Cap the work in flight, not just cancel it.
Unbounded fan-out is the default failure mode of Go concurrency: every go statement is one more task in flight, with no ceiling and no owner. This article is about giving both.
Five Go HTTP stacks, one benchmark rig, real numbers.
Go ships a production-capable HTTP server in its standard library, and since Go 1.22 the built-in router handles method matching and path parameters on its own. The framework question has narrowed, not disappeared.
When llama-server beats Ollama
Ollama and llama.cpp are often compared as if they were rival inference engines. The real choice is between a managed model service and a toolkit you operate directly.
When to move from Ollama to vLLM
Ollama is one of the easiest ways to run a local language model, but convenience can conceal the moment when a local experiment becomes a shared inference service that needs better scheduling and observability.
Protocol security is who may act, not the model.
Prompt injection gets most of the security attention in LLM systems, and it deserves attention, but it is not the whole problem once agents start calling tools and delegating work to other agents.
Long-running A2A tasks outlive chat sessions.
Most AI agent demos still behave like chat completions with extra steps: you send a prompt, wait a few seconds, and get an answer back in one response.
A2A is not dead. It is just not universal.
Google’s Agent2Agent protocol, usually shortened to A2A, had a strange first year.
MCP gives agents tools. A2A gives agents peers.
AI agent architecture is starting to split into two layers.
A2A turns agents into network peers.
The A2A Protocol, short for Agent2Agent Protocol, is an open standard for communication between independent AI agent systems.
Stop duplicate side effects
Idempotency in distributed systems is the property that saves you after the network lies, the queue retries, the client panics, and the operator hits replay. In production systems, duplicate delivery is normal. Duplicate side effects are the bug.
Serve and swap LLMs without restarts.
For a long time, llama.cpp had a glaring limitation:
you could only serve one model per process, and switching meant a restart.
Self-hosted AI search with local LLMs
Vane is one of the more pragmatic entries in the “AI search with citations” space: a self-hosted answering engine that mixes live web retrieval with local or cloud LLMs, while keeping the whole stack under your control.
What actually happens when you run Ultrawork.
Oh My Opencode promises a “virtual AI dev team” — Sisyphus orchestrating specialists, tasks running in parallel, and the magic ultrawork keyword activating all of it.
Run GGUF models with the CLI and server
llama.cpp is a C/C++ inference engine for GGUF models: llama-cli for interactive chat, llama-completion for scripted prompts, and llama-server for an OpenAI-compatible HTTP API.
Airtable - Free plan limits, API, webhooks, Go & Python.
Airtable is best thought of as a low‑code application platform built around a collaborative “database-like” spreadsheet UI - excellent for rapidly creating operational tooling (internal trackers, lightweight CRMs, content pipelines, AI evaluation queues) where non-developers need a friendly interface, but developers also need an API surface for automation and integration.
Comparison of Chunking Strategies in RAG
Chunking is the most under-estimated hyperparameter in Retrieval ‑ Augmented Generation (RAG): it silently determines what your LLM “sees”, how expensive ingestion becomes, and how much of the LLM’s context window you burn per answer.
