Essays
Writing from Conor Bronsdon — essays and episode deep-dives on AI infrastructure, agents, and the people building all of it, drawn from the conversations on the show. New pieces land on the newsletter first.
· on the newsletter ↗
Stop reading diffs. Run cheap checks first, get a second model family’s review, and make every finding show its evidence.
Companion episode · EP 56 AI CodingAI EngineeringAI Evaluation & Reliability
The essays develop a thread from the show into a written argument. Some focus on a guest's approach to a problem; others connect several conversations to a question about building and operating AI. Entries marked “on the newsletter” open the original publication there.
Two ways in
Two routes into the writing are Your AI Agent has an Amnesia Problem, linked to Richmond Alake's memory conversation, and the essay on Intercom's model choice, linked to Fergal Reid's episode. Read the agent memory and open source AI hubs for the surrounding conversations, or use fine-tuning vs RAG vs prompting to frame an experiment of your own.
For conversations about building and evaluating agents, browse the best AI agent podcast episodes from this show and others.
10 in all, newest first
· on the newsletter ↗
Laurie Voss on product engineering, customer context, and the path for junior developers as agents take on more coding work.
EP 74 : AI Codes: Product Engineers Decide What to Build | Laurie Voss, Arize AI CodingAI Engineering
· on the newsletter ↗
What happened when I hired an AI agent to write a personal essay, and what the collaboration still required from me.
· on the newsletter ↗
Open source flips the classic disruption story for AI agents — and changes who captures the agent layer.
EP 56 : 250,000 Lines of Code/Week: Inside an AMD VP's Agent-First Workflow | Anush Elangovan Open Source AIAI AgentsEnterprise AI
· on the newsletter ↗
Agents forget everything between sessions. Why memory is becoming the next battleground in the AI stack.
EP 53 : Agent Memory: The Last Battleground in the AI Stack | Richmond Alake, Oracle Agent MemoryAI AgentsContext Management
· on the newsletter ↗
More context isn't better context. How context poisoning degrades agents — and what to do about it.
EP 52 : Context Poisoning is Killing Your AI Agents: How to Stop it | Michel Tricot, Airbyte Context ManagementAI AgentsRAG & Retrieval
· on the newsletter ↗
Behind the headline: what Block's restructuring actually says about AI and the future of work.
EP 48 : How Block Deployed AI Agents to 12,000 Employees in 8 Weeks w/ MCP | Angie Jones AI AgentsEnterprise AI
· on the newsletter ↗
What an agent-first working day actually looks like when a builder hands most of it to an AI coworker.
EP 50 : I Built an AI Coworker That Runs 90% of My Day | Sterling Chin, Postman AI Agents
· on the newsletter ↗
The economics of model choice at production scale, from the conversation with Intercom's Fergal Reid.
EP 49 : How Intercom Cut $250K/Month by Ditching GPT for Open Models | Fergal Reid Open Source AIModel TrainingAI Evaluation & Reliability
· on the newsletter ↗
Where AI development heads next — Gemini, robotics, and space — from the conversation with Google DeepMind's Paige Bailey.
EP 47 : Gemini 3 & Robot Dogs: Inside Google DeepMind's AI Experiments | Paige Bailey AI CodingAI AgentsAI Energy & Data Centers
The newsletter
New essays and episodes from Conor Bronsdon, on the newsletter first. 1,200+ subscribers. Up to once a week, no spam, unsubscribe anytime.