satoshi 001 (@satoshiwhynot) header
satoshi 001 (@satoshiwhynot) profile picture

Building AI agents & quant systems. Sharing practical AI tools, visual guides & LLM workflows for solo creator

On-chain AI
Joined September 2019
satoshi 001 retweeted
GTA 5 came out in 2013. I was in school. Now I have bills, back pain and a 212 GB download. Goodbye Los Santos. Vice City, I'm home. 🌴
8
3
22
773
f*ck, your second brain just grew hands. Obsidian has 7 official skills for Claude Code and Codex: linked notes, visual maps + headless sync and publishing. your research can live in files you can actually open and edit. repo + setup below
🤖 Made with AI
Obsidian Skills now has seven skills that teach your agents how to fully operate your vaults and wikis 1. obsidian-markdown Create and edit Obsidian Flavored Markdown with wikilinks, embeds, callouts, properties, and other advanced syntax 2. obsidian-bases Create and edit Obsidian Bases (.base) with views, filters, formulas, and summaries 3. json-canvas Create and edit JSON Canvas files (.canvas) with nodes, edges, groups, and connections 4. obsidian-cli Interact with Obsidian vaults via the Obsidian CLI including plugin and theme development 5. obsidian-headless Sync and publish vaults from the command line with Obsidian Headless, without the desktop app 6. defuddle Extract clean markdown from web pages using Defuddle, removing clutter to save tokens 7. knap Render Markdown templates from JSON or CSV data using Knap, including batch Markdown file generation
1
2
29
F*ck yes. You can give Claude Code a Fable 5.2 brain and let cheaper models do the grunt work. Opus 5.5 builds. Sonnet 5.5 checks the changes. Haiku 5.5 handles focused research in parallel. Fable gets called in for the decisions that could send the whole job down the wrong path. For prompts up to 100k tokens, Haiku's standard API input and output rates are 80 times lower than Fable's. Paying Fable to handle every little lookup gets expensive fast. Set it up like this: Run claude update. You need v2.1.321 or later and Fable access. Start with claude --model opus --advisor fable. Confirm the advisor is enabled. Paste the setup prompt in my first reply. It prepares the user-level defaults, a Haiku researcher and a Sonnet reviewer, so they're available across projects. It also asks Claude to back up the affected files and show you the diff before applying changes. Tell the main agent when to consult Fable: before committing to a difficult plan, when the same error keeps coming back, and before signing off on a complex task. Check the transcript for actual advisor calls. Every consultation reads the conversation, so keep an eye on usage as the session grows. Don't skip step four. In Anthropic's GPQA test, Haiku and Sonnet made zero calls to an Opus advisor across 198 questions per run. No system instruction asked them to consult it. Full setup prompt and official docs in the first reply. My Fable 5.2 guide is quoted below.
Claude Fable 5.2 is the most insane AI model in history... 99% of people will never touch it or burn money on it: "Too expensive", "Not for me" Use it right, and it becomes your second BRAIN, for a fraction of the price In this article, I show you how:
2
35
Fable 5.2 brought in $28,460 in 48 hours from a Meta Ads campaign. Total ad spend was $214.80. I gave it the product page, a few creatives and a $250 spending limit, then left it running. When I checked the dashboard, I genuinely thought it was showing a revenue forecast. I refreshed it twice before opening the orders. I spent longer deciding what fucking tea to order than setting up the campaign. Now I’m sitting here with a cold cup of tea and a very annoying amount of confidence.
Claude Fable 5.2 is the most insane AI model in history... 99% of people will never touch it or burn money on it: "Too expensive", "Not for me" Use it right, and it becomes your second BRAIN, for a fraction of the price In this article, I show you how:
1
2
33
MCP CASE STUDY - ANTHROPIC: 150 000 TOKENS TO 2 000 The agent accessed the tool it needed for copying meeting notes from Drive into Salesforce, however it loaded the whole tool catalog at the beginning. By performing this task on an as-needed basis, we’ve saved 98.7% of time spent waiting. Prior to upgrading as the result of our guide, it’s best to check the following information on how your agents load for each task. 1. Monitor the input, output, cache read and turn counts for each task. 2. Make sure the instructions that will be consistently reused will always remain at the start of each task, and the specific task details will change afterwards. 3. Make sure to test the small model for sorting, extraction and formatting. 4. Load the required tools for the task as needed. 5. Set strict usage limits for tokens and turning points in order to prevent persistence on a task where there has been consecutive problems. 6. Batch jobs that can wait, if your API allows this behavior. Write up the policy in TOKENS.md and implement the limits in your implementation. The video demonstrates the MCP case study and you can find the entire guide on the website.
Anthropic measured it: one agent burns 4x the tokens of a chat, a team of agents 15x The people running 12 agents on one plan aren't paying more, they're rereading less 6 levers, 6 token leaks and a TOKENS.md you can copy ↓
1
31
I gave an AI agent Blender. It gave my orange 48 suction cups The agent wrote a Blender Python script to build the peel and tentacle, then animated the camera around the model The .blend file stays editable. If I want a darker material or a different camera angle, I can change those directly A close-up caught overlapping cups in the first pass, so the agent fixed the spacing before the full render. If you want to try AI + 3D, these 8 repos cover different parts of the process: MCP for Blender Connect a coding agent to Blender so it can inspect and change objects, materials and lighting github.com/ahujasid/mcp-for-… TRELLIS.2 Generate a 3D asset with PBR materials from a reference image github.com/microsoft/TRELLIS… Hunyuan3D Generate geometry and textures from images, with separate shape and texture stages github.com/Tencent-Hunyuan/H… ComfyUI Build repeatable image and video workflows for concepts and reference material github.com/Comfy-Org/ComfyUI BlenderProc Script scenes, camera positions and render passes. Particularly useful for synthetic datasets github.com/DLR-RM/BlenderPro… TripoSR Reconstruct a 3D object from a single image github.com/VAST-AI-Research/… Motion Canvas Create 2D motion graphics and explanatory animations with code github.com/motion-canvas/mot… Gaussian Splatting Reconstruct a real scene from photos so you can view it from new angles github.com/graphdeco-inria/g… Get a good still before animating. Check the back of the model too. The camera won’t politely avoid it
2
51
I want my video edited before my tea goes cold. That’s the appeal of an AI team for me. Give the agents a clear brief, hand over the repetitive work, and review what comes back. I’d happily spend less of the afternoon sorting files and more of it on the final cut. The Fable 5.2 guide below covers the team setup and prompts, including a separate reviewer to check the work before you approve it.
Claude Fable 5.2 is the most insane AI model in history... 99% of people will never touch it or burn money on it: "Too expensive", "Not for me" Use it right, and it becomes your second BRAIN, for a fraction of the price In this article, I show you how:
3
55
The creator who shared this Fable 5.2 demo says it’s already made and saved him over $10,000 Then you watch the agents work through Meta Ads: campaign objective, budget, targeting, copy, creative, preview. The demo ends on "Campaign published" That’s a business task people pay to get done The article below shows how to structure the team: Fable handles planning and coordination, while other models take execution, review and repetitive work Give the team a complete brief, approved assets and clear spending limits The full setup, model choices and copy-paste prompts are in the article below
Claude Fable 5.2 is the most insane AI model in history... 99% of people will never touch it or burn money on it: "Too expensive", "Not for me" Use it right, and it becomes your second BRAIN, for a fraction of the price In this article, I show you how:
2
3
47
satoshi 001 retweeted
CLAUDE CODE NOW LETS YOU PROMPT ITS OWN UI You can ask for custom colors and animated backgrounds directly in the desktop app. Start with the theme-examples plugin, or describe the theme you want and have Claude build it with mods. Try this: “Create a dark theme with warm text, muted violet accents and a slow-moving starfield. Keep code blocks and diffs high-contrast. Make the animation subtle and respect reduced-motion settings.” Then refine it like any other design: brighter text, less movement, a quieter background. Official mods docs: code.claude.com/docs/en/plug… T3 Code users got custom themes too, in Nightly. Agents can change its colors and layout through theme colors and CSS. T3 Code Nightly: github.com/pingdotgg/t3code/… The Mythos theme below is a T3 Code JSON with a dark palette and animated stars. Use it as a reference for that app: Mythos theme JSON: gist.github.com/lubluniky/63…
You can now build custom themes for Claude Code on desktop with mods! Change the colors, add an animated background, and make it your own. Just tell Claude how you want it to look, or get started quickly with the theme-examples plugin. We'd love to see what you make!
2
2
9
53
CLAUDE CODE NOW LETS YOU PROMPT ITS OWN UI You can ask for custom colors and animated backgrounds directly in the desktop app. Start with the theme-examples plugin, or describe the theme you want and have Claude build it with mods. Try this: “Create a dark theme with warm text, muted violet accents and a slow-moving starfield. Keep code blocks and diffs high-contrast. Make the animation subtle and respect reduced-motion settings.” Then refine it like any other design: brighter text, less movement, a quieter background. Official mods docs: code.claude.com/docs/en/plug… T3 Code users got custom themes too, in Nightly. Agents can change its colors and layout through theme colors and CSS. T3 Code Nightly: github.com/pingdotgg/t3code/… The Mythos theme below is a T3 Code JSON with a dark palette and animated stars. Use it as a reference for that app: Mythos theme JSON: gist.github.com/lubluniky/63…
You can now build custom themes for Claude Code on desktop with mods! Change the colors, add an animated background, and make it your own. Just tell Claude how you want it to look, or get started quickly with the theme-examples plugin. We'd love to see what you make!
2
2
9
53
An AI rebuild can look perfect on video and still fail when you type two spaces into search. That’s the test I want to see after the agents finish. REA lets your coding agent inspect the original app. The next step is turning those observations into a behavioral spec another agent can implement—and tests that expose what it missed. The workflow: → Pick one feature you can compare against the original. → Record what it does. Separate evidence from assumptions. → Write the behavioral spec before coding. → Give a separate implementation agent only that spec and the permitted test data. → Run both versions on identical inputs. Keep the failing cases. Check result ordering, pagination and errors too. A search box returning something is a low bar. My article walks through the full Opus 5.5 + REA workflow, with the prompts and a worked offline search example:
3
8
227
I can unscrew this shark's eyeball without opening a 3D editor 😂 This is my hammerhead version, built from Cutaway's open-source Three.js examples. three lens rings pull apart around a 10-blade iris both hammer lobes slide out with 18 riveted armour plates the glass flank opens over the drive gears and glowing coils the lower jaw drops and the 64 glass fangs separate from their rows around 0:15, the whole model is apart, then snaps back together Cutaway gives Claude Code a workflow for rebuilding a reference picture in Three.js. The result is editable geometry, so you can change the model, materials and camera moves. Skill, examples and setup: github.com/Nyxera0x/cutaway
2
10
281
Your AI clone can look perfect until someone types two spaces. Give the original and the rebuild the same inputs. Compare the results, including ordering and errors. 18 GitHub repos for investigating software, rebuilding a feature and checking what actually matches: [01] analysis and reconstruction REA: MCP tools for inspecting apps, binaries and runtime behavior. github.com/morluto/rea Craft Anything: study the original, build your own version, compare and refine. github.com/storytold/craft-a… Ghidra: free disassembly and decompilation. github.com/NationalSecurityA… GhidraMCP: connect your coding agent to Ghidra. github.com/LaurieWired/Ghidr… [02] Android and .NET jadx: decompile Android DEX into readable Java. github.com/skylot/jadx Apktool: decode Android resources and rebuild modified APKs. github.com/iBotPeaches/Apkto… ILSpy: inspect and decompile .NET assemblies. github.com/icsharpcode/ILSpy dnSpyEx: debug .NET programs and edit their assemblies. github.com/dnSpyEx/dnSpy [03] runtime and network behavior Frida: instrument an app while it runs. github.com/frida/frida mitmproxy: inspect HTTP and HTTPS traffic. github.com/mitmproxy/mitmpro… x64dbg: step through Windows programs in a debugger. github.com/x64dbg/x64dbg RAD Debugger: a graphical debugger for native programs. github.com/EpicGames/raddebu… [04] deeper binary analysis radare2: a reverse-engineering toolkit for the terminal. github.com/radareorg/radare2 rizin: another command-line framework for binary analysis. github.com/rizinorg/rizin ImHex: inspect binary data with a hex editor and pattern language. github.com/WerWolv/ImHex angr: use symbolic execution to explore program behavior. github.com/angr/angr [05] find the next tool awesome-ai-reverse-engineering: a catalog of AI-assisted RE tools with maintainer reviews. github.com/zhuyansen/awesome… awesome-reverse-engineering: morluto's collection of RE tools and learning resources. github.com/morluto/awesome-r… The workflow: app → evidence → SPEC.md → independent rebuild → test against the original → fix mismatches Start with one feature you can test. Give the implementation agent only the behavioral spec, then check its output against the original. Keep the failing cases. I put the full Opus 5.5 + REA workflow in the article below, with the prompts and a worked search example showing how to measure the match ↓
1
5
275
Dan Koe laid out how one person could build a creative business with AI and charge $10K-$75K a month across 1-3 clients. I broke down his playbook into 3 visual guides, from brand strategy to the first client.
4
10
208
If you're worried about AI taking your job, these 7 GitHub repos are worth trying before 2027. They let you hand over work you might still be doing manually, from tailoring a resume to editing a video. career-ops (80k stars) Scans job boards, scores each role against your CV, and tailors your resume for every application. github.com/career-ops-hq/car… Agent Reach (107k stars) Gives your agent access to YouTube, Reddit, GitHub, LinkedIn and the open web. One command to set it up. github.com/Panniantong/Agent… OfficeCLI (32k stars) Lets agents create and edit Word documents, Excel spreadsheets and PowerPoint presentations without Office installed. github.com/iOfficeAI/OfficeC… Hypit (26k stars) Turns any viral video into an editable workflow, including footage, captions, b-roll and effects. github.com/hypit-ai/hypit AI Website Cloner (37k stars) Give it any website URL and it turns it into a clean Next.js app you can edit. github.com/JCodesMore/ai-web… Strix (72k stars) Autonomous AI agents that look for vulnerabilities in your code and demonstrate them with working exploits. github.com/usestrix/strix Paperclip (110k stars) Run a company of AI agents from one dashboard, with an org chart, budgets and approvals. github.com/paperclipai/paper… If you'd watch more videos like this, give it a like and send it to a friend. Tell me in the comments which agent you want me to visualize next.
14
31
332
Grok Bot can turn a recording of your workflow into a draft skill. I'd try it on a weekly report, then test what happens when a source is missing before putting it on a schedule. This guide has the prompts for that setup
1
9
104