HashiCast is the DevOps podcast produced by HashiCorp, an IBM company, and its intro calls it the self-proclaimed number one podcast about DevOps practices, tools and practitioners. That wink is a fair preview. It is a company podcast, but it plays like engineers comparing notes. Several strands share the name. The Changelog is the release strand, covering new releases across the HashiCorp portfolio from the previous month. Developer advocates Rosemary Wong and Michael Kosir are its recurring voices, and the titles read like a release ledger: 'CHANGELOG - HCP Terraform & Vault 1.21.2', 'CHANGELOG - AWS re:Invent & Terraform 1.14', 'CHANGELOG - Agent Skills & Terraform Enterprise 1.2'. The talk goes past the bullet points. One conversation explains why a coding agent needs the Terraform MCP server to reach current documentation, private registry modules and new provider versions, then notices an agent's output skipping ephemeral resources before conceding that this is not an AI podcast. Another follows a move from agent to the Vault Secrets Operator on Kubernetes, with a candid account of old worries that an operator's controller could leave things in a bad state. Tag policy compliance gets a blunt moment: delete untagged resources and see who screams, alongside a flat recommendation to tag every cloud resource. HCP Vault Radar comes up as HashiCorp's secret scanning and discovery tool. In an episode with Kareem, a senior developer advocate on the infrastructure side, Terraform Actions is presented as a way to trigger an Ansible playbook or a webhook for final instance configuration, without the templating friction of cloud-init scripts inside Terraform. That episode also works through writing a Terraform provider for a self-hosted service, operation by operation. Sign-offs point to changelogs and blogs in the show notes, invite feedback through a pull request, and ask people at AWS re:Invent or Microsoft Ignite to come say hi. Away from releases, Ming Zhao, a software developer at IBM focused on open-source technologies, joins for a conversation on document processing for AI. His case is about retrieval. Keep a document's structure through conversion, and the chunks headed for a vector database can follow that structure rather than slicing a wall of text. There is also a detour into which cheeses are acceptable. Keepin' It Secure is a separate strand, billed as the HashiCast show about security trends, cloud adoption challenges and security innovation, hosted by DevOps Rob and DevOps Adil. Rob says it grew out of their monthly one-to-ones, which they decided to open up, and a Batman nickname passing between them keeps the tone loose. For release tracking, start with the Changelog; for AI data preparation, the Ming Zhao episode; for the wider security picture, Keepin' It Secure.