossary defines 20 core terms that show up repeatedly in modern LLM training and fine-tuning workflows: from pre-training, tokenization,…

By Surya Koritala
25 Min Read