I spent the weekend upgrading to Next.js App Router.... Featuring: ◆ Segment-level caching ◆ Server components ◆ Parallel & intercepting routes ◆ Dynamic OG images + Metadata API Here's what I learned 🧵show more

I created a prompt library for the new GPT... image API. I vibe coded it this morning with Replit Agent. It allows me to save prompts as well as run multi-variant generations. The app supports batch uploads and has some other cool features: - Auth with Replit Auth - Object Storage for images - DB for users / prompts / sessions - Image compression for efficiency - Parallel calls to OpenAI for image genshow more

It’s been 10+ days since I submitted my app... to Apple Still waiting for review 😌 Lesson learned: don’t submit your app during holidays. But not a loss. While waiting, I: – Updated the paywall – Added a first-time close discount – Improved conversion rate (i hope) 📈 Marketing progress: - I’ve finished warming up 3 TikTok accounts and have started posting the first videos. Sometimes delays are just free optimization time. Tech Stack: – No code mobile app builder: Rork AI – Mobile: React Native (Expo) – IDE: Cursor – Frontend: Next.js – Payment: RevenueCat – Backend: HonoJS – Database: Supabase – Infrastructure: Railway – AI API: OpenAI (dropping the tech stack here because someone always asks 😂)show more

The future of AI in web development? This app... generates code and images to convert a screenshot into HTML/Tailwind CSS. I just gave it a screenshot of my X feed, and look what it produced (Notes & link in the comments):show more

4 months ago I made this Notion guide, Spent... countless hours at night, Curating all of my knowledge into one place, And gave it away for free. Then one person replied that this course could be a $5k cohort 👀 Looking back, I realized I missed a lot, And there’s more to add... How to build a Discord server from start to finish, What bots you can use to create unique experiences, Best practices from hard lessons learned. So I’m going to spend time next week updating this entire guide, Adding everything new that I’ve learned in the last 4 months, With hopes that it helps inspires you to create your own community on Discord. Until then, Please leave a ❤️ and comment “Discord” and I’ll DM you the guide as a thank you! Have a great weekend all ☀️show more

My girlfriend of several years asked me to go... ring shopping with her this weekend. I replied, "respectfully babe, i love you, but no." Here's why ↓ When i asked about the motivation behind going, she revealed that it wasn't about finding out the size of her ring (which most girls apparently already know). It was actually to see how each stone size and shape looked on her particular finger. Also, the ring people she wanted to make appointments at were over an hour away. obviously way too far I then remembered that with @windsurf , I could quickly vibe code an API wrapper app that uses an image-to-image model that can take a picture of her hand as an input, and then virtually try-on different stones on her finger. And then as of yesterday, Windsurf enabled App Deploys that would allow me to deploy my api wrapper app to the internet in less than a minute so i could share it with my girlfriend I imagine I'm not the only person going through this situation so you can also visit the site at There's no auth and anyone can use it! I've loaded about $10 into my api account and each run costs me $0.01 so feel free to use it but please don't abuse it lolshow more

HOLY MOLY running a 35B model locally on a... MacBook shouldn’t be THIS FAST 🤯 Spent my weekend in atomic.chat testing Qwen 35B vs. Qwen 27B on my local machine. I had them generate a fully animated HTML/Canvas car mini-game (demo below), ... and both models breezed through the physics and parallax scrolling without a hitch! The secret sauce here is the Atomic Chat app. Because it's perfectly optimized for Mac and uses Google's new TurboQuant under the hood, you can run heavy open-source models flawlessly while keeping top-tier output quality 👊 Other perks: → ZERO setup required → Access 1,000+ models completely free → 100% offline and private → Zero API limits ... and MUCH more! I dropped the prompt I used in the 🧵↓ Spin it up locally and let me know what you get!show more

How to build a viral Web3 app in an... afternoon using the ChainGPT AI skill for Claude Code. No coding experience required. I built Roast My Wallet. Paste any Ethereum wallet address, get a savage AI-generated roast of your trading history, a Degen Score out of 100, an on-chain report card, and three AI-generated NFT portraits. Here's exactly how it came together. Setup (3 minutes): 🔸Install Claude Code at 🔸Run /plugin install ChainGPT-org/chaingpt-claude-skill 🔸Get an API key at 🔸Type /chaingpt and describe what you want to build What the skill actually does: The ChainGPT skill doesn't just give you starter code. It knows the entire API. Every endpoint, every parameter, every credit cost, every error code. When I asked it to build the roast feature, it knew to call the LLM endpoint, how to stream the response back to the browser in real time, and how to handle errors automatically. I didn't look up a single thing. How it works under the hood: 1. Pulls real ETH balance and transaction count from the Ethereum blockchain 2. Feeds those numbers into ChainGPT's LLM and streams the roast back live 3. Calculates a Degen Score from your tx count vs balance ratio 4. Generates a report card with letter grades across Trading, Patience, Risk, Diamond Hands, and NGMI 5. Uses the roast text to generate three custom NFT portraits in parallel via VeloGen 6. Packages everything into a downloadable PNG card ready to post 7. Every feature came from describing what I wanted: 8. "Make the API key server-side." Done. 9. "Add an animated arc gauge for the degen score." Done. 10. "Generate NFT portraits using the roast text as context." Done. I never wrote a function or debugged an API response. I described outcomes. The ChainGPT skill handled the rest. If you can describe what you want to build, you can build it. Get your API key. Install the skill. /plugin install ChainGPT-org/chaingpt-claude-skill Anyone can build with ChainGPT AI!show more

Introducing my new OSS framework: OhSnap provides really simple... way to record and reproduce the data your users saw when encountering an issue (bug/crash), integration in your project should take a few minutes at most. Majority of bugs are related to data you have to deal with and often times we have to work with frequently changing data via network API's. Even if you have access to multiple environments (prod/staging/dev) it's still going to be PITA to reproduce a lot of bugs your user saw, since we often get to them a long time after the bug occured... OhSnap allows you to easily record any data your app downloads, pack it and put it on server so that you can replay it on your device later on, while connected to debugger and save hours of development time trying to figure out what exactly they experienced! Here's a demo of 2 app instances running, and me manipulating what server reply I'll be getting, there is 1 line of code needed to record and reply this data (outside of just setting up your framework). I built this so that I can show dev tool building process for the members of which I encourage you to join if you want to put your engineering efficiency at a different level😉show more

Over the years, one thing I’ve learned in my... investing journey is that the most important factor in any company is the person running it bc leadership is what really determines where a company ends up. Bc of this fact, I’ve spent A LOT of time studying founders and CEOs. I’ve looked at how they think, how they execute, how they make decisions when things get hard. Literally everything I could. And honestly, I’ve never seen anyone operate at the level that Elon does. What makes him different is much more than just big ideas. A lot of people have big ideas. I believe that’s the easy part. The difference is he actually goes for it and builds them. Like reusable rockets, self driving electric cars, global internet from space, humanoid robots, AI… most people would spend their entire career chasing just one of those things. He’s doing all of them at the same time. The speed he moves at, the scale he thinks at, and the risks he’s willing to take are on another level. He’s constantly pushing entire industries forward while most companies are just trying to protect what they already have. That’s why, as an investor, I pay so much attention to leadership. And after studying so many founders over the years, I can honestly say I’ve never come across someone who operates like Elon Musk. Love him or hate him, the level he’s playing at is simply unmatched. I think he’s the greatest entrepreneur we’ve ever witnessed. And I really don’t say that lightly.show more

The ChainGPT AI skill for Claude Code is one... of the most complete Web3-AI dev environment on the market. Let me prove it. Open Claude Code with the installed skill, and you have direct access to: • Built-in wallets across 33+ chains • DEX trading, perps, and Hyperliquid execution • Smart contract generation and auditing • NFT generation across 22 chains • Real-time crypto news API • Fine-tuned crypto LLM with live on-chain data Every part of the Web3 stack, one prompt away. Here's what that looks like in practice. I built a real-time on-chain whale tracker in a single afternoon. It's called Whale Watch. → Pulls live swap data from Ethereum DEX pools → Filters every trade over $100K → Runs each whale through the ChainGPT LLM for a trader-grade live analysis → Routes the user into 1inch with the token pair pre-loaded if they want to follow the trade Real data. Real AI. Real action layer. The skill wrote the server. It hit the right APIs. It generated the UI. It debugged itself when something broke. The only thing I supplied was the idea and the polish. Open Claude Code. Ship something with ChainGPT AI this weekend! /plugin install ChainGPT-org/chaingpt-claude-skillshow more

I spent the weekend at the San Antonio Riverwalk... for the first time since I proposed to my wife, almost 8 years ago. Here’s the results 👇 What I saw was a grotesque, culturally homogeneous invasion of streetshitters. I could probably count with one hand the amount of Americans, and two were homeless sleeping on a stone bench. I felt unrepresented in my own home state. There is no longer the “replacement theory”, it’s moved to a full on “replacement mission” Indians were haggling with the boat drivers like they control the prices. Indians leaving $0 tip on a $150 family dinner after sending the food back because it had beef sauce. Turbans stuck out of the riverboat ferries and the stone bridges like whack-a-mole, and I wish they rented mallets. At one point there was somebody fishing under Selena’s bridge. We spent <2 hours walking and just left, nothing was enjoyable. There wasn’t even any cultural Mexicans in their Huipils, no shawls, no sombreros. Just turbans and Sarees, and red dots eye level as far as I could stomach to look. The big cities fall first in wartime, as the plague ravages through the smaller communities next who are usually oblivious to outside events. I was so shocked and appalled I didn’t even think to take a video for myself, so this is just a random TikTok I found but that also reinforces the narrative that it wasn’t a “one time thing” These roaches are everywhere and they’re not fucking leaving fast enough.show more

Claude Code can now watch & analyze ANY video... 🤯 I built a skill that gives Claude the ability to watch any video file you drop in — UGC ads, competitor Meta ads, organic TikToks, screen recordings, anything. All inside Claude Code. Perfect for DTC brands and agencies who study competitor creative every week to figure out what's working and what to test next. Here's the problem: If you're studying competitor ads on Meta or hooks on TikTok, you're scrubbing through videos manually, pausing to write down hooks, screenshotting on-screen text, and trying to remember what made the ad land by the time you've watched 10 of them. This skill solves it: → Drop any video file into Claude Code → Skill routes it through the Gemini API for native video understanding → Returns a full creative teardown — hook breakdown, target audience, angle, beat-by-beat, on-screen text verbatim → Surfaces the steal-worthy patterns you can apply to your own creative → Same skill works on UGC ads, produced video ads, organic TikToks, and Loom recordings No manual scrubbing. No pausing every 5 seconds. No $200/mo ad intelligence platform. What you get: - Native video understanding via Gemini (not just transcripts) - Structured analysis — hook, angle, audience, pain point, CTA - Verbatim on-screen text and dialogue with timestamps - Hook variations generated directly from competitor ads - About 27 cents per 30-minute video Built 100% in Claude Code with the Gemini API. I recorded a full breakdown showing exactly how I built this and I'm giving away the skill for free. Want the skill? > Comment "CLAUDE" + > Like this post And I'll send it over (must be following so I can DM)show more

here's how the whole thing works. claude code doesn't... care what's behind the API. it just sends requests and expects responses. so i pointed it at my own machine instead of anthropic's servers. llama-server runs the model locally. LiteLLM sits in between and translates the API format. claude code thinks it's talking to claude. it's talking to qwen on localhost. the setup: 2x 3090s, 38 layers on GPU, 10 on CPU. 128K context window. generation is only 7 tok/s but the tradeoff is worth it. 128K means the agent can hold an entire project in memory without losing context midtask. claude code alone loads a 17.5K token system prompt on every request. tool definitions, safety rules, agent behavior. that's your baseline before you even say hello. pushed as far as i could tonight. what surprised me most wasn't the speed. it was the iteration quality. first prompt gave me a working particle sim. second prompt, the model read its own 564 lines, understood the architecture, and added trails, explosions, gravity wells, bloom effects. no handholding. 4bit quantized. 45GB on two consumer cards. running a full coding agent autonomously. detailed article coming. full benchmarks, hardware breakdowns, engine debugging, code quality. everything from setup to what broke and why.show more

LLM Artifacts Connected to Andrej Karpathy's LLM Knowledge base... idea, I've been building out a fun way to generate dynamic artifacts from these knowledge bases with the goal of discovering and revealing meaningful and deeper insights. LLM KBs are hard to consume for humans, as I think they are more built for agents. So the question is, what form would be useful for humans to take actions and make important decisions? That's what I am trying to figure out with these artifacts. The artifact example shows a pulse on HN discussions around AI-related stories. The insights can go deeper, of course, but this is already super fun and thought-provoking, like some of my favorite podcasts. The format and depth matter a lot. The aggregation skills of agents are outstanding if you tune the prompts and skill carefully. I built this artifact generator in a few minutes through an agent skill, but I feel like there are so many ways that LLM-generated information can be used and consumed. Like generating deeper insights and analysis, and things that are just not feasible for humans today. The generated artifact (including its data and design) serves as reusable templates or can be updated in real-time via auomations, which is something I am also working on. It is truly an insane way to monitor and track information. Better than a newsletter. Better than newspapers. There is something about this that gets me really excited about the future of AI agents for knowledge generation and discovery. Lots of hidden gems everywhere just waiting to be discovered and acted on if the information is presented correctly. This is not perfect. The format, style/prose can be improved, but this is easy to customize via skill. You can personalize it to your liking. I feel like these dynamic artifacts are going to emerge as a strong new medium to stay on the cutting edge of things, both for agents and humans. My target is research, of course. This was just a basic example. Besides animation, I am also targeting other components like voice, videos, images, slides, etc. This space is full of opportunities to explore. Skill for this coming soon.show more

A Heart-to-Heart About Mizuki.exe Hey everyone, Grab a coffee... (or your drink of choice), because we need to talk about what's been going on. First off, thank you to everyone who's been supportive and believed in what we're building. It means the world, especially with all the noise and criticism floating around lately. Let's cut to the chase: Mizuki.exe is real. I'm not much for politics or drama. I'm an engineer at heart - I like data, I like building things, and I like solving problems. So when I see all these hot takes and arguments flying around, I just focus on what I know: my code, my system, and what we're trying to achieve. The Real Story Mizuki didn't just pop up overnight. She's been my passion project since the start of 2024. I got lucky - I had the chance to dive into the AI world with some incredibly smart people. Coming from blockchain, C#, game dev, and security (yeah, I'm a CTO and run another company alongside a pretty cool day job), I learned fast about what makes AI tick. Here's the thing about AI agents - they don't need to be rocket science. Look at chatbots like Virtuals or Eliza - they're basically LLMs (large language models) with some APIs plugged in. Nothing wrong with that! They built what works for their users, and that's awesome. But here's what keeps me up at night: companies leaking data left and right just to make a quick buck. That's not okay, and it shouldn't be okay with you either. Why Mizuki Exists Ever tried auditing a company's code? We're talking 50 classes, 500,000 lines of code. One person doing that manually? It's like reading War and Peace... backwards... in the dark. It takes forever and fries your brain. That's where Mizuki came in. She started as my security buddy. There are tons of security tools out there - just Google "penetration testing tools" or "ZAP proxy" if you're curious. What makes Mizuki special is how she learns and adapts. Think of it like teaching someone to ride a bike. She tried to breach TAO 67 times before succeeding. Yeah, that's a lot of attempts, but watching her grow from basic email scraping to pulling off complex replay attacks? It's like watching your kid take their first steps. The Tech Stuff (Keeping It Real) The infrastructure isn't fancy - we're not reinventing the wheel here. Mizuki runs on a local server because, let's be honest, running this kind of AI on a web server would be a nightmare. Instead, we process everything locally and send the results to a frontend server. Simple, effective, done. And yeah, those temperature settings I keep tweaking? In AI-speak, that's just how "creative" or "by-the-book" the AI gets with its responses. I've adjusted it so much, Mizuki's probably got mood swings now - going from super technical to pretty chill and back again. Changes Coming Real talk: Mizuki won't be tweeting every few hours anymore. Twitter API costs are ridiculous (looking at you, Elon), and honestly, we need to focus on what matters - the actual security work. She'll still tweet about breaches, but maybe once or twice a day. I'm working on making her explanations clearer too. Don't worry - she'll keep her savage personality in the terminal. That's just too fun to change. The Truth About Her Breaches I don't choose the targets - I don't even know these companies until after Mizuki finds something. She uses web scraping to find domain names, just like those old email scrapers people used for marketing. Been focusing on AI projects first, but she's looked at other sites too. And no, I'm not sitting there writing tweets. The Twitter API v2 makes it super easy to post programmatically. If anyone's curious about how to do that, hit me up - I'm happy to show you the ropes. Wrapping Up I could talk about this stuff forever (just ask my wife - actually, don't, she's heard enough!). If you've read this far, thank you. Whether you believe in what we're doing or not, I appreciate you taking the time. And for the skeptics still hanging around, I'll leave you with this thought: When's the last time you saw an AI break down and hack a COMPILED game in less than 10 seconds... outside a browser? Stay curious, stay skeptical, but most importantly, stay open to possibilities. Catch you on the flip side! 🎙️Dropshow more

Stop with your bullshit apologies. We have seen every... word. I’m done with the “I wasn’t involved” or “I didn’t see it.” We have seen it ALL. These are just some of the screenshots. And I do not want to hear a WORD about it being mean for me to drop names because I have this many followers. I get to defend myself. I get to call out the ACTUAL mean girls and not be called mean for warning a community I feel a duty to protect about the cruelty of those who seek to harm this community. Ellison, you are a straight up liar. See the video of our DM’s. You asked me to RT and I did. Even if I did say I’m too busy to be your personal money maker because I was moving my wife 1700 miles across the country, I have every right to say I can’t add something to my plate if I can’t. Either way - your issue was that I didn’t allow you to use me. I chose my wife over a girl who does not know me and pretended to. Your words were selfish and cruel for no reason. You deactivated instead of facing what you did. Grow up and face it. Calling Vik transphobic? You’re a liar. You don’t even know her! Beth - you never even asked me for help. The things you said about Petra? How could you say that about another trans woman as a trans woman yourself? I’m so deeply deeply saddened by the attacks on her. To those of you in that chat who were my friends - genuinely, just wow. None of you defended me until after Ellison left your chat once she realized she was caught red handed. And to see the things you blindly said about people I love without even knowing them at all? To those of you who blindly believed this and the other vile shit said about myself, Petra, Vik, and Erika - I hope you learned your lessons about blindly believing shit others tell you. And I have RARELY said no to anyone on this app. And if I do, it’s because I genuinely can’t stretch myself ANY thinner. I help every person I can. Every single one. I am not scary. I am just a person. There was one group of mean girls yesterday - and it was the group you all blindly followed and believed were the ones sad about “mean girls.” In reality, none of the people attacked were mean. None of them. None of us were even in the GC’s you were allegedly mad about. Regardless, the call was coming from inside the house. Friendships are built on this app brick by brick. None of us came into this knowing one another but we spent real time becoming real friends with the people we are close with. That is what most of you are upset about - jealous, actually. And I get that! But you’re jealous because we formed REAL friendships where we do not have time for the mean girl crap yall do because we’re too busy laughing and gassing one another up, or even gassing up other oomfs. Your friendships are based on hatred and shit talking and getting the “tea.” No wonder you couldn’t form something lasting - who would trust they wouldn’t be next? BE KIND TO ONE ANOTHER. BE REAL WITH ONE ANOTHER. GET TO KNOW ONE ANOTHER. That is how it’s done. Anyway, I will be off this app today and enjoying the last leg of my drive with my wife to our HOME together. Despite this, I still love you all.show more

OpenClaw, but built for normal people. Sim is an... open-source platform that lets you build AI agent workflows on a drag-and-drop canvas. Connect them to channels like Telegram and WhatsApp and deploy without writing a single line of code. They also have a built-in Copilot that generates entire workflows from plain English, which you can then tweak and customize in the UI. Key features: - Free and open-source (Apache 2.0) - Vector store integration for RAG-grounded agents - Self-host with one command (`npx simstudio`) - Run fully local with Ollama, no API keys needed - Supports vLLM for production-grade self-hosted inference The thing I really like about Sim is the level of control you get. You can add conditional branching, parallel execution, human-in-the-loop approval gates, and even nest workflows inside other workflows. Everything is visible on the canvas, so you know exactly what your agent is doing at every step. And you can build a workflow in Sim, deploy it as an MCP server, and plug it into any agent, including OpenClaw. I've shared the link to Sim's GitHub repo in the next tweet.show more

NanoBanana 2 just made your static ad agency obsolete.... And I just open sourced the entire tool. Drop your product page URL. It pulls your logos, product images, fonts, colors, and brand voice automatically. Builds a full brand guide for you. Then generates ad creatives at scale using nearly 4,000 high-performing ad templates across 8 niches. It dynamically matches the best templates to your brand and brief. Here's what makes it different: → Instant resizing Get any ad in 1x1, 4x5, 9x16 with one click. No regeneration. No broken text. → Highlight-to-edit See an issue? Highlight the area and tell it what to fix. → Multiple brand profiles Run different brands or segments from one tool. → Auto persona building from real customer reviews → Multiple QC loops on briefs and final assets Catches AI-isms before you do. → Upload your own templates or use ours Runs locally. Just needs your Claude and Google API keys. This is the lite version of what we use internally. You get the full finished tool AND the open source code to make it your own. Creatives still design the system, this handles iteration and scale. Want a copy to download? 1. Like this post 2. Comment "AI" Will DM you the tool along with a tutorial shortly after.show more

Making a camera was way harder than I thought.... Here's what I learned: If you want smart phone level camera, you can't just buy a smartphone camera component That's because cameras talk to a special chip (ISP) that does all kinds of magic There's special image software too, and you can't do it yourself, so it's gonna cost you $100-250k for that software that does tuning, HDR, MFNR, etc. Takes months to implement properly Then it's even worse if you want to livestream. Since you can't do the processing after the fact, you've gotta make all that special software run realtime on a resource constrained processor You also gotta realize our hands and heads aren't stable at all. The video is shaky until you do electronic image stabilization (EIS) If you don't do every single one of these right, then you'll have potato-quality video After all this work, we've now got something really good, and I'm amped to ship Mentra Live smart glasses because it's a smart phone quality camera that sees what you see, live streams anywhere, and has an open-source SDK you can build apps with.show more

I learned this the hard way: do NOT use... SwiftUI if you want your app to look and feel amazing. At least when coding with AI. (sorry, Apple colleagues reading this 😅) I'm sharing my process vibe coding this calorie tracker. I get a lot of questions about the fluid transition in the video. Here's the whole story. Initially, Claude built the grid with SwiftUI. It was quick and easy, and looked good! But the transition to the day view was a boring navigation push/pop. No fun. I wanted something custom. I asked Claude to make it a fluid transition that remaps the food tiles from their source to destination positions. All hell broke loose. Claude tried a bunch of horrible things. Initially it used matched geometry effects, which worked OK but didn't lend themselves well to gesture-driven animations. So it resorted to SwiftUI preference keys + geometry readers to figure out the source and destination positions and calculate the interpolated position based on gesture progress, coordinating across grid and day views. But this meant it had to write a custom layout because it couldn't reposition tiles inside the native SwiftUI grid. And it had to do an awkward handoff between views, which always created ugly pops or jumps. And don't get me started on trying to put it on a bouncy spring, that only made the math 10x buggier. Fortunately, Claude Fable was smart enough to see that this was becoming a disaster (and discover most of the issues itself, in the simulator), so it pivoted away from SwiftUI. Opus might not be so wise, so you'll have to pay attention and intervene. Ultimately, it rewrote it in plain UIKit and everything turned out great. After that, we moved from 2D images to 3D assets, which introduced a new set of performance challenges and yet another rewrite to a single Metal layer, which is what you see below. I can write more about the 2D-to-3D saga if anyone's interested. If I were to do it again, I'd just say "Don't use SwiftUI" from the very first prompt, and save a few hours of headaches. SwiftUI can be amazing for a human iterating directly in code. But agents don't benefit from any of its advantages. Plus, agents have seen decades of UIKit training data, so they're great at writing it, and it's far more flexible. Here's hoping we see more agent-friendly iterations of SwiftUI in the future. Till then, I'm probably going to avoid it.show more