~/if.codes$ls posts/

Notes on building AI that actually works.

RAG and semantic search, LLM integrations, agent pipelines and the backends underneath — practical posts written while shipping.

⌘K

Work with meTaking projects

I build what I write about — RAG and semantic search, AI integrations and agent pipelines, on backends in Go and Python.

Free · no commitment · I read every request myself

[0.0213, -0.0087, 0.0441, …] · 1536 dims

Latest · RAG & searchOct 20, 2026 · 6 min read

GPT-4o and GPT-4.1 leave the API on January 14: moving to gpt-6-luna without breaking your function calls

OpenAI shuts down four GPT-4o and GPT-4.1 snapshots, and their fine-tunes, on January 14, 2027. Its named replacement, gpt-6-luna, is a reasoning model with one function-calling catch in Chat Completions. Here are two migration paths, a replay test and a token-bill check.

#openai#migration#llm-api#function-callingRead the post

Infrastructure

Oct 19, 2026 · 5 min readSwapping Docker Desktop for Apple's container: one VM per container on your Mac, what runs unchanged from Docker Hub, and where you still need DockerApple's container CLI has been stable since 1.0 in June and is at 1.5.0. A hands-on swap: install, build, run Postgres from Docker Hub, reach it by name, and the gaps (no Compose, slower cold starts, memory that is not handed back) that decide whether Docker stays.#docker#macos#containers#devtools

Infrastructure

Oct 19, 2026 · 5 min readGitHub's agentic autofix is GA: batching 25 CodeQL alerts into one draft PR, what it bills, and how to review the fixesAgentic autofix for code-scanning alerts went GA on October 15: Copilot explores the repo, fixes, re-runs CodeQL and opens a draft PR for up to 25 alerts at once. Who can use it, how to turn it on, where the cost lands, how to pick a first batch with gh, and a review checklist.#github#codeql#security#code-review

Infrastructure

Oct 18, 2026 · 7 min read'Yes to all' is a permission policy: writing Claude Code allow, ask and deny rules that hold up against compound commands and wildcardsEvery 'Yes, and don't ask again' writes a rule. Turn that habit into a reviewed .claude/settings.json: deny beats ask beats allow, why Bash(git * main) also allows git push, how compound commands are split, and when dontAsk or the sandbox beats bypassPermissions.#claude-code#permissions#security#developer-tools

AI integrations

Oct 17, 2026 · 6 min read710 ms to 150 ms for 100 requests: what Go 1.27's HTTP/1 connection reuse fixes, and the body-draining habit to keep for older clientsGo 1.27 drains unread HTTP/1 response bodies for you, so a closed-but-unread body no longer kills connection reuse. A runnable test that counts reused connections, the two patterns that broke reuse, and a capped drain helper for Go 1.26 and older.#go#http#performance#networking

AI integrations

Oct 16, 2026 · 6 min readClaude Code subagents get a 5-minute cache even on a subscription: when to set it to 1h, and how to checkOn a subscription your main Claude Code conversation gets a one-hour prompt cache, but subagents, workflows and compaction stay at five minutes. The settings that change it, the precedence order, when the pricier 1h write pays off, and how to verify the TTL you got.#claude-code#prompt-caching#subagents#llm-costs

All posts

// what I write about — and build

Three topics. Also three things I build for clients.

Every topic has posts to read and a free quote if you want it built for your data and systems.

RAG & semantic search

Search and assistants that answer from your own documents — with sources, and a handoff when they’re not sure.

Embeddings & vector search
Hybrid keyword + meaning search
Assistants with citations

AI integrations

LLMs wired into the systems you already run — tickets, CRM, catalog, inbox — with costs and quality you can see.

Summaries, drafts, classification
Translations and content generation
APIs, webhooks, your own model

Agent pipelines

Multi-step agents as inspectable workflows — they plan, act and check, and a person approves anything that matters.

Workflows you can open and trace
Human approval steps
Event-driven, retry-safe backends

New posts by RSSOne entry per post, no tracking — works with any feed reader.Subscribe to the feed

// about

One engineer. I write about what I build.

Software engineer working on RAG, AI integrations and agent pipelines on Go and Python backends. Author of the Redelay framework and the FlowDSL workflow engine. Remote, EU time zones.

GoPythonRAGembeddingsLLM APIsagentsKafka / NATSMongoDB

RSS feed

// free quote

Read something you need? I’ll quote it for free.

RAG, AI integrations, agents or the backend underneath — tell me what you have and what should change. I read every request myself.

Free · no commitment

Tell me what you have. I’ll tell you what it takes.

1Describe the projectA few sentences is enough — about two minutes.

2I review itI read it myself and may ask a follow-up question.

3You get a free quoteScope, approach and estimate — yours to keep, no strings.