aethertrace (@aethertrace_) header
aethertrace (@aethertrace_) profile picture

AI content creator | Researcher DMs always open

Joined July 2026
Photo by aethertrace (@aethertrace_)Photo by aethertrace (@aethertrace_)Photo by aethertrace (@aethertrace_)Photo by aethertrace (@aethertrace_)Photo by aethertrace (@aethertrace_)Photo by aethertrace (@aethertrace_)Photo by aethertrace (@aethertrace_)Photo by aethertrace (@aethertrace_)Photo by aethertrace (@aethertrace_)Photo by aethertrace (@aethertrace_)
aethertrace retweeted
holy sh*t, Obsidian just dropped an official Grok Bot plugin and people are quietly turning their vaults into 24/7 autonomous Second Brains your notes don’t have to be a dead archive. read, write, link, build Bases, draw Canvas maps, and sync without the desktop app. the 4-step setup: 1. ADD THE PLUGIN open the official plugin below → add to GrokBot. 7 skills: obsidian-markdown, obsidian-cli, obsidian-bases, json-canvas, obsidian-headless, defuddle, knap. 2. CONNECT YOUR VAULT local → settings → computer → your computer → execution on this computer. keep “ask every time.” the Obsidian CLI needs the app running with CLI enabled cloud → sync a vault copy to the bot’s computer. Obsidian Headless needs Node.js 22+ and an Obsidian Sync subscription for syncing 3. START IN /Inbox let it read selected notes, but write ONLY inside /Inbox. check one note, its properties, and its [[wikilinks]] first 4. GIVE IT A NIGHT SHIFT “every night at 11 PM, process raw dumps in /Inbox into atomic markdown notes with YAML properties. link to relevant existing notes. keep every write inside /Inbox. update /Inbox/Inbox.base and leave a change log.” test one run before scheduling it stop rebuilding your context from scratch in every chat. give your agent a vault it can actually work with build your second brain in 5 minutes ↓
Andrej Karpathy’s LLM Wiki + Grok Bot A Second Brain that turns your knowledge into finished work Put your notes on a night shift before another week slips away The complete setup and prompts below ↓
51
61
4
740
74,402
aethertrace retweeted
This is f*cking insane. Anthropic senior engineer: “Right now I’m running Fable 5.2 as the second brain while Opus 5.5 builds everything. Anthropic engineers are running dozens of 24/7 agent sessions in parallel this way. That’s the new agentic engineering stack.” Fable 5.2 → Orchestrator → Parallel Sessions → Auto-Verification → Shipped Work In this workshop she shows exactly how to get 100% value from Claude Code at minimum cost right now 99% people are still babysitting one chat Anthropic is running parallel sessions that plan, verify, and ship without constant supervision This 17-minute breakdown is worth more than most $3,000 agent engineering courses Watch it today before it gets buried Then build an AI team with Fable 5.2 & an AI employee for your company ↓
3
5
47
2,711
aethertrace retweeted
holy f*ck, everyone is quietly turning GrokBot into a 24/7 personal CFO while most people still bleed money on subscriptions, surge pricing, and forgotten trials... your bank account is paying a laziness tax 20 money-defense setups to build: 01–05 Refunds → late orders, fare discrepancies, flight delays, double charges, app store claims 06–10 Cuts → zombie trials, unused seats, SaaS downgrades, ISP negotiations, renewal alerts 11–15 Deals → price drops, restocks, coupons, flight fares, cashback checks 16–20 Cashflow → unpaid invoices, missed brand DMs, failed renewals, receipts, tax folders one money goal. strict rules. a paper trail subscriptions never forget to charge you. give your money a team that remembers to check start with the leak costing you the most today the underlying finance playbook + article ↓
Andrej Karpathy’s LLM Wiki + Grok Bot A Second Brain that turns your knowledge into finished work Put your notes on a night shift before another week slips away The complete setup and prompts below ↓
6
13
1
115
8,173
aethertrace retweeted
No f*cking way, do it ASAP... Build your own GTA 6 clip studio with Fable 5.2 20 tools + Fable 5.2 + GTA 6 = that’s all you f*cking need A new wave of millionaires is on the horizon $$$... gameplay → moments → vertical edits → captions → review → publishing queue 01 / CAPTURE + INGEST 1. OBS Studio: record gameplay and commentary github.com/obsproject/obs-st… 2. yt-dlp: import authorized sources github.com/yt-dlp/yt-dlp 3. MediaInfo: inspect codecs, audio and frame rates github.com/MediaArea/MediaIn… 4. rclone: back up the originals github.com/rclone/rclone 02 / FIND THE MOMENTS 5. faster-whisper: transcribe commentary github.com/SYSTRAN/faster-wh… 6. WhisperX: align words to timestamps github.com/m-bain/whisperX 7. PySceneDetect: find shot boundaries github.com/Breakthrough/PySc… 8. Silero VAD: locate spoken segments github.com/snakers4/silero-v… 03 / CUT + REFRAME 9. AutoClip: propose highlight candidates github.com/artbyjazi/autocli… 10. OpenShorts: an alternative clipper with API + MCP github.com/mutonby/openshort… 11. OpenCV: build custom action and HUD tracking github.com/opencv/opencv 12. FFmpeg: cut, encode and combine the output github.com/FFmpeg/FFmpeg 04 / CAPTIONS + FINISH 13. pysubs2: create timed subtitle files github.com/tkarabela/pysubs2 14. libass: burn readable captions into clips github.com/libass/libass 15. Pillow: make covers from actual frames github.com/python-pillow/Pil… 16. Auto-Editor: propose dead-time cuts for review github.com/WyattBlue/auto-ed… 05 / QUEUE + MEASURE 17. n8n: connect jobs and approval steps github.com/n8n-io/n8n 18. PostgreSQL: track sources, jobs and edit versions github.com/postgres/postgres 19. YouTube API: upload private drafts developers.google.com/youtub… 20. Metabase: build performance dashboards github.com/metabase/metabase Use Fable 5.2 to build the glue between the tools. Start with one clipper, then add custom control where it matters Build your workflow based on your own video materials or those from streamers. Automate routine tasks and keep every draft ready for review master prompt below, start building, everything is in your hands ↓
Opus 5.5 can run a whole motion design studio Sell a 20-second launch clip built from real product screens Copy the motion engineering prompts and delivery process Build your offer before the next product launch ↓
26
11
116
4,987
aethertrace retweeted
This is f*cking insane. Anthropic engineer: “Fable 5.2 is already smarter than you know how to use The bottleneck was never the AI. It’s you.” In 19 minutes he shows how to get everything out of Claude with no extra tools and no extra cost Fable 5.2 → Prompt Engineering → Memory Engineering → Skills → Agents → Harness Engineering → Everything You’re already paying for all of it 90% of users just never find it Worth more than most $2,000 Claude courses Bookmark and watch it today Then build an AI team with Fable 5.2 & an AI employee for your company ↓
3
4
40
737
aethertrace retweeted
This is f*cking insane. The creator of Claude Code, Boris Cherny, was asked if people should still bother learning to code. His answer: “In a year or two, it’s not going to matter.” In 73 seconds he lays out what actually works. Play with the tools. Learn them deeply. Stay on the frontier. Don’t be scared On his team everyone ships now. PMs, designers, even the finance guy He hasn’t touched a single line of code by hand since last November. He ships 10 to 30 pull requests a day Writing Code → Prompting → Agents → Harness → Loops → Everyone Ships Claude Code is the harness. Tools, memory, rules and loops around the model. His team rewrites it with every new model and already cut 80% of the system prompt Save this before it gets lost in X’s algorithms… Every guide and repo below is FREE: Claude Code Docs code.claude.com/docs/en/feat… Equipping Agents With Agent Skills anthropic.com/engineering/eq… Defending Code Reference Harness github.com/anthropics/defend… Agent Skills Open Standard github.com/agentskills/agent… Claude Platform Release Notes platform.claude.com/docs/en/… How We Contain Claude anthropic.com/engineering/ho… Claude API Skill platform.claude.com/docs/en/… Anthropic CLI github.com/anthropics/anthro… OSS Scanner github.com/anthropics/oss-sc… Knowledge Work Plugins github.com/anthropics/knowle… Claude Code Base Action github.com/anthropics/claude… He compares it to the printing press. Writing used to belong to a few scribes. Now everyone can write. Building software is next 99% will keep debating whether to learn to code The 1% who build a harness will ship 30 pull requests a day Bookmark it and start with the first link right now Then read how to build an AI team with Claude & an AI employee for your business ↓
2
7
34
3,126
aethertrace retweeted
holy sh*t, SpaceXSI engineer, Lauren Tan just published a GrokBot playbook for automating your work your to-do list is starting to look like a hiring list i turned her method into 20 copy-paste bot job setups: 01–05. Clients → leads, meeting prep, replies, follow-ups, reviews 06–10. Content → topics, research, drafts, repurposing, calendars 11–15. Product → feedback, bugs, specs, release checks, release notes 16–20. Admin → inbox, meeting notes, documents, reports, reminders each gets one job, a clear output and a way to check its work the setup: tell Dr Eggbot the goal, explain how you do it, teach it your judgment calls. get one task working, then turn the process into a skill and routine start with the task you keep putting off save this, then read the article below on building a Second Brain for your company with Grok Bot ↓
Andrej Karpathy’s LLM Wiki + Grok Bot A Second Brain that turns your knowledge into finished work Put your notes on a night shift before another week slips away The complete setup and prompts below ↓
7
4
1
42
1,335
aethertrace retweeted
This is f*cking insane. Former Google CEO Eric Schmidt just shared the easiest way to get rich in 2 years: “If you really want to make money, it’s actually easy. Found an agentic AI company.” Here’s the 2-year window + the 12 best free tools to build & launch one: In 27 seconds he lays out the play: build an agent that actually does something — if it’s better than anyone else’s, you can build an incredible company and own it yourself Fable 5.2 → Prompts → Memory → Skills → MCP → A Company You Own Yourself The next 1–2 years are the agentic window. And it won’t stay open Save this before it gets lost in X’s algorithms Everything below is FREE. Start today: Building Agents With the Claude Agent SDK anthropic.com/engineering/bu… Claude Managed Agents platform.claude.com/docs/en/… Claude Agent SDK (TypeScript) github.com/anthropics/claude… Claude Agent SDK Demos github.com/anthropics/claude… Mem0 github.com/mem0ai/mem0 Agent Skills Specification agentskills.io Model Context Protocol Servers github.com/modelcontextproto… Composio github.com/ComposioHQ/compos… Claude Commerce Agents github.com/anthropics/commer… Claude for Financial Services github.com/anthropics/financ… E2B Agent Sandboxes github.com/e2b-dev/e2b Langfuse github.com/langfuse/langfuse All of this costs $0 99% will spend the entire agentic period scrolling threads The 1% who actually ship agents will own the companies that come out of it Bookmark it and start with the first link right now Then read how to build an AI team with Claude & an AI employee for your business ↓
4
6
1
70
5,139
aethertrace retweeted
Before you blame Fable 5.2, learn these 25 official Anthropic commands 25 essentials to get more out of Fable 5.2: less setup friction → custom tools → agent workflows steal the cheat sheet ↓ 01 /fast → lower response latency code.claude.com/docs/en/fast… 02 /autocompact → tune auto compaction code.claude.com/docs/en/mode… 03 /skills → curate your skill set code.claude.com/docs/en/skil… 04 /skill-doctor → find context-heavy skills code.claude.com/docs/en/skil… 05 /reload-skills → refresh skills without restarting code.claude.com/docs/en/skil… 06 /plugin → extend your toolkit code.claude.com/docs/en/plug… 07 /plugin-authoring → build a custom mod code.claude.com/docs/en/plug… 08 /reload-plugins → apply plugin changes live code.claude.com/docs/en/plug… 09 /update-config → edit settings by intent code.claude.com/docs/en/sett… 10 /permissions → set precise tool rules code.claude.com/docs/en/perm… 11 /auto-mode-setup → draft your environment rules code.claude.com/docs/en/auto… 12 /fewer-permission-prompts → reduce repeat approval prompts code.claude.com/docs/en/comm… 13 /add-dir → work across folders code.claude.com/docs/en/perm… 14 /import → bring your existing setup over code.claude.com/docs/en/comm… 15 /teleport → bring cloud work into your terminal code.claude.com/docs/en/clau… 16 /remote-env → choose your cloud environment code.claude.com/docs/en/clau… 17 /run-skill-generator → teach Claude how to run your app code.claude.com/docs/en/skil… 18 /dataviz → design readable charts and dashboards code.claude.com/docs/en/comm… 19 /slides → turn a brief into a presentation code.claude.com/docs/en/arti… 20 /artifact-diagramming → create clearer diagrams code.claude.com/docs/en/arti… 21 /output-style → change how Claude responds code.claude.com/docs/en/outp… 22 /doctor → diagnose setup problems code.claude.com/docs/en/skil… 23 /list-agents → find reachable agents and sessions code.claude.com/docs/en/comm… 24 /workflow-authoring → write custom agent workflows code.claude.com/docs/en/work… 25 /workflows → watch, pause and resume workflows code.claude.com/docs/en/work… start here: /skills → /plugin → /workflows save this, then build an AI team with Claude & an AI employee for your company ↓
29
19
1
144
13,641
aethertrace retweeted
This is f*cking insane. CEO of Anthropic, Dario Amodei: “I have engineers at Anthropic who say: I don’t write any code anymore.” Reality is changing for everyone. You need to adapt It’s even better if you turn a $0 open-source Claude Code stack into an AI development team The real skill is building the operating system around the model Save this before it gets lost in X’s algorithms Every resource below is FREE: STAGE 1 — Build AI Fluency & Repeatable Workflows (No code needed) > Anthropic Academy Official free courses with certificates — from AI Fluency to Claude Code anthropic.skilljar.com > OpenAI Academy Self-paced tracks on AI foundations, workplace workflows, and delegating to agents academy.openai.com > Learn Prompting Structured open-source curriculum on prompting, context design, and AI safety learnprompting.org STAGE 2 — Turn AI Into a Coworker (Tools, Browsers & Automation) > Hugging Face MCP Course Free hands-on course with certification on Model Context Protocol — connecting AI to real apps and data huggingface.co/learn/mcp-cou… > n8n Free Courses Automate entry-level ops, finance, and research workflows visually using AI agent nodes docs.n8n.io/courses > Browser Use Open-source framework letting AI agents navigate websites and finish web tasks like a junior hire github.com/browser-use/brows… STAGE 3 — Build & Orchestrate Real AI Agents > Google × Kaggle 5-Day AI Agents Intensive Hands-on crash course covering agent architecture, tools, memory, and evaluation kaggle.com/learn-guide/5-day… > LangChain Academy Free step-by-step courses on stateful agents, memory, and human-in-the-loop control with LangGraph academy.langchain.com > OpenAI Agents SDK Official lightweight Python framework for multi-agent handoffs and sandboxed execution github.com/openai/openai-age… STAGE 4 — Ship Reliable Systems & Understand How They Work > 12-Factor Agents (HumanLayer) The engineering playbook for building AI agents reliable enough for real production work github.com/humanlayer/12-fac… > AI Engineering Hub Tiered, hands-on real-world builds across RAG, MCP, and multi-agent apps github.com/patchy631/ai-engi… > LLMs From Scratch Code a GPT step by step on a regular laptop to truly understand what runs under the hood github.com/rasbt/LLMs-from-s… All of this costs $0 Bookmark it and start with the first link right now Then read how to build an AI team with Claude & an AI employee for your business ↓
1
3
1
35
876
aethertrace retweeted
This is f*cking insane. Anthropic CEO Dario Amodei just told high school kids the truth most people are still ignoring: “AI can already do entry-level law. Entry-level finance. Entry-level consulting We can’t deny it will eliminate jobs. Probably a large number of them.” In a few minutes he names the jobs that disappear first and the one skill that keeps you ahead His prediction: 50% of entry-level white-collar jobs disrupted in the next 1 to 5 years Prompts → Agents → AI Coworkers → Entry-Level Work Done By AI Every past shift gave people decades. This one hits every stage of a career at the same time His advice is one line: “Of course, learn to use AI.” The people who start this year won’t compete with AI. They’ll be the ones running it Everything below is free. Start today: STAGE 1 — Build AI Fluency & Repeatable Workflows (No code needed) > Anthropic Academy Official free courses with certificates — from AI Fluency to Claude Code anthropic.skilljar.com > OpenAI Academy Self-paced tracks on AI foundations, workplace workflows, and delegating to agents academy.openai.com > Learn Prompting Structured open-source curriculum on prompting, context design, and AI safety learnprompting.org STAGE 2 — Turn AI Into a Coworker (Tools, Browsers & Automation) > Hugging Face MCP Course Free hands-on course with certification on Model Context Protocol — connecting AI to real apps and data huggingface.co/learn/mcp-cou… > n8n Free Courses Automate entry-level ops, finance, and research workflows visually using AI agent nodes docs.n8n.io/courses > Browser Use Open-source framework letting AI agents navigate websites and finish web tasks like a junior hire github.com/browser-use/brows… STAGE 3 — Build & Orchestrate Real AI Agents > Google × Kaggle 5-Day AI Agents Intensive Hands-on crash course covering agent architecture, tools, memory, and evaluation kaggle.com/learn-guide/5-day… > LangChain Academy Free step-by-step courses on stateful agents, memory, and human-in-the-loop control with LangGraph academy.langchain.com > OpenAI Agents SDK Official lightweight Python framework for multi-agent handoffs and sandboxed execution github.com/openai/openai-age… STAGE 4 — Ship Reliable Systems & Understand How They Work > 12-Factor Agents (HumanLayer) The engineering playbook for building AI agents reliable enough for real production work github.com/humanlayer/12-fac… > AI Engineering Hub Tiered, hands-on real-world builds across RAG, MCP, and multi-agent apps github.com/patchy631/ai-engi… > LLMs From Scratch Code a GPT step by step on a regular laptop to truly understand what runs under the hood github.com/rasbt/LLMs-from-s… Total cost: $0 You don’t need another year of school You need to start telling AI what to do 95% will graduate and fight AI for the same entry-level job The 5% who learn to run agents this year will manage the ones doing that job Save this and send it to someone graduating this year Then read how to build an agent that actually does the job in the "Context, Harness, Loops, Prompts" guide below ↓
Your agent can spend all night repeating the same mistake Prompt, Context, Loop, and Harness Engineering Find the failing layer and learn exactly what to fix The four-layer playbook for agents that finish the job ↓
4
4
1
27
1,154
aethertrace retweeted
A single prompt card mass-produces 30+ specialist agents. Fable 5.2 is wild for spinning up self-improving agent teams. Feed it one prompt and it designs, recruits and continuously upgrades 30+ specialized agents on its own while you sleep. The prompt card lays out context, harness, loops and prompt rules for the whole team.
34
10
142
8,353
aethertrace retweeted
This is f*cking insane. Elon Musk just put a number on Grok 5 reaching AGI for the first time: “Grok 5 will be the smartest AI in the world by a significant margin, on every metric, without exception I calculate like 10%. And it’s really going to feel sentient.” In 8 minutes he walks through what SpaceXAI is training right now. 6 trillion parameters. Text, images, video and audio in one model that builds its own tools Grok 3 → Grok 4 → 6 Trillion Parameters → One Model For Every Sense → AGI He has never put AGI above zero before Worth more than any $3,000 AI course Save this before it disappears from your feed Every resource below is free and built for running Grok Bot teams right now: Grok Bot Getting Started Official guide to create your first bot x.ai/bot/guides/getting-star… Grok Bot for Engineering Run a fleet of engineer bots x.ai/bot/guides/grok-bot-for… Grok Bot for Product Managers A team that reports to you x.ai/bot/guides/grok-bot-for… Grok Bot for Marketing Research competitors and draft campaigns x.ai/bot/guides/grok-bot-for… Grok Bot for Founder-Led Sales Research and outreach with bots x.ai/bot/guides/grok-bot-for… Grok Bot Templates Share and reuse bots as recipes x.ai/bot/guides/templates-fo… Grok Bot Marketplace Ready-made bots you can start using today x.ai/bot/marketplace Awesome Grok Resources Community list of Grok tools and guides github.com/xai-org/awesome-g… All of this costs $0 Bookmark and watch it today Then read how to build your own team of GrokBot agents in the harness engineering guide below ↓
Founders using Grok Bot are already one step ahead of you Give your one-person company an AI workforce Reduce routine outsourcing and reclaim founder time Copy the full Grok Bot harness below ↓
26
143
8
636
49,940
aethertrace retweeted
this is f*cking insane. a 20-year-old problem in black hole physics just got solved by Fable 5.2... Anthropic’s internal model did the calculation physicists couldn’t finish b = −1.48 × 10⁻³ λ² The negative sign means the transition stays first-order and happens a bit below the Hagedorn temperature. A physicist gave the task to the model. It got the answer. They checked it on known special cases. A completely separate run with a different method produced the exact same number. They later connected this internal model to Claude Fable 5.2. Claude did original theoretical physics work that humans had been stuck on for years... Still needs independent verification. But the fact that it happened at all is hard to ignore Paper: hexagonmath.org/2611.00175v1 Save this ↓
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:
8
5
48
4,095
aethertrace retweeted
HOLY SH*T, OpenAI co-founder Ilya Sutskever went back to the University of Toronto and told the graduates: "The day will come when AI will do all the things we can do" Not some of them. All of them. Anything a person can learn, AI will be able to do too In 72 seconds, he explains how he can be so sure: your brain is a biological computer, so why can't a digital one learn everything it can Some people give it 3 years, some 5, some 10. He says it's hard to predict, but AI will keep getting better, maybe not so slowly Biological Computer → Digital Brain → Agents → Harness → AI That Can Do Every Job Later in the speech he borrows an old line: you may ignore politics, but politics won't ignore you. He says it applies to AI many times over Save this before it gets lost in X’s algorithms... Everything below is FREE: LLM course with roadmaps github.com/mlabonne/llm-cour… Neural networks from zero to hero github.com/karpathy/nn-zero-… Build a tiny neural net engine github.com/karpathy/microgra… Train your own GPT github.com/karpathy/nanoGPT The best ChatGPT $100 can buy github.com/karpathy/nanochat Build a ChatGPT-like LLM from scratch github.com/rasbt/LLMs-from-s… 21 lessons on generative AI github.com/microsoft/generat… 50+ agent tutorials github.com/NirDiamant/GenAI_… 12 principles of agents that actually work github.com/humanlayer/12-fac… Production-grade agents github.com/NirDiamant/agents… Batteries-included agent harness github.com/langchain-ai/deep… Autonomous agent loop github.com/snarktank/ralph Agent memory github.com/thedotmack/claude… Claude cookbooks github.com/anthropics/claude… Anthropic Academy anthropic.skilljar.com All of this costs $0 AI will learn every skill you have. The only edge left is being the one who runs it 95% will wait to see which jobs survive, the 5% who build a harness now will be the ones AI works for Save it and start with the first link right now Then read how to build an agent that actually does the job in the "Context, Harness, Loops, Prompts" guide below ↓
Your agent can spend all night repeating the same mistake Prompt, Context, Loop, and Harness Engineering Find the failing layer and learn exactly what to fix The four-layer playbook for agents that finish the job ↓
1
6
34
788
aethertrace retweeted
Fable 5.2 is insane for building self-evolving agent swarms… send it this prompt and it will design a 40-agent team that plans, builds, reviews, benchmarks, remembers failures, and improves its next run then read the full "Context, Harness, Loops, Prompts" guide ↓
Your agent can spend all night repeating the same mistake Prompt, Context, Loop, and Harness Engineering Find the failing layer and learn exactly what to fix The four-layer playbook for agents that finish the job ↓
33
20
1
261
27,853
aethertrace retweeted
This is f*cking insane. Boris Cherny, the creator of Claude Code, was asked if people should still bother learning to code. His answer: “In a year or two, it’s not going to matter.” In 73 seconds he lays out what actually works. Play with the tools. Learn them deeply. Stay on the frontier. Don’t be scared On his team everyone ships now. PMs, designers, even the finance guy He hasn’t touched a single line of code by hand since last November. He ships 10 to 30 pull requests a day Writing Code → Prompting → Agents → Harness → Loops → Everyone Ships Claude Code is the harness. Tools, memory, rules and loops around the model. His team rewrites it with every new model and already cut 80% of the system prompt Save this before it gets lost in X’s algorithms... Every guide and repo below is free: Claude Code overview code.claude.com/docs/en/over… Best practices for Claude Code code.claude.com/docs/en/best… Anthropic on building effective agents anthropic.com/research/build… Long-running agent harnesses anthropic.com/research/effec… Claude Agent SDK github.com/anthropics/claude… Skills system code.claude.com/docs/en/skil… Hooks documentation code.claude.com/docs/en/hook… Subagents guide code.claude.com/docs/en/sub-… Memory in Claude Code code.claude.com/docs/en/memo… Claude Code in Action course anthropic.skilljar.com/claud… Awesome Claude Code github.com/hesreallyhim/awes… Claude Code GitHub Action github.com/anthropics/claude… Agent plugins collection github.com/wshobson/agents Claude cookbooks github.com/anthropics/claude… He compares it to the printing press. Writing used to belong to a few scribes. Now everyone can write. Building software is next 95% will keep debating whether to learn to code. The 5% who build a harness will ship 30 pull requests a day Bookmark it and start with the first link right now Then read how to build an agent that actually does the job in the "Context, Harness, Loops, Prompts" guide below ↓
Your agent can spend all night repeating the same mistake Prompt, Context, Loop, and Harness Engineering Find the failing layer and learn exactly what to fix The four-layer playbook for agents that finish the job ↓
2
14
2
105
11,254
aethertrace retweeted
HOLY F*CK, ELON MUSK + SPACEX ENGINEERS BUILT THESE 100 GROKBOT AGENTS FOR US, AND YOU CAN TURN THIS INTO A SECOND BRAIN keep your product + customers + voice + decisions in a small context pack → give the relevant brief to the specialist that owns the job 01–10 Research → markets, competitors, validation 11–20 Product → specs, prototypes, landing pages 21–30 Build → coding, QA, security, shipping 31–40 Leads → prospects, signals, outreach drafts 41–50 Sales → call prep, pipeline, follow-ups 51–60 Writing → blogs, newsletters, sharper copy 61–70 Social → clips, posts, product videos 71–80 Growth → SEO, AI visibility, campaigns 81–90 Admin → inbox, meetings, project tracking 91–100 Money → receipts, invoices, cash flow capture what you know. put it to work save this, then read the article below on building a Second Brain for your company with Grok Bot ↓
Andrej Karpathy’s LLM Wiki + Grok Bot A Second Brain that turns your knowledge into finished work Put your notes on a night shift before another week slips away The complete setup and prompts below ↓
5
13
86
4,417
aethertrace retweeted
THIS IS AN ACTUAL F*CKING GOLDMINE Opus 5.5 just killed the jobs of motion designers... anyone can build your own motion design studio with Opus 5.5 + Claude Code one product action → a launch clip → a reusable production workflow give every move a job: 01 assets + brief supply the logo, product screens and reference, then define the shot, copy and reading hold 02 frame timeline pin the entrance, button press and result reveal to specific frames 03 preview + revision inspect the stills, watch the moving scene, then give the agent one precise edit 04 export + reuse deliver the clip and keep the approved entrances, cursor paths and text components for the next project the loop: brief → build → preview → scoped edit → render → review → reuse start with one launch clip + one cover + one revision round save this, then read the full guide for Opus 5.5 Motion Engineering ↓
Opus 5.5 can run a whole motion design studio Sell a 20-second launch clip built from real product screens Copy the motion engineering prompts and delivery process Build your offer before the next product launch ↓
12
41
2
443
46,485
aethertrace retweeted
your Slack workspace could become an entire AI engineering team @vorflux just launched Vorflux 2.0, with AutoBots, 24/7 AI coworkers that own ongoing responsibilities instead of waiting for individual tasks → monitor production and investigate bugs automatically → turn customer calls into engineering fixes → review code, run tests, and prepare PRs → manage projects and track progress → coordinate work across multiple AI agents the part that matters: these agents can identify what needs attention and start working on it without someone constantly assigning tasks think customer feedback: an agent picks up a bug from a recorded call, proposes a fix, gets approval, and follows it through deployment that's the pitch: moving from AI agents that complete tasks to AI coworkers that keep the work moving
🤝 Paid partnership
Today we're launching @vorflux 2.0: AutoBots, 24x7 coworkers that live in Slack. I previously co-founded @Rippling, now $10B+, engg team of 1000s. Engg changes crazy fast now. IDE autocomplete > CLI > Cloud agents > multi-player slack agents Models deliver predictably now. Bottleneck to parallelism is to initiate, keep track and ping the right people for approvals. Hire from our directory of opinionated AutoBots or make a new one with a goal. Bring your Codex/Claude Code subscriptions and Vorflux is practically free. We recently closed a $15M seed round for @vorflux backed by @ycombinator, @peakxvpartners, @alliancedao, along with incredible partners including @parkerconrad, @immad, @typesfast, @jake_zeller, @balajis, @nivi, @metakovan, @lmrankhan, @nikitabase, @0xrwu, @ayushjaiswal, @mattshumer_, @eshamanideep, @sreeramkannan, @dvcoolster, @nusimow, @TeddySolomon11, @ashtoncofer, @rvivek, and others. Lmk which autobot you’d like to try and we’ll double your credits to try. Chapters: 0:00 Introducing Vorflux 2.0 0:22 From manager to director 0:44 How an AutoBot works 1:04 Cloud setup 1:25 On-call AutoBot 1:42 Project manager AutoBot 2:03 Customer experience AutoBot 3:01 Designer AutoBot 3:52 Social AutoBot 4:51 Open and frontier models 5:23 Always-on agents
20
4
55
1,514