Opus 4.6 vs Gemini 3 pro (1/9) prompt: Build... a single-file HTML/CSS/JS (no libs) demo that uses SVG to simulate a plant growing: stem extends, leaves sprout + unfurl with springy/windy “physics”, then seamlessly loops forever. my dear opus...what have you done...show more

Opus 4.6 vs GPT 5.4 (High) (1/9) prompt: Build... a single-file HTML/CSS/JS (no libs) demo that uses SVG to simulate a plant growing: stem extends, leaves sprout + unfurl with springy/windy “physics”, then seamlessly loops forever. For the initial impressions I'm really impressed by GPT, for speed they both felt about the same, but gpt was still half cheaper than opus. I also much prefer the design and animation that gpt produced, physics on the leaves are super cool and it also loops pretty nicely whilst opus just fades out the plant. Still got a bunch of tests to run but this is really exciting.show more

Gemini 3 Pro (left) vs ChatGPT 5.1 thinking vs... Claude Opus 4.5 (right) Prompt: Multi-ball heptagon physics Write html, css, and javascript to display 20 balls bouncing inside a spinning heptagon. each ball should have a unique color, a number label, gravity, and collision detection with the walls. all code in one file.show more

GPT 5.2 vs. Gemini 3 Pro Kimi 2.5 vs.... Opus 4.6 Prompt I used: "Build a first-person downhill skiing simulation in a single HTML file. The 'skier' must automatically accelerate to 93 mph within 10 seconds. As the speed increases, the edges of the screen must 'blur' and the camera must vibrate realistically to simulate the extreme speed of an Olympic run."show more

Gemini 3.1 Pro vs Opus 4.6 opus is consistent... and relaible Prompt : Build a premium interactive isometric 3D cozy room using Three.js in one html code block, with all objects modeled in code (no external assets) and subtle ambient animations and you can do 2 things of your choice to make it bettershow more

1-bit Kimi K3 performs at Opus 5 level on... 3D physics! We ran our Atomic Chat quant of Kimi K3 locally on 4x B200 against three cloud models and gave them all the same task, to build a giant anvil drop test as a single HTML file with real physics Outputs: K3 1bit (local): 15.8K tokens, $0 API cost Kimi K3 (API): 15.3K tokens, $0.30 API cost Opus 5: 22.8K tokens, $0.77 API cost GPT 5.6: 14.5K tokens, $0.72 API cost All four got the physics right. But only Kimi made a working winch. The drum turns and the chain drags the flat car off the pad. Opus 5 drew the most detail, road markings and sparks on the hit. And you can run a model at this level on your own box now. That still feels insane to usshow more

look what a single consumer GPU just built. gave... Qwen3.5-35B-A3B one prompt: build a cloud GPU marketplace with pricing cards, deploy templates, and a benchmark leaderboard. it planned the layout, wrote the animations, populated the data, and served it. one shot. one HTML file. then i told it to iterate. split the hero, add a floating GPU with neural network animation. glassmorphism on the cards. done. done. done. three rounds, no confusion, no regressions. 4-bit quantized. 19.7 GB. single RTX 3090. full coding agent claude code harness running on localhost. no API calls leaving my machine. no subscription. no rate limits. earlier today i pointed it at my own production website. it curled the HTML, found every broken link, and told me "pretty shell, empty core. would not recommend." then built a better version from scratch. local inference stops being a demo when you actually steer it. the models are there. they understand intent. but you have to meet them halfway with good prompts, clear context, and real project structure. that's the skill gap now. not the models. the steering. more experiments coming. i genuinely cannot stop playing with this thing.show more

this is f**king dangerous someone figured out how to... make Opus 4.8 run on Fable 5's brain with one prompt access to the best model is never guaranteed. It disappeared once already this year. but you can use it forever. here's how: 1. ask Fable 5: "write the operating manual your replacement will run on" (procedures, failure modes, a 5-question self-test) 2. save the output as one .md file and drop it into a new Claude Project as the project instructions 3. switch to Opus 4.8 and now your everyday model runs off the smart one's method, no top-tier price save and bookmark this no matter what full extraction prompt is in the article below: ↓show more

Hold up, here is the prompt: works with almost... any model. enjoy :) Role & Objective: Act as an Elite UI/UX Front-End Engineer specializing in Apple-tier micro-interactions and advanced CSS. Your task is to program a perfectly centered navigation bar in a strictly SINGLE HTML file containing all HTML, vanilla CSS, and vanilla JavaScript. No external libraries or frameworks (No Tailwind, React, etc.). Design Concept - "True Liquid Glass": CRITICAL INSTRUCTION: Do NOT generate standard, flat "glassmorphism" or basic frosted glass. I require a physically accurate "Liquid Glass" aesthetic. It must look like wet, poured clear resin, combining the high-gloss specular highlights of classic macOS Aqua with the volumetric spatial depth of modern Apple VisionOS. 1. The Liquid Glass Material & Lighting (CSS): - Deep Refraction: Use `backdrop-filter` with extreme blur (e.g., 50px) and over-saturation (200%). - Specular Highlight: Create a curved, semi-transparent white gradient on the top half using a pseudo-element (`::before`) to simulate a hard light reflection on a wet, rounded 3D surface. - Caustics & Volume: Use multi-layered inner and outer `box-shadow` properties to simulate light refracting at the bottom edge and casting a realistic ambient drop shadow. - Interactive Glare: Implement a soft radial-gradient spotlight inside the glass that dynamically tracks the user's mouse cursor (X/Y coordinates) using JavaScript and CSS variables (`mix-blend-mode: overlay`). 2. Navigation Layout & Elements: - Center the pill-shaped navigation bar perfectly in the middle of the viewport. - Include 3 main navigation items with minimalist, inline SVG stroke icons and text labels: "Home", "Call", and "List". - Add a subtle vertical divider line after the main buttons. - Next to the divider, add a Dark/Light Mode toggle button containing inline SVG Sun and Moon icons. 3. Animations & "Apple Magic": - Sliding Active Pill: Create a solid background "pill" that sits *behind* the active navigation item's text/icon. When a different item is clicked, this pill must dynamically recalculate its width and slide to the new position. - Spring Physics: The sliding transition MUST use an exact Apple-style bouncy spring easing curve (e.g., `transition: all 0.5s cubic-bezier(0.34, 1.2, 0.64, 1)`). - Tactile Feedback: Buttons and icons must physically press down slightly (`transform: scale(0.92)`) when clicked (`:active`). - Theme Switch: The Sun and Moon icons must smoothly rotate, scale, and cross-fade during the transition. 4. Background Environment (Crucial): - Glass needs light and color to refract! Create a full-viewport, smoothly animated mesh gradient background using 3 large, heavily blurred, floating color blobs. - Implement full Dark/Light mode logic using CSS variables (`:root` and `[data-theme="dark"]`). Toggling the theme must seamlessly transition the background blob colors, glass opacity, shadow intensity, and text colors. Output ONLY the pristine, production-ready code. Prioritize maximum visual fidelity and silky-smooth 60fps animations.show more

hey if you have a 3060, or any GPU... with 8GB or more sitting in a drawer right now, that thing can run 9 billion parameters of intelligence autonomously. and you don't know it yet. 2 hours ago i posted that 9B hit a ceiling. 2,699 lines across 11 files. blank screen. said the limit for autonomous multifile coding on 9 billion parameters is real. then i audited every file. found 11 bugs. exact file, exact line, exact fix. duplicate variable declarations killing the script loader. a canvas reference never connected to the DOM. enemies with no movement logic. particle systems called on the class instead of the instance. fed that list as a single prompt to the same Qwen 3.5 9B on the same RTX 3060 through Hermes Agent. it fixed all 11. surgically. patch level edits across 4 files. no rewrites. no hallucinated changes. game boots. enemies spawn, move, collide. background renders. particles fire. and here's what nobody is talking about. this is a 9 billion parameter model running a full agentic framework. Hermes Agent with 31 tools. file operations, terminal, browser, code execution. not a single tool call failed. the agent chain never broke. most people think you need 70B+ for reliable tool use. this is 9B on 12 gigs doing it clean. the model didn't fail. my prompting strategy did. the ceiling is not the parameter count. the ceiling is how you prompt it. this is not done. bullets don't fire yet. boss fights need wiring. but the screen that was black 2 hours ago now has a full game rendering in real time. iterating right now. anyone with a GPU from the last 5 years should be paying attention to what is happening right now.show more

been getting a lot of qs on how i... vibe coded this demo. and the answer is: stop acting like a coder and start acting like a PM! my exact playbook👇 1. logic first ignore the UI/graphics. focus strictly on how it works. do a massive brain dump of every requirement into a plain txt file and drop it in your root folder. make it exhaustively detailed. 2. the stack · Nano Banana 2 (tiles/textures/3D refs) · Tripo AI (turning the 2D pics into actual 3D models) · Cursor + Opus 4.6 (doing the heavy lifting) · Netlify (deploy) 3. hire the AI PM feed your brain dump to an LLM. tell it to write a technical spec sheet like a senior PM would. drop that in the root too. 4. the scrum master have it break the entire project down into a markdown checklist by phases. notice we still haven't touched the codebase yet. that's the point. 5. let it cook go to your agent in Cursor and literally just prompt "Start Phase 1". test locally. bugs? tell it to fix. works? "Start Phase 2". loop this until you're done. Now you can focus on visuals, refine them with the agent. And yes, it can deploy it for you too if you give it perms. stop rushing to the editor. architect the idea and let the AI sweat the code. Open sourced repo in the qt post belowshow more

1 Neural Network + Obsidian + Karpathy’s 1-file method... = the most unhinged second brain build of 2026. It remembers everything you’ve ever done, and it costs $0 on top of what you already pay. The base is Karpathy’s append and review: 1 giant note, new thoughts stack on top, old ones sink, every few days you reread and pull the survivors back up. No folders, no tags, no plugins the rereading IS the system, because review is what turns storage into thinking. The flaw: past 10,000 lines, no human rereads anything. That’s where the neural network takes over. You keep the note in Obsidian 1 vault, everything dumps to the top: ideas, links, meeting fragments, half-thoughts. You never organize, you only dump. It all lives as plain markdown on your own disk, and that detail is the whole trick. Because now you point Claude Code at the vault folder, and it reads every line you’ve ever written. “What did I think about pricing in March.” “Find the 3 ideas I keep circling.” “What did I drop that deserves a second look.” It answers from YOUR notes, with quotes, in 15 seconds. Then once a week, 1 prompt closes the loop: read the last 7 days, surface the 5 entries worth pulling back up, flag anything that contradicts what I wrote a month ago. The model does the sinking and surfacing Karpathy did by hand, and the note stays alive instead of turning into a graveyard. Week 1 feels like nothing. Week 4 you hit the first “I already solved this in January.” Month 3 you consult your past self more than Google. Most second brains die in 11 days under 40 plugins and 200 folders. This one is 1 file and a loop, and it compounds because dumping takes 0 discipline. Notion stores what you thought. This thing argues back.show more

THREE 3090s ON ONE BOARD GIVE YOU 72GB OF... VRAM AND KILL YOUR $200 CLAUDE CODE AND $200 OPENAI BILL people are pulling three used 3090s off ebay for around $2,100 total and stacking them in one tower to build a dedicated ai rig. that pools 72gb of vram for less than what a single rtx 5090 retails for alibaba shipped qwen 3.6 27b in april under apache 2.0. on realworldqa vision it scores 84.1 against claude 4.5 opus at 77.0. on ifbench instructions it lands at 76.5 against claude's 58.0 a single 3090 already runs qwen 3.6 27b with eight gigs of headroom. three of them in parallel handle larger models like deepseek r1 70b and qwen 235b without breaking a sweat a heavy ai user pays $200 claude code, $200 chatgpt pro plus $40 cursor and gemini. that's $5,280 a year and the rig pays itself off before month nine on $8 a month in electricity setup is one shell command for ollama, one to pull the model, one environment variable to point claude code at localhost. cli stays identical, nothing leaves the network, requests stop costing money bookmark this and read the article belowshow more

💔 A Plea from a Father Exhausted by War... 💔 I am a father of a family of eight. I lost my job because of the war and have become unable to provide even the most basic necessities for my children. We are living through extremely difficult days, and only God knows the bitterness of what we are enduring. My fear for my children grows with each passing day. My youngest child is suffering from a severe skin rash, and the doctor has warned me that treatment must be provided urgently before it spreads to my other children. Yet I stand helpless before this pain, unable to afford the treatment or even the basic needs of my family. For a very long time, we have not received a single donation. Honestly, I am exhausted from pleading and waiting. Sometimes I ask myself: Have we been forgotten? Why does no one feel the suffering of my children? What have they done to deserve this pain and deprivation? If you were in my place, what would you do? I feel ashamed to ask for help, but I have no other choice left. My hope is in God, and then in kind-hearted people who still care about the suffering of others. We are not looking for luxury or comfort. We are simply seeking to live with dignity and to provide treatment and food for our children. I kindly ask you to visit my donation link and look at my family's situation with compassion. I am writing these words with a heart filled with pain that cannot be described. If you are able to help, please do not hesitate. And if you cannot donate, sharing this post may help it reach someone who can. I wish you and your families health, safety, and peace. Thank you to everyone who extends a helping hand, and may God reward you abundantly. 🤲❤️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

Today we’re introducing Google AI Threat Defense - a... comprehensive AI-powered cybersecurity solution designed to help continuously monitor for and stop AI-powered threats before they can impact your business. Here’s how it works: 1. AI Threat Defense uses our cybersecurity platform Wiz to scan and prioritize what applications and systems have the highest security risk. 2. Gemini and other frontier AI models can then autonomously perform continual deep scanning of your applications - starting with those at the highest risk - to identify security vulnerabilities. 3. The capabilities of CodeMender - a new software repair agent - are then used to verify and accelerate the patching of vulnerabilities. 4. And our Wiz autonomous agents continuously test your systems to find unknown vulnerabilities before adversaries do so that you can remediate them before you are attacked. While other model providers focus on using AI to find and flag vulnerabilities, Google AI Threat Defense actively prioritizes your most critical real-world risks and accelerates their remediation using a variety of models since no single model finds a superset of the vulnerabilities found by all other models.show more

A WEB STUDIO CHARGES $35,000 FOR AN ANIMATED SITE.... THE SAME BUILD NOW COSTS $12 - CLAUDE CODE WRITES, HIGGSFIELD RENDERS. Every agency billing $100-149/hr is just three departments. Here's each one, collapsed into a single agentic session. SYSTEM 1 - THE MOTION STUDIO (Higgsfield) Cinematic clips pulled from 30+ generative models - hero shots, transitions, ambient loops. → This used to be a motion artist on retainer. Now it's a prompt. SYSTEM 2 - THE DEV TEAM (Claude Code) Scaffolds the site, writes the GSAP ScrollTrigger timelines and Lenis smooth-scroll, extracts frames, optimizes every asset. → A full scroll-driven build with zero hand-coded keyframes. SYSTEM 3 - THE DESIGN DEPT (baked-in cinematic layer) Six effects with no config: film grain, particles, vignette, glass cards, color tints, scroll pacing. → The polish that justified the invoice - now it ships by default. Three departments. One operator. One pass. What used to take a designer, a motion artist, and a developer through weeks of handoffs now runs in a single session - for a Claude subscription and a few dollars of Higgsfield credits. The studio was never the talent. It was the overhead. And the overhead just became three systems. Reply "web-site" to this post and I will send you the step-by-step Playbook 👇show more

I just ran Gemma 4 31B on @CerebrasSystems at... 1,800+ tokens/sec and it's multimodal. For context: that's 35x faster than a typical GPU endpoint, and the first token (reasoning included) lands in 1.5 seconds. This isn't a benchmark slide, I recorded the inference live. Prompt I used: "Create a simulation of an iPhone. Include at least one working dummy note taking app, a functional notification pulldown, high quality graphics, single HTML file, any libs via CDN." - Generation time: 3 seconds. - Notes app worked. - Notification panel worked. - Rendered first try. This is what wafer-scale inference unlocks, not just "faster," but a different category of product. When generation is this fast, you stop waiting and start iterating in real time. Why this matters: Gemma 4 31B is Google DeepMind's flagship open weight model, Apache 2.0 licensed, dense (not MoE), and built for efficiency over raw parameter count. It scores close to Claude Haiku 4.5 on the Artificial Analysis Intelligence Index (30 vs 29) but runs ~18x faster on Cerebras. It's also the first multimodal model on Cerebras's platform, meaning you can now feed it screenshots, documents, charts, and UI states at wafer scale speed. # Applications I'm most excited about: - Screenshot → Insight: Drop in a dashboard or document screenshot, get structured findings back instantly. no waiting, no batching. - Live UI generation: Full interactive interfaces (like my iPhone sim) generated and rendered in under 2 seconds. - Screenshot -> Patch: Feed it a broken UI + console error, get a minimal code fix and verification steps back. - Computer use & agentic loops: See -> reason -> act - verify, fast enough to keep a human in the loop instead of waiting on the model. - Long context summarization: Full research reports condensed into decision ready summaries you can read and requery in one sitting. The bigger unlock isn't the speed number itself, it's that agentic and multimodal loops (see -> reason -> output -> tool call -> verify -> retry) finally run in real time instead of feeling sluggish. As Logan Kilpatrick (Logan Kilpatrick) put it: "If every model was doing 2,000 tokens per second, you wouldn't build the same product and just have it be faster, you'd build different products." Gemma 4 31B is live now on Cerebras Inference Cloud in public preview. If you're building multimodal, agentic, or real time apps, this is worth testing today. What would you build with such insane inference throughput?show more

my team didn't want me to give this away... for free. But I'm going to do it anyway it's the SEO & AI search dashboard I built in Claude Code it connects to your Google Analytics (GA4) and Google Search Console and Claude Code builds it in 5 minutes and I made a Notion document and a skill file so you can build this in Claude Code yourself in literally minutes the dashboard has three tabs: 1. AI Search - How much traffic is coming from ChatGPT, Perplexity, and Gemini ETC. It aggregates the GA4 data and gives single number 2. Paid ads - which keywords rank top 3 for but still pay for ads on, you should cut these to save budget 3. Organic overview - sessions, conversions, top landing pages, demographics. The single view for what is working I built this because this is how I drive our SEO and AEO forward it gives me the insights I need to allocate budget and prioritize what content to work on next I decided to give it away because most companies have no idea AI search is already sending them traffic like this post and comment "AEOdashboard" and I'll send it overshow more

Good morning Oscar; I’ve been pondering on how to... respond to these screenshots without context and eh; I am totally out of words. 1. Crown has been in school for EIGHT terms, you have paid for two and mostly on condition that you know where I stay, or act positive to your sexual advances or other random that fit your narrative then. Refusing to comply would mean no support. And that I was totally okay without financial support from you. 2. Don’t lie on my dad’s name like that. You almost broke my little brother’s hand fighting for crown in front of both of them. You have insulted my parents to their faces more than once and you are shocked that my dad isn’t besties with you. 3. Oscar, you paid that 1.2M after not talking to your child for a year. How are you shocked that she was in a different school and her school fees paid???? You sat in that meeting and said you are comfortable with crown in fort portal, she went to fort portal and then you got angry???😂😂 4. You can’t do the bare minimum like ask about your child’s school. You know crowns school, if you were intentional; you’d know that it is among the TOP THREE schools in fort portal. You can’t fail to consistently pay 1.2M for school fees, then refuse to pay 850,000 and then fight to pay 2.5M. How does that make sense??? 5. You have both my dad’s numbers, mom’s numbers, grand mom, aunty, twin brother. My dad told you, his home is ALWAYS open for you to visit your child. If you wanted to see your child, you would, but of course, please use me as your excuse. My family called for a meeting, you sat and agreed to terms that you couldn’t comply to thereafter and somehow I am to blame for your deadbeat tendencies?? 6. You want to have “full access” but won’t do the bare minimum to achieve that. You are angry that you showed up to my house at 10PM and no one opened? You are angry that you said “I love when you are sweet to me” because I was polite and I blue ticked you???? You are mad because I said crown needed consistency?? Be for real! 7. Crown used to call you, and you wouldn’t pick, she would reach out and never hear from you, you made crown pack a suitcase twice and wouldn’t show up and communicate. I don’t have to tell your child one single thing about you, never have, to her, you are a good person, she will make prayers for you, and ask God that one day you’ll call her, and yet, you never do that. 8. When I was still staying with crown, I welcomed you into my home, I even told you, I’ll make sure I’m not in the house when you are visiting and will come back after crowns sleeping hours. Just go, bond and do homework, put her to bed. You did that for three weeks, you then started showing up late in the night, not doing the homework, and being sexual. 9. YOU ARE A DEADBEAT FATHER!!!! The worst kind actually because you even have the audacity to try and make it look okay. I won’t be doing a back and forth with you, you are a very bad person. And No, I’m not happy that my child isn’t without her biological dad, but I’ll say this again, she is GOOD!!!! I work hard for that reason. My parents numbers will always be available if you want to visit or see her, you know my home where she is. But that said; you don’t pay school fees, you don’t offer support, you don’t check on her, you don’t pick her calls. I won’t be doing this back and forth with you, you are a deadbeat!!!! And like I told you; GO TO COURT AND HAND IN THESE SCREENSHOTS WITH THEIR CONTEXT!!! I would absolutely absolutely LOVE THAT!!!!🫱🏿‍🫲🏾🫱🏿‍🫲🏾🫱🏿‍🫲🏾🫱🏿‍🫲🏾show more

It's easy to make up figures, but Russia shows... you facts, what does Zelensky show to back up his claims? Nothing. For every 1 Russian casualty 20 Ukrainians fall. Russia dominates Ukraine 9 to 1 is fire power. A single FAB-3000 could eliminaтe 100+ soldiers. That's reality. You may not like it, CNN didn't say so, so that makes me wrong to you, but yet you show nothing to counter these figures. Drone warfare is fairly 1:1. Ballistic, heavy bombs, artillery is incomparable and you know it but can't accept these facts. I'm not here to celebrate over dеаd men, it hurts me to see Ukrainians dуing in such numbers. Russian soldiers are contracted, trained, well paid with high morale. Ukrainians? They are dragged off the streets into vans, 2 weeks of training if lucky and then off to the front. Do you criticize that? Or it's justified by Zelensky? Again, outside the physical domination by Russia, it is also psychologically more dominant. You either lie to yourself after reading this or you just ignorant. Western media has done more damage for the average Ukrainian than it has made gains in the PR warfare. Russia's win is slow, calculated, hold, devastate the lines and advance a few KMs. It's a conflict of attrition, many have been so brainwashed they believe the American war in Iraq is standard warfare or else Russia is losing. "Kiev in 3 days". Does Russia need to commit a genocide to win? Is that what you are asking. Be real with yourselves.show more