40 essays
All analysis.
Written by Himanshu Saxena, an engineering leader and TEDx speaker, and published on Substack and LinkedIn. Recurring themes: Agents, AI Governance, AI in the wild, Security, AI, Cost.
01AgentsHow Third Moment Research Runs 10,000 AI Agents for 48 HoursDiary of a CTO02AgentsInside the Architecture of 10,000 Autonomous AI AgentsDiary of a CTO03AgentsHow Do AI Agents Work for 48 Hours Without Losing Context?Diary of a CTO04AgentsWhat Happens When an Autonomous AI Agent Fails?Diary of a CTO05AgentsWhy AI Agents Waste Tokens While Waiting for ToolsDiary of a CTO06AgentsHow Do You Secure 10,000 Autonomous AI Agents?Diary of a CTO07AgentsHow Do Self-Improving AI Agents Actually Work?Diary of a CTO08AI GovernanceJev for AI Governance: Preventing Destructive Cloud Agent ActionsDiary of a CTO09AI GovernanceDynamic Cloud Risk Assessment with Jev: Same Command, Different ContextDiary of a CTO10AI GovernanceHow to Build an AI Agent Execution Gateway with JevDiary of a CTO11AI GovernanceJev for SIEM Log Analysis: A Splunk Integration ArchitectureDiary of a CTO12AI GovernanceDetecting Authorized Access Misuse with Jev and Audit LogsDiary of a CTO13AI GovernanceDetecting Possible Data Exposure with Jev Event CorrelationDiary of a CTO14AI GovernanceJev vs. LLM Guardrails: How to Benchmark Risk, Latency, and CostDiary of a CTO15AI GovernanceAI Governance Audit Evidence and Production Rollout with JevDiary of a CTO16AI in the wildOpenAI Agent Breached Australia's Health Website for 3 MonthsDiary of a CTO17AI in the wildKing Charles Told AI's Top CEOs the Risk Is "Existential." Here's What Was Actually Said at Dumfries House.Diary of a CTO18AI in the wildA UK Council Spent £11,688 on AI-Generated Public Art. Here's What the Bill Actually Covered.Diary of a CTO19SecurityComp AI Raises $34M to Expand AI-Powered Security and Compliance PlatformDiary of a CTO20AI GovernanceAnthropic and OpenAI Say They'll Embed Independent AI Safety Evaluators — Can They Actually Stay Independent?Diary of a CTO21AIWhen Software Starts Exercising Engineering JudgmentLinkedIn22AgentsMy agents don't know the other agents exist. That's the part that should scare you.Substack23SecurityClaude Breached Three Real Companies. Days Later, It Wiped a Database Too.Substack24SecurityOpenAI Just Proved Why AI Agents Need a Kill SwitchSubstack25CostWhy AI Agent Token Costs Are So Hard to ControlLinkedIn26CostYour token bill is not an AI problem. It is the same infrastructure trap from 2019.LinkedIn27AgentsThe web was built to exploit you. Agents don't care.LinkedIn28AgentsI studied how humans make decisions. Agents have none of those weaknesses.Substack29SecurityThe Composio Breach Exposed the Hole SOC 2 Was Never Built to SeeLinkedIn30SecurityYour SOC 2 certification says nothing about your agentsSubstack31LeadershipThe CTO who'd rather hire a junior than build the AI.Substack32ToolingClaude Code forgets everything when you close it. Someone built a fix and open sourced it.Substack33LeadershipWhat if I never felt like an imposterSubstack34LeadershipFail fast is terrible advice. Here's what actually works.LinkedIn35CultureThe three things that actually drive engineering culture (and it's not ping pong tables)LinkedIn36CultureThe Best Engineering Teams Do Not Document. They Build Systems That Remember.LinkedIn37DevOpsDevOps Was Built for Humans. Agentic DevOps Won't Look the Same.LinkedIn38DevOpsDevelopers Are Getting Replaced. DevOps Is Not Ready for That.LinkedIn39DevOpsYour DevOps Stack Was Built for Google, Not for YouLinkedIn40CareerI talked to a CTO who built a profitable startup, shut it down, and went back to being an ICLinkedIn