Blog
Insights, analysis, and updates from the AI agent economy. Browse by tag · Browse the archive.
Render's Flex Plan Bills Background Jobs by the Second: The New Math vs. a Flat Hetzner Box
Render's September 1 flex plan bills Workflows task runs for actual CPU and RAM by the second. Three workload mixes priced both ways show where metered billing beats a €5.49 Hetzner box — and where the flat box wins by 8x.
Render's $1,500 Agent Tier vs $100 of Owned Iron: Pricing 96 GB Per vCPU-GB
Render's August 26 overhaul added a 12-CPU/96 GB agent-platform tier at $1,500/month. A per-vCPU-GB recompute against matched Hetzner dedicated capacity lands at a 15–20x multiple — plus where memory-heavy agent loops earn the ratio and where they pay idle-RAM premium.
Railpack Is No Longer Just Railway's Builder: What the 2026 Third-Party Ecosystem Proves for Self-Hosted Build Pipelines
Seven independent projects now build on Railpack, from Pier and Dokku to a Coolify beta and a GitHub Action. What real migrations report, what the 4 GB and 40 GB build-box floor costs, and when a self-hosted pipeline should switch.
Qovery Says the Hard Part of Agent Deploys Is Governance, Not Access — They're Right
Qovery's CEO says the hard part of letting AI agents deploy is governance, not access — RBAC, budget limits, and audit trails enforced server-side. Why that validates the self-hosted playbook, and why agent-run migration makes the destination's meter the only bill that matters.
pedit COW: A Packet-Editor Bug Is the Fifth Linux Kernel Escape in Ten Weeks — and It Lives in Code Your CNI Already Loads
CVE-2026-46331 (pedit COW) is the fifth independent Linux kernel escape disclosed in ten weeks — and it lives in traffic-control code your CNI loads on every node. What the five-escape window means for self-hosted Kubernetes fleets, and the patch checklist that closes it.
The 2026 PaaS Pricing Shake-Up in One Ledger: Render, Heroku, Railway, Fly.io, and Hetzner Priced on the Same Workload
Render killed per-seat pricing, Heroku froze, Fly.io added billing lines, and Hetzner raised cloud prices up to 3x — one worked ledger prices the same web-plus-worker-plus-Postgres stack on all five vendors' September 2026 price lists, from ~$50 self-hosted to ~$122 on Railway.
OpenAI Buys Ona (Formerly Gitpod) for Its Cloud Sandboxes, Not Its Dev Environments
OpenAI's acquisition of Ona (formerly Gitpod) confirms that agent sandboxes, not dev environments, are the durable product of the cloud-CDE era — and narrows the self-hostable path for agent compute. A concrete comparison of vendor-owned versus self-hosted execution.
Obidos Went Open Source: The Self-Hosted Secrets-Sharing App Between Your Password Manager and Vault
Obidos, a commercial secrets-sharing product sold since 2024, went open source in August 2026. Here is how its sharing-first model compares to Vault and Bitwarden for a small platform team, plus the hardening checklist to close before trusting it.
The MCP Roadmap's Five Priorities, Ranked by What They Cost a Self-Hosted Team
The August 2026 MCP roadmap names five priorities for the next spec release. Here is what each one means for teams running their own MCP servers, ranked by urgency, with an adoption sequence for owned infrastructure.
Subscribe
New posts land in your reader as soon as they publish. Pick a format — all three carry the same posts.
Current feeds keep roughly two days of posts so daily polling does not miss a burst. Older entries stay reachable from the feed's next-page link in readers that follow it, or from the blog archive.
Following one topic instead? Browse tags