at this point i’ve airdropped thousands of dollars to design everydays holders in the past year alone. a few thousand in zora creator coin thing. and a few thousand months back in comic sans coin. and the real market value of the NFTs hasn’t even been recognized yet.

me: i vibe coded an app for you to use you: 🦗🦗🦗 me: i vibe coded a way to send you money you:

i vibe coded a thing. now, every time i claim my vested creator coin from zora, it gets distributed to everyone that holds design everydays NFTs. the more you hold, the more you earn. the earlier the NFT you hold, the more you earn. first distribution just went out.

the question i get most is "how do i make better designs with AI?" the truth is, you're going to get UI slop no matter what. there's no magic bullet other than designing the old fashioned way (for now) and using the Figma MCP for dev specs. but here's what you can do that you couldn't do before: 1. schedule 5-7 calls with users 2. use granola to record and take notes 3. paste transcripts into cursor 4. prompt the AI to synthesize the conversations and extract insights for design improvements and generate personas and just like that, you have instant user research and design artifacts that the agent can reference every time you brainstorm or build a new feature. i know that's not the design you were asking for, but it's the design you need.

i've been using gpt-5 in cursor for almost a week. the good: > one-shotted a problem a problem i'd been debugging for a week in claude code > really fast for a thinking model > always calls the right mcp, and uses them generously the bad: > absolute yapper. i ain't reading all that, bro.

vibe coding viral games with @songadaymann

we’re really going to get GTA 6 before we get a better siri aren’t we

notion sucks. it’s always been so confusing to me. but notion MCP is so underrated. why? first, it’s one of the easiest MCPs to set up. you don’t have to write a single line of code and is usually one click connection on most AI chat apps. second, a notion hub is a single artifact that any model or ai chat app can work on. it’s a single source of context and memory for claude, chatgpt, any IDE or CLI. populate it with to-do’s, notes, work documents, events, anything and you can ask any model to organize it, read it, or make changes to it. it’s the best way i’ve found to no-code an actionable assistant. a few months ago i posted about using chatgpt projects to create a fitness coach. here’s how i’m using notion mcp to do the same now: o3-pro is the smartest planner. in chatgpt, i asked o3-pro to cycle a new training program tailored for me. then i instructed it to create a notion doc for the program, schedule and meal plan. claude code is a great systems architect. i asked claude opus in my CLI to read the notion and architect a hub for the program, where ill be logging workouts and a meal diary. it created a front page calendar with sub pages for workout instructions and a place to log sets and reps. i asked it to create a system prompt with instructions on how to read and populate the notion. then asked it to place it somewhere in the notion that any agent can read before making changes. now i can open chatgpt or claude on my phone when i’m at the gym and ask “what’s my workout today?” and it guides me through each exercise, logs my performance, and calculates progressions for my next workout. every chat is a complete throw away. the only thing that matters is what’s in the notion. you might not find the fitness use case to be as valuable for you, but maybe it sparks another idea on things you can track cross-model. and the third reason the notion MCP is so valuable is because it’s a database that’s very easy to check for mistakes or manually make changes. i don’t have to completely rely on the model. — the notion MCP has completely reinforced what a “product” and “interface” means with AI. it’s required my brain to see them in completely new ways. interoperability was always the most exciting aspect of blockchain for me. that you could port your data to any app, and apps are just different access points. the same applies here with chat apps and models. i can create a private database for myself and access it with any of the new models. i can start in chatgpt, finish in claude, and even fine tune my own model to do the same from my device. the app doesn’t matter. the model doesn’t matter. but the context is always there.

visual design is wildly overrated interaction design is wildly underrated

a lifelong friend has built an impressive ai visibility service to help startups get more traffic from AI mentions. if you want to increase the amount of times chatgpt mentions your company or product, lmk and i’ll introduce you.