November 2025 (version 1.107) What's new in the Visual Studio Code November 2025 Release (1.107).

Copilot in vscode is also adopting it (experimental setting for now): code.visualstudio.com/updates/v1_1...

Does being a celebrant for weddings and funerals count as a trade? I reckon I’d enjoy the challenges of building trust and psychological safety quickly and supporting people through the big life milestones

Agreed that’s a great read! I don’t love the anthropomorphism (“it’s become my colleague” feels a little close to the people starting relationships with a bot) but his point about the entire industry being built on the $150/hr assumption rings so true

Software Development in the Time of Strange New Angels Five months ago, my lifelong profession of software development changed completely.

this is a really valuable read

"If you're a developer focused on a single technology, you're about to discover you were never really a developer - you were a highly paid translator between requirements docs and React components"

davegriffith.substack.com/p/software-d...

Photo of a nice bar with lots of indoor plants and an arrow pointing to where we will be with the text “web devs meetup over in this bit”. If you need help finding it let me know!

I’ve booked another Perth Web Devs meetup - casual drinks at Market Grounds in the city on Tuesday Oct 14. Come hang out 😊

perthwebdevs.com/meetups/pert...

In theory, everyone has access to what is close enough to production-grade cameras

But to make a pro movie/video, you also need things like lighting, setup, characters, story... the stuff that professionals learn how to do over years! All that differentiates production-grade

So far, AI is so much more a cultural phenomenon than a technological or economic one.

I helped with the registry for a niche programming language back in the day and now I’m glad it never took off 😅

I think a lot of potential efficiency gains will be tasks where no kickoff is required - where a trigger happens, a prompt is programmatically generated, and the agent gets to work. My guess is the majority of “agents running in parallel” won’t have a developer orchestrating/prompting each of them

For types of work - I think for some of the “jobs to be done” in a developers day where AI can achieve some automation there’ll be some tasks where the agent needs a human kickoff (bespoke prompting) and a human review. And some where it’s one or the other, or neither.

Photo of a conference room with a few hundred people. Behind the stage is a big window overlooking matagarup bridge in Perth

I’m up next at #dddperth, giving a talk “Paths Beyond Senior”. This view out over Matagarup Bridge and the river is beautiful

Kate strides across the stage in a dress that has code snippets printed all over it

Kicking off @dddperth.bsky.social for 2025, wearing a dress printed in code snippets that she has based her talk off 🤯
#DDDPerth

Thoughts on Motivation and My 40-Year Career I’ve never published an essay quite like this. I’ve written about my life before, reams of stuff actually, because that’s how I process what I think, but never for public consumption. I’ve been pus…

I found this post from @charity.wtf honest and refreshing and maybe even inspiring. First time I’ve thought about startup founding in a positive light since the last time I tried/failed 😅

charity.wtf/2025/07/09/t...

A defining characteristic of fascism, that you can find in any textbook on the subject, is picking a powerless minority to scapegoat to unify its base while distracting from the crimes of the government and its allies.

This comes to mind for no particular reason.

Literally, they're in the middle of a heart, bounded by names, it's so cute

And no obviously this weirdo little website isn't the objective arbiter of truth, just an interesting, well, proof that I'm right. Heh.

Also: The space around Janelle Monae makes a little heart, aww. Kind of on brand for them these days.

Yeah given where these packages are mostly used if they stole developer credentials rather than tried to change crypto transactions when running in a browser this would have been so much worse. xeiaso.net/notes/2025/w...

I like this take. I’m not an expert but it feels like the pace of model improvement has slowed. But the applications of the technology (AI coding agents being the one I spend the most time with) still have plenty of room for improved performance even if we never see a smarter model

Two boys playing frisbees

Frisbee time

Is the LLM response wrong, or have you just failed to iterate it? Many "errors" in search-assisted LLMs are not errors at all, but the result of an investigation aborted too soon. Here's how to up your LLM-based verification game by going to round two.

This is a good post on how to get better at using AI. It's the product of pretty intensive research on what consistently works although I don't surface that research directly in it.

Bluesky and AI, so obviously comments on this are off. mikecaulfield.substack.com/p/is-the-llm...

Yuck… I’m sorry you’re getting that. I’ve been subscribed to your blog posts for a while and it’s probably the most thoughtful approach I’ve seen for how to use LLMs usefully in the everyday non-software-development use cases. Thank you

Excited to be preparing this talk for Web Directions' Engineering AI conference in Sydney next month.

I'll be delving into Developer Experience and Codebase Health and how AI tools is part of the picture now