SumGuy's Ramblings
The art of wasting time.
Docker, self-hosting, AI/LLM, Linux, and DevOps — explained by someone who learned the hard way. No fluff, no enterprise jargon, just practical stuff that actually works on real hardware.
Recent Posts
-
Jitsi Meet Self-Hosted
Self-host Jitsi Meet as your private Zoom alternative: encrypted video calls, no SaaS bill. Here's how to deploy it with Docker without it falling over.
13 min read -
COSMIC vs KDE vs GNOME in 2026
Three Linux desktops, three philosophies. COSMIC is opinionated Rust. KDE is maximally customizable. GNOME is workflow-first. Here's how to pick.
10 min read -
Hailo-8 vs Coral: AI Accelerators for the Edge
Hailo-8 vs Google Coral: which edge AI accelerator actually wins for Frigate NVR, Whisper transcription, and local home lab inference workloads in 2026?
11 min read -
Matrix: Synapse vs Dendrite vs Conduit
Three Matrix server implementations — Python's Synapse, Go's Dendrite, Rust's Conduit. Which one runs your federated chat without eating your homelab alive?
12 min read -
AI Swarm Audited My 840-Post Blog
I pointed a parallel swarm of AI agents at 840 technical posts to fact-check and refresh them — the architecture, the token bill, and the guardrails.
12 min read -
Mini PCs With 2.5GbE: Picking the Right Box
A 2026 home lab guide to mini PCs with built-in 2.5GbE: NIC chipsets, idle power consumption, thermals, and which small form-factor box fits your use case.
13 min read -
Minisforum vs Beelink vs Trigkey Homelab
Three mini PC brands dominate homelab 2026: Minisforum, Beelink, Trigkey. Here's how to pick the right Ryzen or N100 box for your cluster.
16 min read -
OPA & Gatekeeper: Policy as Code
Open Policy Agent and Gatekeeper enforce rules across Kubernetes — block privileged pods, require labels, restrict images. Rego basics and real homelab constraints.
11 min read -
SigNoz vs Uptrace Self-Hosted Observability
Two OpenTelemetry-first observability platforms backed by ClickHouse. Here's how SigNoz and Uptrace compare for self-hosted homelab and small-team use.
13 min read -
Used GPU Buying Guide for Home Lab LLMs
Your no-BS guide to buying a used GPU for local LLM inference in 2026. Budget tiers from $200 to $1500 plus, real VRAM math, and tips to dodge the scams.
12 min read -
Longhorn vs Rook-Ceph
Two ways to give your K8s cluster persistent block storage — Longhorn for simplicity, Rook-Ceph for scale. Real homelab tradeoffs, not vendor marketing slides.
11 min read -
Gateway API vs Ingress in 2026
Ingress is the annotation-soup classic. Gateway API is the role-oriented future with proper L4/L7 separation. Here's when to migrate and when to wait.
11 min read
