New Blog; Who Dis?

I appreciate that the post before this one says it’s the first one, even though there’s one “before” that. Both of those were placeholders generated by Claude Code while it was designing this theme. It came up with the Ink Console name too; it’s as good a name as any.

Enough of that, though. This post is written by me, and all future posts will also be written by me, unless otherwise specified. They’re going to be about AI, because what the hell else am I going to be writing about in 2026? AI, LLMs, agents, harnesses, runtimes, tools, frameworks, maybe hardware, and projects I’m working on with all of that. Ooh, that’s a good idea.

Projects I’m working on

New workshop

So the theme for this blog was actually extracted from a “design language” that Claude and the Anthropic frontend-design skill came up with for the website for the Claude Code workshop that I’ve been working on with my longtime friend and collaborator Dylan Beattie. Check it out at claude.rendle.dev

We’re going to be running public workshops online that anyone can join, usually using a two-half-days format, sometimes on weekends. And for businesses, we can run a single day either online or at a location of your choosing, for up to 20 attendees, with the content customised for your teams and your projects.

I personally believe a live workshop is a better fit for teaching this material, purely because it’s all moving and changing so fast that a pre-recorded video course might be out of date a week after we finish editing it. This way, everything stays right up-to-date, even if we have to update the demos or the slides the night before the workshop.

We’ve run an in-person one for a customer in Sweden and it went really well. The attendees were a mix of developers and engineering-adjacent people, and by the end of the day everyone completed their final projects - I won’t spoil what that is, but it was a lot of fun.

Dmon

I’ve been working on my own agent harness and routing system, built with .NET 10 and Microsoft.Extensions.AI. There’s an ulterior motive that I’m not ready to share just yet, but I can tell you it’s heavily focused on privacy and local models like Gemma 4 that you can run on a Mac Mini with 48GB of unified memory. The stuff I’m learning while working on that is the reason I wanted to start this blog.

I think that the end-game of all this is more specialised small models running on commodity hardware, like Apple Silicon or Nvidia’s new Arm-based CPU/GPU/thing. Maybe not this year or next, but the performance and the memory available, once the chip shortages are resolved, are going to keep going up, and the models are going to keep getting more efficient. The Mac Studio with the M3 Ultra chip and 512GB of unified memory was “only” £10k; rumour has it the M5 Ultra version will go to 768GB. At some point that capacity, combined with falling prices, will make it possible for everyone to run frontier-level models on a relatively low-power system that sits in their house like a HomePod or an Echo does today.

You can give these local models access to your emails, your calendar, your messages, all your documents and other data without it ever leaving your house. Nothing uploaded to The Cloud, no worries about what Zuckerberg or Altman or whoever might do with it. Your model, your data, your hardware. That’s what I’m trying to build. Obviously I’m hoping to make a bit of money off it somehow, but the core of it is open source and will stay that way.

(If that sounds like something you might want to help with, or even invest in, please hit me up on LinkedIn).

Incredibly personalised apps

I got back on stage to do stand-up comedy again recently, after a 25 year absence from that scene. Back in the day, we used to write new material on our arms or on scraps of paper. This time I made an iPhone app I called BitTimer, where you could paste a Markdown bullet list of “bits” and it would step through them one by one, with a timer on the top half of the screen. It’s a native iPhone app, SwiftUI, Xcode, all that stuff. It took maybe half an hour of my time, and two or three hours of elapsed time.

Now I’m building ADHD Buddy, for iPhone and Apple Watch, to remind me to do things like take medications or do breathing exercises or drink water or whatever. It’s a bit more complicated but it’s still just scratching an itch. I might publish it. We’ll see.

Games

I’ve always wanted to make games, and I’ve had a few ideas bouncing around my head for years. Turns out the combination of a Claude Max subscription, an excess of idle time waiting for all those serious things to finish, and the modern browser as a game engine makes it really easy to prototype those ideas. It also turns out that most of them are maybe not such great ideas (yes, zero-G pool, I’m talking about you). But there’s a couple that I’m pretty happy with. Yesterday I made Terra, which you might like if you’re a geography nerd.

I am available for children’s parties work.

I am currently looking for a new job where I can put all this stuff I’ve been learning to good use. I’m open to full-time or contract, or one-off projects I can build for you that your actual staff never seem to have time for. Again, please reach out on LinkedIn until I add a Contact form to this site.

Oh, and if you are having a children’s party, I can juggle.

I think that went okay. Honestly, after a year of working with LLMs I’m kinda glad I can still write. Okay, it’s gone 1am, I should probably sleep now. See you soon.