ZuluVox @ZuluVox

Karpathy method + Claude Code reading your whole Obsidian vault is the smartest second brain on earth. The method is simple and brutal. If you can’t build a thing from scratch, you don’t know it. Tutorials are fake learning and your brain deletes them in 3 days. Most people ignore this. They build a second brain that just sits there, folders of notes nobody reopens, dead text. Point Claude Code at the vault and it wakes up. 5,000 notes, one mind. It reads all of it and answers in your own words and your own proofs, not a model’s guess. Then the loop closes. Want to understand neural nets? Skip the 3-hour video and ask Claude Code to build a tiny one. 200 lines from scratch. Watch it train, break a layer, watch it fail, fix it. It clicks in 20 minutes instead of 3 weeks. The second it lands the note gets written. One idea per file, linked to 10 others, dropped into the vault while the memory is still hot. Now it compounds. Month 1: is 60 notes. Month 6 is 900. Every new note pulls in old ones, so you ask anything and the answer comes from your brain, not the internet. Before: 40 tabs, 6 half read PDF, 0 retained. After: build it once, own it for life. Setup takes 4 minutes. Plain text, no lock-in. A second brain nobody reads is a graveyard. Yours just started thinking.

Spike 1% @SpikeCalls

a week ago

x.com/i/article/2069…

17 92 477 1.3M 2K

Watch

@RealEntheos @grok is this true?

A Google Cloud engineer just built a full app with Claude from scratch with zero errors 😭 24 minutes. completely free. worth more than any $500 vibe-coding course here's what he covers: - the 14% you lose to CLAUDE.md before typing a word > going from zero to deployed app in one session > Claude replacing an entire engineering team > no coding background needed to start if you've been using Claude for months and never left the chat window, you skipped the most important 30 minutes of the entire experience that's exactly why I put together a guide on Claude rules most people have no idea exist instead of another show tonight, watch this make sure to bookmark it before it gets lost in your feed the guide is in the article below

Khairallah AL-Awady @eng_khairallah1

a month ago

x.com/i/article/2060…

19 23 133 130K 274

there's a free site with ready-to-use Claude Code templates and almost nobody knows it exists sara found it. 9,700 downloads. 1,900 github stars. zero mainstream attention. one npm install and claude code stops being a blank slate graphql architect. prompt engineer. business analyst. hackathon strategist. llms maintainer. search specialist. pre-built agents for every task. you pick the template, it loads the context, claude knows exactly what role it's playing before you type a word pause at 0:10 and scroll through the template library. that's not a prompt. that's a configured specialist agent ready to deploy the difference between vibe coding and actually shipping is knowing which tool to load before you start npm install -g claude-code-templates bookmark this before it gets behind a paywall

obssnnn @0xObssnnn

2 weeks ago

x.com/i/article/2068…

6 9 99 622K 191

The guy who BUILT Claude Code is running 10–15 parallel AI agents like an engineering team. Not prompts. Systems. His secret isn’t some hidden feature. It’s a simple file: CLAUDE.md And it changes everything. Every time Claude makes a mistake → it writes a rule. Every correction → permanent memory. Every session → smarter than the last. > “Update your CLAUDE.md so you don’t make that mistake again.” That’s the loop. No repeated errors. No wasted tokens. No babysitting. Just compounding intelligence inside your own codebase. While most people: Rewrite the same prompts Fix the same bugs Start from zero every time He’s building a self-improving engineering system. And it gets crazier: • 10+ agents running in parallel • Research, coding, testing — all split into sub-agents • Clean context, zero clutter • Complex problems = more agents, not more thinking He hasn’t written SQL in 6+ months. Claude just pulls from BigQuery via CLI. This isn’t “AI-assisted coding.” This is AI orchestration. And the gap is already showing. Claude Code is now contributing to ~4% of all public GitHub commits. If you’re still using AI like a chatbot… You’re not behind. You’re playing a completely different game.

Nainsi Dwivedi @NainsiDwiv50980

2 weeks ago

x.com/i/article/2069…

17 27 95 58K 66

This is a new paradigm for interacting with Claude that is significantly more "inline" with all the other human activity org-wide. Once you do all of the under the hood engineering work to make this "just work" (e.g. across tools, integrations, compute environments, memory, security, etc.), Claude basically joins the team in a seamless way - you can talk to it as you would talk to a person and it can help with a very large variety of workloads. Imo this is the 3rd major redesign of LLM UIUX. The first paradigm was that the LLM is a website you go to, the second was that it is an app you download to your computer. This third one is that it is a self-contained, persistent, asynchronous entity with org-wide tools and context, working alongside teams of humans. It really takes a while to wrap your head around it, but it works and it is awesome.

Claude @claudeai

2 weeks ago

Introducing Claude Tag, a new way for teams to work with Claude. In Slack, Claude joins as a team member with access to the channels and tools you choose. Tag Claude in and delegate tasks to it while you focus on other work.

2K 2K 28K 20.2M 12K

his “AI server” is 4 mac minis stacked next to the toaster linked by free software into one machine that runs 235B models no single computer could load regulated clients pay him $3,500/month each to use it he has 6 of them that’s $21k/month from a tower of silver boxes in his kitchen the stack runs exo - open-source, $0, turns four separate minis into one pool of memory. the monitor shows all four nodes firing together, one model spread across the whole cluster the thing nobody tells you about local AI: a model either fits in your machine’s memory or it doesn’t run. that’s the wall everyone hits. it’s why people assume serious AI means a $40k server exo walks straight through that wall. it splits one massive model across every box in the chain. four minis at $599 each suddenly run what a single high-end machine can’t touch and the people who need this aren’t chasing the newest model they’re the ones the cloud locked out: → a law firm with privileged case files → a medical practice under HIPAA → an accounting firm holding client financials they need AI nowhere near a third-party server. a stack they can point to during an audit ends every objection their models run on his cluster. the data never leaves equipment he controls the breakdown: → 4 mac minis: $2,400 total → exo software: $0 → electricity: ~$25/month → what he charges per client: $3,500/month → 6 active clients the kitchen stack paid for itself in the first 3 weeks month 1: 2 clients, $7k month 3: 4 clients, $14k month 6: 6 clients, $21k he’s not running a startup with an office and a logo he’s running four silver boxes between the coffee machine and a bowl of fruit, clearing $21k/month from hardware that cost $2,400 clients picture racks and cooling fans reality is four minis humming next to where he makes breakfast the cloud companies charge a fortune for compute and access to your own he charges $3,500 a month for the one thing they can’t sell - a machine that never phones home

regent0x @regent0x_

3 weeks ago

x.com/i/article/2066…

13 30 260 1.3M 1K

🚨 Anthropic’s Research Lead just dropped a masterclass on AI agents. "99% of our engineers run swarms of 300+ self-improving agents". The key? Closing the loop so models can verify their own work. In 20 minutes, they unpack continuous Claude loops, plan modes, and dynamic workflows. Skip the $300 courses. This is free, and better. Then read this ace Loop Engineering article by Rari 👇

rari @0xwhrrari

3 weeks ago

x.com/i/article/2065…

46 160 917 2.4M 3K

Andrew Ng: "100% of my tasks now run through AI agents, the hype actually passed my expectations, loops are the next step" "In 3-6 months, everyone will be using self-improving loops, prompting alone is over" In a 30-minute talk, Andrew Ng breaks down how to build self-improving agentic systems from scratch Agent loops Feedback Memory Self-improvement Worth more than most $500 agentic courses

rari @0xwhrrari

3 weeks ago

x.com/i/article/2065…

46 160 917 2.4M 3K

Most people use Obsidian to store notes. He used it to build a brain that thinks for him while he sleeps. Inside Obsidian he runs Claude. He stopped saving notes to read later. Now he dumps raw thoughts into one inbox and Claude sorts them. Each idea moves on its own: Inputs, then a stage called The Crucible, then a commit gate. By morning the weak ones are cut and one is tagged to build. Today it picked "AuthConcierge." He didn't choose it. The vault did. Then he clicks a tab called The Galaxy. His whole brain as a 3D world, every project a planet you fly through and click to open. He doesn't search his notes. He flies over them.

Moysei @0xMoysei

2 weeks ago

x.com/i/article/2069…

15 172 1K 1.9M 4K

I made $16,000 last month. Never showed my face. Never hired a creator. Never touched a camera. Just 40 videos. A $20 Claude subscription. And a workflow that most people still don't know exists. 3 of those videos pulled 2 million views combined. Back in the day, the biggest barrier to UGC was finding the right creator — negotiating rates, waiting on revisions, hoping they delivered. Not anymore. 40 videos. A few hours of work. $16,000. Here's the exact five-step build I used.

Stijn Feijen @spwfeijen

a week ago

x.com/i/article/2069…

213 132 459 564K 1K

INSTEAD OF WATCHING AN HOUR OF NETFLIX TONIGHT. This 60-minute Cambridge lecture by Demis Hassabis will teach you more about the future of AI than most people will learn in the next 5 years. Bookmark it and give it an hour, no matter what.

🎯21 hacks to never hit Claude's limit. ✨ Stop wasting your daily window and get maximum output with minimum tokens.Use these power-moves to make Claude faster, cheaper, and more accurate. 👉 HACK #1: Convert files before uploading Don’t upload PDFs, screenshots, or PPTX files raw. Copy-paste the text into a Google Doc, download it as .md, upload that instead. 👉 HACK #2: Plan in Chat, build in Cowork Don’t open Cowork & say “Create a financial model.” Open Chat first. Plan the structure. Then move to Cowork with “Build this exact file.” 👉 HACK #3: Say “Ask me questions” Write this: “I want you to [task] to accomplish [criteria]. Read my folder. Ask me questions using AskUserQuestion before you start.” 👉 HACK #4: Stop redoing the whole thing When section 3 is wrong, say “Only redo section 3. Keep everything else.” Add “No commentary. Just the output.” 👉 HACK #5: Edit your original message In Chat, click Edit on your original message instead of writing a new one. Change the text. Hit send. It regenerates. The old exchange gets replaced. (Works in Chat only, not Cowork.) 👉 HACK #6: Batch tasks into one message Instead of 3 separate prompts (“Summarize this” + “Then edit this” + “Suggest a headline”), send one prompt: “Summarize, list points, suggest a headline.” 👉 HACK #7: Reuse the same prompt structure I keep a prompt library. Swap only the variable part. I use the same 30-word prompt for 80% of my Cowork sessions. 👉 HACK #8: Pick the right model Grammar check? Reformat? Use Sonnet or Haiku. Save Opus + Extended Thinking for the real work. 👉 HACK #9: Keep your files short Cowork reads your folder before every task. Keep it under 2,000 words. Too many words = burning tokens. 👉 HACK #10: Restart, don’t follow up When Cowork gets it wrong, don’t type “No, I meant…” Click Restart the conversation from here on an earlier message. 👉 HACK #11: Summarize every 15–20 messages Ask Claude to summarize the conversation every 20 prompts. Copy the summary. Open a new session. Paste it as your first message. 👉 HACK #12: Use Projects for recurring files Stop uploading the same PDF to five different chats. Use Projects instead. Upload the file once. 👉 HACK #13: Turn off features you don’t need Web search, connectors, Explore mode — they all add tokens. Default: everything off. Turn features on per task. 👉 HACK #14: New topic = New chat You asked about a LinkedIn post. Then a proposal. Keep it clean. The same chat confuses the model. New topic, new chat. Always. 👉 HACK #15: Don’t dump your whole folder Only include files Claude needs for this specific task. For tasks that don’t need files, select 0 folders. 👉 HACK #16: Schedule your recurring tasks Don’t run the same weekly digest. Use the /schedule plugin. Set it once. It runs on its own. 👉 HACK #17: Stop using Claude for things it can’t do Claude can’t make images. Need images? ChatGPT. Need real-time search? Grok. 👉 HACK #18: Speak your prompts for more context Use wispr.ai to speak your prompts for richer context in one take. Fewer follow-ups, fewer reloads. 👉 HACK #19: Set Up Preferences Go to Settings → General → Personal Preferences. Turn off Memory. 👉 HACK #20: Prompt Claude Code tightly Use very specific prompts: “Build a chart from this CSV. Save as chart.png.” 👉 HACK #21: Spread across the day Claude uses a rolling 5-hour window. Split it. Morning, afternoon, night. Stop using all your window before usage has rolled off. ❤️ Like 🔁 Retweet 🔖 Bookmark Follow @HeyAnjula for more such posts

INSTEAD OF WATCHING NETFLIX TONIGHT. Spend 1 hour with this. Claude AI FULL COURSE that teaches you how to BUILD and AUTOMATE anything. The people who watch this tonight will wake up tomorrow with a new skill. Watch it and Bookmark it now.

David, co-creator of MCP at Anthropic: "There's more to the protocol than what most people use it for." In a 26-minute talk, he shows how shallow most people's use of these tools really is. Most people are scratching the surface. The new high-paid role lives in the depth they never reach. That gap is exactly where the money is moving. Watch the talk, then see the role it's creating in the article below. Save this before it scrolls away.

Zephyr @Zephyr_hg

a week ago

x.com/i/article/2069…

0 7 56 112K 150

Anthropic engineers just showed how to build agents that can run for days with loops "At Anthropic, 30%+ of our code is already written by loops. That's why we ship so fast" In this 40-minute workshop, they break down the whole stack: Agent loop Harness Memory Sub-agents This is worth more than most $500 vibe-coding courses Bookmark and watch the workshop today, then read the article below

rari @0xwhrrari

3 weeks ago

x.com/i/article/2065…

46 160 917 2.4M 3K