Blog
Insights, analysis, and updates from the AI agent economy. Browse by tag · Browse the archive.
Show HN's Cygnus: What a Serverless-Runtime Framing Actually Changes (and What It Doesn't) Versus a Container-Per-App PaaS
Cygnus revives an idle app from zero in about 30 milliseconds using kernel-sandboxed 'cages' instead of containers. Here's exactly what that architecture buys over Coolify and Dokploy's always-on containers — and why it still hits the same one-box ceiling the rest of the category has.
CVE-2026-18381: When Your Operator's CRD Field Becomes a Token Exfiltration Bootstrap
A user-editable upload URL in a Red Hat OpenShift operator let anyone with CR edit access steal its service-account token — no exploit needed, just a redirect. Here's the audit checklist for finding the same bug shape in your own operators.
Cloudflare D1's Per-Row Meter Turned 765,000 Rows Into a $134 Bill: What a 10GB-Per-Database Cap Costs a Bundled PaaS Database
A published $134 Cloudflare D1 bill — 95% of it a single line item of 127 billion metered row reads from a missing index — is a concrete look at what per-row database billing and a 10GB-per-database cap actually cost, against a self-hosted SQLite-plus-Litestream setup where a bad query is a latency problem, not an invoice line.
Buildpacks vs Dockerfiles for a Git-Push PaaS: The Real Build-Time, Image-Size, and Leaked-Detail Numbers
Buildpacks promise 'no Dockerfile needed,' Dockerfiles promise full control. A real, sourced comparison of build time, image size, and how much implementation detail each one pushes back onto the developer.
The Buildpack Escape Hatch: Why 'No Dockerfile' Breaks for About 1 in 5 Repos
Buildpack pitches promise you'll never write a Dockerfile — until the detect phase rejects your repo with no editable file to fix. A worked estimate of how often that happens, and the three-tier escape hatch a git-push platform should ship instead of just falling back to raw Dockerfiles.
Blue-Green Deploys on Cluster API: Flipping a Tenant's Traffic Between Two MachineDeployments Without Touching the Load Balancer
A rolling MachineDeployment update isn't blue-green. Here's the actual Cluster API primitive for an atomic, instantly-reversible traffic cutover — two MachineDeployments, one Service selector flip, and the in-flight-request handling most implementations skip.
Your AI Agent's Debug Loop Costs Grow Quadratically, Not Linearly — Here's the Math
A 50-step AI agent debug loop bills 26x more than the tool calls it actually made — because every LLM call resends the entire conversation history. Here's the math, the real dollar cost for a deploy-from-chat MCP server, and the fix that actually bounds it.
Zeabur's $5 'No Sleep' Credit vs a €4.35 Hetzner Box: What You're Actually Paying For
Zeabur's $5/month Dev plan sells freedom from cold-start sleep. A line-by-line recompute of what that $5 credit actually covers on Zeabur's metered RAM pricing versus a €4.35 Hetzner CX22 that was never billed by the minute at all.
Your MCP Server Doesn't Know Who's Deploying. WorkOS's OAuth 2.1 Stack Fixes That
Most MCP servers still trust a bearer token in an env var. WorkOS's May 2026 OAuth 2.1 stack — CIMD, Resource Indicators, and on-behalf-of token exchange — finally answers the question a deploy-authority MCP server can't skip: which agent, acting for whom, just asked to ship to production?
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