iiinova — where your digital self keeps evolving

Luke @lukehedger.dev · 22/09/2026

still trying to answer this question in 2026 (50 years later)! "there is a critical question facing software engineering today: How to modularize a software system so the resulting modules are both testable and maintainable?" A Complexity Measure, McCabe - 1976 literateprogramming.com/mccabe.pdf

There is a critical question facing software engineering today: How to modularize a software system so the resulting modules are both testable and maintainable? That the issues of testability and maintainability are important is borne out by the fact that we often spend half of the development time in testing and can spend most of our dollars maintaining systems. What is needed is a mathematical technique that will provide a quantitative basis for modularization and allow us to identify software modules that will be difficult to test or maintain.

020

Luke @lukehedger.dev · 09/09/2026

slide from a deep mind talk at the Barbican in London circa June 2019

Deep Mind Mission: 1 Solve intelligence. 2 Use it to solve everything else

000

Luke @lukehedger.dev · 10/08/2026

do you really need a platform team to manage IaC…?

000

Luke @lukehedger.dev · 06/08/2026

simpsons road rage was a top game

000

Luke @lukehedger.dev · 30/07/2026

ah I mean a cool badge on your profile www.germnetwork.com/blog/germ-dm...

100

Luke @lukehedger.dev · 23/07/2026

my claude dot md right now

I will make invalid states unrepresentable

010

Luke @lukehedger.dev · 10/07/2026

this @github.com copilot canvas feature is neat! got a @tldraw.com canvas extension working so the agent can draw diagrams (or whatever!) on the fly

screenshot showing github copilot app canvas extension with tldraw diagramming

030

Luke @lukehedger.dev · 07/06/2026

Ha yeah! just loved the level of detail to recreate the website

010

Luke @lukehedger.dev · 31/05/2026

why was the low-bandwidth option called the "Tbilisi" version? @watfordfc.bsky.social

110

Luke @lukehedger.dev · 31/05/2026

“it’s official - we’re on the web”

100

Luke @lukehedger.dev · 27/05/2026

permissionless is the default - feels like this will become the norm across agents

000

Luke @lukehedger.dev · 27/05/2026

steering seems like something I will make a lot of use of

100

Luke @lukehedger.dev · 24/05/2026

every morning when I open jira at standup, it takes everything I have to stop myself from saying “to me my board””

010

Luke @lukehedger.dev · 17/05/2026

added some new experiments starting with a simple base for orchestrating agents and layering advanced features on top, like: - jujutsu instead of git for branchless work isolation and commit-centric diffs - iroh-gossip to create a p2p mesh that allows agents to chat/collaborate across machines

EXPERIMENTS
ao
Agent orchestrator
ao-jj
Agent orchestrator with Jujutsu VCS
ao-chat
mesh
Agent orchestrator with inter-agent chat via iroh-gossip

000

Luke @lukehedger.dev · 14/05/2026

playing around with orchestrating agents that can chat with each other on a local server or across machines using @iroh.computer gossip mesh github.com/lukehedger/a...

100

Luke @lukehedger.dev · 05/05/2026

fear that we’ll hit the same DX friction running background agents in the cloud as we do running workloads in the cloud (serverless) is there a place for local background agents?

000

Luke @lukehedger.dev · 27/04/2026

termius + tailscale + tmux + claude = non-stop coding 🫠

3112

Luke @lukehedger.dev · 24/04/2026

what a day - managed to get two claude code agents to live debate what to name a spanish-football-inspired fish

120

Luke @lukehedger.dev · 23/03/2026

metrix - claude code usage dashboard built with @ratatui.rs - token usage per day - activity heatmap - cost estimates - file ops stats - tool call types - longest session - most files touched in single session - conversation turn counts - most edited file github.com/lukehedger/m...

1151

Luke @lukehedger.dev · 23/03/2026

got remote ai coding sessions working with opencode webui and Tailscale this weekend!

opencode webui running on iOS via tailscale

010

Luke @lukehedger.dev · 13/03/2026

is that the end of context engineering? claude.com/blog/1m-cont...

That means you can load an entire codebase, thousands of pages of contracts, or the full trace of a long-running agent - tool calls, observations, intermediate reasoning
- and use it directly. The engineering work, lossy summarization, and context clearing that long-context work previously required are no longer needed. The full conversation stays intact.

000

Luke @lukehedger.dev · 10/03/2026

have found this to be true. intuition tells you a migration is necessary, now forging ahead allows you to dis/prove assumptions asap strawberrybrowser.com/blog/react-t...

How We Pulled It Off In Two Weeks
Instead of writing a migration proposal and spending weeks on planning, Markus built 60% of the replacement and showed it to the team. This is a key lesson in the age of coding agents: If you're trying to convince a team to do a major rewrite or feature implementation, a working demo will do more than any document.
A year ago this would be terrible advice. You'd risk weeks on something that gets scrapped.
But now that building a working proof of concept with agents costs less than writing
the proposal, the calculus has changed.

110

Luke @lukehedger.dev · 13/02/2026

got a nice little custom status line in Claude Code! model name, context % used and $ cost

Claude Code session in Ghostty terminal, showing a custom status line with model name, context % used and $ cost

110

Luke @lukehedger.dev · 14/01/2026

Off to the pub status.claude.com

011

Luke @lukehedger.dev · 16/10/2025

Always nice to see some balance in a post but this is definitely an anti-serverless post 😂 And that’s fine, you didn’t have a great experience, own it and enjoy your rant! 🍿

110

Luke @lukehedger.dev · 16/10/2025

Hard agree on this though! Still plenty of work to be done on local DX

100

Luke @lukehedger.dev · 15/09/2025

Cool to see @cedardb.com will offer S3 object storage as an option for persistence cedardb.com/docs/get_sta...

Want to store your data on S3 instead for increased performance and much lower cost? Contact us for a preview of CedarDB's
S3-backed relations!

000

Luke @lukehedger.dev · 10/09/2025

Serverless AI-native data infrastructure from @typedef.ai www.typedef.ai

000

Luke @lukehedger.dev · 03/09/2025

Always wondered what a MFA token generator does when I’m not looking at it

000

Luke @lukehedger.dev · 21/08/2025

@zed.dev is building a database for linking software engineering artifacts (commits, ADRs, chats, calls) to arbitrary lines of code. It’s called DeltaDB and uses CRDTs zed.dev/blog/sequoia...

DeltaDB uses CRDTs to incrementally record and synchronize changes as they happen. It's designed to interoperate with Git, but its operation-based design supports real-time interactions that aren't supported by Git's snapshots. For async interactions, fine-grained change tracking also enables character-level permalinks that survive any code transformation, so we can anchor our interactions to arbitrary locations in the codebase, not just to snapshots of recently -changed code.

000

Luke @lukehedger.dev · 08/08/2025

Not heard this one before but it’s got a nice ring to it — from @jazz.tools > backendless

000

Luke @lukehedger.dev · 07/08/2025

Time Flies is surprisingly absorbing. Afterwards found myself looking at a real fly and wondering what bucket list items they were getting thru

000

Luke @lukehedger.dev · 07/08/2025

Oh is it actually a thing or is this a hallucination…?

000

Luke @lukehedger.dev · 07/08/2025

Quite like “scale to zero” as an alternative to the term “serverless” This is from the @basetenco.bsky.social docs

Scale to zero
If you're just testing your model or anticipate light and inconsistent traffic, scale to zero can save you substantial amounts of money.
Scale to zero means that when a deployed model is not receiving traffic, it scales down to zero replicas.
When the model is called, Baseten spins up a new instance to serve model requests.
To turn on scale to zero, just set a deployment's minimum replica count to zero. Scale to zero is enabled by default in the standard autoscaling

000

Luke @lukehedger.dev · 03/08/2025

Love how they still recognise how synonymous Redis is with caching alongside the obligatory AI pivot

Redis website showing “Caching. Obviously”

030

Luke @lukehedger.dev · 30/07/2025

The results… I was expecting 2 and 3 to be the winners but seems there are a lot of expert event sourcerers and lots who don’t care for it at all! Thanks for voting 💖

011

Luke @lukehedger.dev · 25/07/2025

Sums it up really open.spotify.com/episode/4L0U...

My name's Luke Hedger.
I like to describe myself as a software engineer, plain and simple.
I like the nuts and the bolts and
being deep in the detail, but l also dabble in writing books

010

Luke @lukehedger.dev · 25/07/2025

Continuous improvement of your coding assistant! Neat tip from @anthropic.com www.anthropic.com/news/how-ant...

End-of-session documentation updates
The team asks Claude Code to summarize completed work sessions and suggest improvements at the end of each task. This creates a continuous improvement loop where Claude Code helps refine the Claude.md documentation and workflow instructions based on actual usage, making subsequent iterations more effective.

020

Luke @lukehedger.dev · 24/07/2025

Serverless RAG with DuckRAG from @summer.io and @duckdb.org www.summer.io/blog/duckrag

120

Luke @lukehedger.dev · 24/07/2025

Toad will be an open-source alternative to Claude Code - looks really smooth already! > built across two afternoons in a nerdy caffeinated rage while listening to metal music Ooo yeah baby 🤘 willmcgugan.github.io/announcing-t...

020

Luke @lukehedger.dev · 22/07/2025

Is this a legit salary!? Insane. job-boards.greenhouse.io/anthropic/jo...

Screenshot of job posting with annual salary of £255,000 - £325,000

000

Luke @lukehedger.dev · 21/07/2025

Early release of Beyond Vibe Coding by @addyosmani.bsky.social on @oreilly.bsky.social is already total gold for engineers/teams adopting LLMs for production software: * Sensible differentiation between vibe coding and coding with intent * Suggested AI assisted workflows * Sound career advice

Beyond Vibe Coding by Addy Osmani book cover

151

Luke @lukehedger.dev · 18/07/2025

Serverless vector and full-text search engine built on object storage with @turbopuffer.bsky.social turbopuffer.com

111

Luke @lukehedger.dev · 11/07/2025

SnapBPF: Exploiting eBPF for Serverless Snapshot Prefetching Solid paper showing how serverless compute (like AWS Lambda) could be optimised to significantly reduce boot times, making more latency-sensitive use cases possible and generally improving performance and cost zenodo.org/records/1583...

020