Tilly The Coder

@Matildevoldsen

Building www.zxn.studio Laravel | YouTuber | Glassy Filament wirepackages.dev | pinkary.com/@TillyTheCoder | Podcast with @techenby

Commit activity

134 commits in the last 30 days

longest streak 21 dayscurrent 21

3,673 XPMaximum rank (UID 0)

Pinned

feat: NativeWatch, error tracking for NativePHP apps Later this week I'll be shipping: - native-desktop and native-mobile SDKs: install with composer, crashes get reported - groups errors into issues by fingerprint, reopens them when they come back - native iOS/Android crashes…
NativeWatchNativeWatch: error tracking for NativePHPCatch PHP, JavaScript and native iOS and Android crashes in your NativePHP app, see who they hit, and fix them with AI.

Yesterday

I won’t be applying to speak at conferences or events next year as I’m currently waiting for a date for my gender-affirming surgery. I don’t want to commit to events when there’s a possibility the dates could overlap. I’ll also likely be a little quieter for a few months at…
Do you let your clanker file a bunch of PRs, or keep committing to main/dev branch? I think this may vary by company/personal project but curious to hear what folks do.
At the beginning I've always been direct to main, once it gets to a MVP state then I usually do PRs. With and without clankers.
If I'm solo, direct to main. No need to add more complexity. As Grug said, complexity is bad. Really, really bad.
Have you used AI to write most of or an entire tutorial or article?
Poll options

2d left

I usually draft my original idea, let AI scaffold it, review it, rewrite or expand some paragraphs, review with AI, run editorial checks I built, read it again, decide if it's good... and sometimes bin it 😅
I have to admit that sometimes I even have most of my README.md's written by AI for me — even though I actually quite enjoy technical writing. But I’m also in a software factory mode right now.

Wed, Nov 11

Tip: You have a platform where analysing video's is important, consider using Gemini 4/3.8 flash as it's so much cheaper to watch and analyse your video than other option's for the quality. Been testing a lot of models recently
Using fable 5.2 for some work, but really trying my best to use all of my tokens on my 5x plan. So essentially hoping to have all of it spent today and use a reset
Where I’m staying in London internet is really slow. So spending £10 on a e-sim for a month so I can have extra 80gb of data. Also why I’ve not shipped as much
Currently ZXN Studio only supports Codex for editing / motion graphics and that, but soon, that will change. Plan is to: - add Claude Code support through agent SDK (from what I've seen if I do it similar to t3code or other's it's fine) - AI credits to any monthly sub including…

Tue, Nov 10

Mon, Nov 9

Do you guys want me to post relevant news from AI world here daily? With sources and that ofc
  • Yes67%, 8 votes
  • No33%, 4 votes

12 votesFinal results

When they introduce a new AI model, do the older models get cheaper like the way Apple products do when new versions are released?

Sun, Nov 8

Sat, Nov 7

Fri, Nov 6

While this is not SuperNative desktop, and I do not have access to the codebase, I'd expect it to be similar to Swift in here, maybe slightly worse/better. This is myself for fun porting ZXN Studio to Swift, also to hopefully increase speed of porting to SuperNative desktop. So…
Up early today. Moved to t3 code, and honestly this time it feels right. Beforehand t3 code didn’t feel right UI/UX wise or was too early on and at that point I preferred to use Claude code and codex. Going to get some SEO articles out
I'm a Pi user myself. I've tried T3code previously and honestly didn't like it. But it was some time ago. I think I will give it a go once more. Both to T3 and to opencode. See if it clicks now