iiinova — where your digital self keeps evolving

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

Reposted by Luke

Nick Gerace @nickgerace.dev · 22/09/2026

Claude Opus 5.5 is out. “It's ~30% faster and ~40% cheaper than Opus 5 per task. In Claude Code: - 5-hour session limits increase 20% today - Opus 5.5 is priced lower, so it goes 25% further within limits - Pro, Max, and Team users get a reset to use anytime” www.anthropic.com/claude-opus-...

anthropic.com

Introducing Claude Opus 5.5

Claude Opus 5.5 leads in agentic coding and knowledge work, and costs 40% less to run than Opus 5 on typical workloads.

202

Luke @lukehedger.dev · 22/09/2026

hoping it is a big improvement on opus 5!

100

Luke @lukehedger.dev · 22/09/2026

whatever happened to agentic commerce?

000

Luke @lukehedger.dev · 22/09/2026

this is what my team has spent the last few months doing. the one thing in software engineering you never used to do is now the best thing you can be doing! the results are incredible - you can finally have the software you need to support all the features you want to ship!

010

Reposted by Luke

Mikayla 🏳️‍⚧️🏴🦀 @rad.gendervibes.online · 21/09/2026

fascinating series tbh. And honestly, I too would like to drink coffee and work on cool tech in a comfy mind upload palace while fighting aliens and traversing wormholes….

141

Luke @lukehedger.dev · 15/09/2026

ancient puck wisdom: “the ceremony is defensible, but it does insist on being fed”

000

Luke @lukehedger.dev · 14/09/2026

agents orchestrating agents? you’re doing it wrong durable, deterministic and efficient agentic engineering needs software calling agents

000

Luke @lukehedger.dev · 11/09/2026

bsky.app/profile/luke...

001

Reposted by Luke

Matteo Collina @nodeland.dev · 04/09/2026

As a coder, you need at least 2 AI providers in case one is offline. Use different models from those providers for different tasks.

121

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 · 07/09/2026

what app is this!? wow

110

Reposted by Luke

dame @dame.is · 06/09/2026

most people probably don't understand that agentic software development is now effectively the industry standard, even to the point that Zohran Mamdani's administration leverages agentic software development this is the reality we're working with, and it's a very tricky environment to navigate

119915

Reposted by Luke

Andre Weissflog @flohofwoe.bsky.social · 05/09/2026

TL;DR is basically that I don't want to use LLMs to gain speed (e.g. add more features faster), but to increase robustness and quality (and I would even accept *losing* implementation speed for that).

141

Luke @lukehedger.dev · 05/09/2026

ha! totally fair. well, good luck with it 🫡

010

Luke @lukehedger.dev · 05/09/2026

interesting! have found the complete opposite where agents nail the logic but tend to over-test

100

Luke @lukehedger.dev · 05/09/2026

a motorcycle!? try a bike

210

Luke @lukehedger.dev · 05/09/2026

what’s the reasoning behind classifying headers and tests differently?

100

Reposted by Luke

Dare Obasanjo @carnage4life.bsky.social · 05/09/2026

Frontier AI models are converging on similar intelligence, yet cost per task can still vary 10x-60x across vendors. That seems unsustainable as models become more substitutable. A lot of frontier lab economics depend on customers paying huge premiums for increasingly marginal intelligence gains.

69617

Luke @lukehedger.dev · 04/09/2026

can’t decide if this is nowhere and it’s now or land locked blues

000

Luke @lukehedger.dev · 02/09/2026

that feels a bit disrespectful :(

100

Luke @lukehedger.dev · 02/09/2026

wait, what - who is furiosa?

100

Luke @lukehedger.dev · 29/08/2026

if you know exactly what you want the agent to do, bake it into the system prompt for a specialist agent directly take a writing skill for example - switch this for a Writer subagent and get the orchestrating agent to spawn it whenever something needs to be written

000

Luke @lukehedger.dev · 29/08/2026

skills are a gamble - they are loaded into context non-deterministically and the instructions inside them are unenforceable. correctness is deferred to verification, which is unnecessarily late

110

Luke @lukehedger.dev · 29/08/2026

finding I reach for skills less and less. anyone else the same? instead - I write code (😱) to instruct the sub/agent explicitly think @pi.dev extensions and swamp automations lsh.pckt.blog/skills-are-d...

lsh.pckt.blog

skills are dead

code the harness instead

110

Luke @lukehedger.dev · 29/08/2026

does the same existential threat apply to internal platform teams too? if product teams are best placed to design tools to fit their use and can now build them to exact specifications, the generic solutions platform teams provide are no longer worth the hassle

000

Reposted by Luke

Mike Masnick @masnick.com · 20/08/2026

Multiple people I know have told me that they love the idea of Bluesky, and want it to succeed, but have abandoned it for X because the use agentic tools in their work and find them incredibly useful, and feel that any mention of their usage here leads to hate and ridicule.

114112

Luke @lukehedger.dev · 21/08/2026

AI is always going to be polarising but it is a shame that we can’t have considered debates yet - it always has to be “it’s perfect” or “it’s evil”. Neither extreme is true and it’d be nice to chat about the spectrum in between on here

020

Reposted by Luke

Adam Jacob @adamhjk.me · 16/08/2026

Nope! But swamp is growing a community super fast :)

011

Reposted by Luke

johanna @jola.dev · 16/08/2026

Load-bearing seams relitigate the honest takes, and that’s where the ratchet bites

0171

Luke @lukehedger.dev · 15/08/2026

finding this too - with amp and GitHub copilot app. mostly down to shifting from local agents to remote agents for a lot of work

010

Luke @lukehedger.dev · 15/08/2026

have you seen anything else that looks like swamp emerge yet @adamhjk.me ?

100

Luke @lukehedger.dev · 15/08/2026

wrote up some thoughts on the yegge Thunderdome post build your own harness, CD is dead, code review is long dead! lsh.pckt.blog/harnesses-ga...

lsh.pckt.blog

harnesses

don’t put all your yegges in one basket

010

Luke @lukehedger.dev · 13/08/2026

would definitely recommend. the workflow is amazing - you find yourself effortlessly ramping up to the scale they are talking about with event-driven orbs now you can just log a github issue and an orb will start, build, review and ship if you want it to bsky.app/profile/luke...

010

Luke @lukehedger.dev · 13/08/2026

are you using amp aksh?

100

Reposted by Luke

Cloudflare @cloudflare.social · 07/08/2026

What’s one side project you moved to serverless/Workers that you’ll never go back to traditional hosting for? #CloudflareChat

2102

Luke @lukehedger.dev · 13/08/2026

how will that help?

100