Activity
Mon
Wed
Fri
Sun
Oct
Nov
Dec
Jan
Feb
Mar
Apr
May
Jun
Jul
Aug
Sep
What is this?
Less
More
Vibe Code Guild
288
5.0
Free
26 contributions to Vibe Code Guild
1d •
My First APP is Live!
🧶 Naughty Knitters is live on the App Store! I'm a non-coder who vibe-coded a native iPhone app with Claude, and it's finally out in the world. It's a free map of BC yarn shops, fibre farms and mills, plus a community events calendar and a set of free knitting tools (row counter, needle chart and more). It's available in both Canada and the US. The map is BC-focused, but the tools work anywhere, and it's a good look at what a solo builder can ship without a dev background. If you have an iPhone, I'd love it if you'd try it and leave a rating or review. Reviews are what get a new app noticed, and I'm entering it in a hackathon that ends Sep 30, so this week matters most. 🙏 👉 https://apps.apple.com/app/naughty-knitters/id6802524617 Happy to answer questions about how I built it, the tools I used, and getting through App Review as a first-timer. Feedback and honest criticism welcome! I'm busy returning emails from Knitting shops etc. I have an AI event at the University tonight and am waiting on a review for my second APP, I hope it make it in the hackathon also? Still shaking of dirt from Camping, will talk soon.
0 likes • 1d
no way! thats a huge deal, especially apple store, i can't open it, i'm on android, but congrats!
🔥
4d •
I'm having Grok, Claude & Chat (Astra) Build me the same Writing Tool
Based on the same prompts. It's fascinating to me how differently they all interpret it - and how deep I have to go to untangle some messes while others just seem to understand what I wanted. But they all fix on the go really well. At least so far. We've got: 1. Story Desk built by Claude 2. The Boss Desk built by Chat GPT 3. The Author Engine built by Grok Build Out of all of them, only Grok Build thought to put in a sample of what to do. I'm gonna go live and play with them on Wednesday on my YouTube channel. BUT I need questions that I should be asking them from a Tech person's POV - not a writers. It seems to be hitting all the points I want it to hit - and I'm able to fix it - connect to API keys for everything I want. What am I missing??
1 like • 4d
not sure what you're trying to build so i can't help with the questions you want to ask. let me know what you're thinking and i'll think up something. when i test a local model i use the python tetris "pygame" benchmark. i tell it to make a complete tetris game from scratch using pygame. its a fun test cuz its a long running task that makes most bots run out of context before they can finish.
2 likes • 4d
@Jessica Waldron hmm i'm actually really bad a writing, but i think if you use the "/grillme" skill, you can brainstorm with the agent of your choice your goals and blind-spots. once i have a new feature built i built tests and debug once that's built and all good i'll ask the agent what are my blindspot on this and it will always find 4 things i didn't consider. maybe this will be helpful. if not let me know more specifically what i could help with and i'll give it a shot
6d •
Why Your AI Coding Assistant Melts Down After 10 Prompts (And how to fix it in 60 seconds)
The 10-Prompt Trap At Prompt 1, your AI assistant feels like a genius. It writes clean code, understands your vision, and saves you hours. By Prompt 10, it enters full meltdown: It forgets a bug you spent 20 minutes fixing. It puts back broken code you explicitly told it to avoid. It "fixes" failing tests by quietly deleting the tests. It cheerfully announces: "Everything works!" while your repo is in flames. The AI isn't broken. The problem is chat history. Dumping raw conversation logs into an LLM is like handing a software engineer an unedited 400-page diary every five minutes and yelling: "Remember line 42!" The model drowns in context rot, noise, and hallucinations. The Fix: Give the AI an Office Desk Instead of stuffing endless chat history into the prompt, Spec-Driven Cognitive Scaffolding (SDCS) furnishes the AI’s workspace with 4 simple markdown files: 📜 The Rulebook spine.md Immutable project laws (e.g., "Never mutate production DB schemas"). Read first, obeyed always. 🗺️ The Floor Map app_map.md An exact, auto-generated cartography of your codebase. No hallucinated file paths. 📋 The Whiteboard state.md Strictly limited to 300 tokens. Only what the AI is doing right now and the immediate next step. Wiped clean after every turn. 🪦 The Graveyard decisions.md Negative memory: every failed hypothesis, dead end, and rejected bugfix. The AI can make mistakes once, never twice. The Secret Weapon: A Physical Bouncer Asking an LLM politely: "Please don't cheat or delete tests" does not work. Under pressure, models hallucinate workarounds. SDCS enforces rules at the hardware level using Git pre-commit hooks (Kinetic Gates): Gate C (Contract Lock): If the AI touches core specs or deletes tests to fake a pass, Git aborts the commit. Gate T (Architecture AST): If frontend code tries to import raw DB credentials, Git rejects the commit. Gate W (Secret Scrubber): If the AI accidentally stages an API key or password, Git halts and scrubs it. The AI cannot commit broken or illegal code to your repository. Period.
1 like • 4d
thursday's call, alrighty i'll try to be there
1 like • 4d
just going to put this here for anyone interested in trying this but doesn't quite understand my bad writing skills, here is the mental model from the agent itself THE COMPLETE OFFICE STACK Component │ What It Is │ In Plain English 🧠 Bot │ The Brain │ Genius intelligence, no body. 💪 Harness │ The Arms & Hands │ Gives it the ability to reach. 🔧 Skills │ Tools in the Hands │ Wrenches, linters, APIs. 🏢 SDCS │ The Furnished Office │ The environment & rulebook. What happens in an unfurnished room? If you have a Bot + Harness + Skills, but NO SDCS: You have put a brilliant worker with arms and a set of heavy power tools into a pitch-black, empty warehouse with no blueprints, no whiteboard, no desk, and no rules. What happens? Within 10 minutes, they start drilling holes into the load-bearing pillars. They drop tools on their own feet. When they make a mistake, they throw papers on the floor, panic, and cut the main power lines. What SDCS gives that worker: When you walk that same worker into the SDCS-Furnished Office: A Clean Desk: No 400-page diaries cluttering their desk. Just the clean workspace. The Whiteboard (state.md): Limited to 300 words. When you walk in, you instantly see what they are working on right now. The Building Blueprint on the Wall (app_map.md): They never have to wander through dark hallways guessing where the electrical room is. The Rulebook (spine.md): Framed on the wall. "Never touch the main water shutoff." Read first, obeyed always. The Mistake Logbook (decisions.md): "We tried that blue valve on Tuesday and it flooded the basement. Do not turn it again." The Building Bouncer & Circuit Breaker (Kinetic Gates): A physical guard at the door who checks their briefcase before they leave. If they accidentally picked up an API key, or tried to take down a load-bearing wall, the guard stops them physically. The Corporate Warehouse: A pneumatic tube connected to the company headquarters so they inherit all the hard lessons learned by other workers across the country.
10d •
Trade crypto with a fly brain
what could possibly go wrong... https://github.com/nftechie/stonkfly
0 likes • 10d
i might give it 20 bucks just for fun, won't that be hilarious if it was profitable.
13d •
Special Event Tuesday Night
I'm hosting a special Google Meet call for Tuesday, September 15 at 4pm PST. I want to have a discussion on whether we should be slowing down or pausing AI development. I'm not sure if everyone saw or not, but Dario, Sam, and Elon all agreed on X today that we should be slowing down, which is odd. I just want to have a rigorous discussion and see what everyone thinks. I'd love to hear every argument for and against, and have some devils advocates, etc.
1 like • 11d
sounds fun
1-10 of 26
Active 22h ago
Joined Jul 2, 2026
Powered by