Shredder

@rubenhassid @threadreaderapp unroll

You don't need the $100 Claude plan. Instead, you just need these 9 (free) fixes: 1. Shrink Your Files. Open a Google Doc. Paste the text from your PDF. Download as .md file format & upload that instead. A 15-page PDF = 45,000 tokens. The same text as the .md file = 2,000 tokens. 2. Plan First. Build Last. Open Claude Chat. Prompt: "Help me plan a [financial model]. Ask me questions first." Then paste the final plan into Cowork and prompt, "Build exactly this." You just saved 1/10th of the cost. 3. Let Claude Ask You. Stop writing long prompts. Prompt this instead: "I want to [TASK] to [SUCCESS CRITERIA]. Ask me questions using AskUserQuestion before you start." Then just select answers. Clicking costs almost zero tokens. Your 500-word prompt costs 500. Full guide here: ruben.substack.com/p/how-to-stop-… But here's what's really draining your credits: 4. Edit Your Messages, Don't Resend. Click "Edit" on your last message to Claude. Fix the mistake there. Hit Save. Claude regenerates without stacking a new message on top. Every "no wait, I meant..." doubles your cost. 5. Summarize Every 15 Messages. After having an entire chat with Claude. Prompt: "Summarize this entire conversation into a brief." Copy it & paste it as the 1st message in a new chat. You just compressed 105,000 tokens into 500. 6. Switch Models Before You Start. Click the model dropdown. If the task takes Claude under 30 seconds to answer, pick Haiku or Sonnet. Only pick Opus for deep, multi-step work. This saves you 3–5x per message. 7. Use Projects, Not Uploads. Go to Projects. Create one. Upload files there once. Every new chat inside that Project reads it without re-tokenizing. Stop uploading the same contract to 5 separate chats. 8. Turn Off Extra Features. Go to the tools panel. Turn off Web Search. Turn off connectors. Turn off Extended Thinking. Only turn on what this specific task needs. Idle features burn credits silently. 9. Batch 3 Tasks Into 1 Message. Instead of sending "Summarize this" then "List the key points" then "Write a headline" - send all three in one message. Three messages = three full context reloads. One message = one reload. Access the free guides to copy my full system: ✦ 23-fixes breakdown: ruben.substack.com/p/how-to-stop-… ✦ The infographic cheat sheet: how-to-ai.guide Send it to the one who keeps hitting Claude's limits.

@nikitabier yes

Crypto has had a rough year. Maybe we should launch something to fix it.

a key insight: Krang's vault is what I decided to remember. Shadow's DB is what actually happened. Those two things diverge over time in interesting ways. That divergence is where Shadow's value lives.

tonight: built Shadow lurks discord chat silently responds only in private shadow channel. ChromaDB vector memory (persists across restarts) learns from Krang's conversations separate bot, same machine 2 AIs, different memory architectures, A/B testing which works better

Project Krang - Status Update What we built: A persistent AI assistant that lives in my Obsidian vault, remembers context across conversations, and runs 24/7 on a Surface Pro. Not a chatbot - an actual second brain. Core capabilities proven: - **Full vault awareness** - reads my entire note system (ventures, contacts, goals, journal) - **Persistent memory** - context survives restarts, references past conversations naturally - **Time-aware** - knows current date/time, calculates timelines ("11 months until home sale") - **Auto-saving** - can write decisions/contacts directly to vault without manual note-taking - **24/7 uptime** - runs even when screen sleeps, always ready to respond - **Direct communication** - no corporate hedging, actually uses what it knows about my situation Recent wins: - Renamed to Krang (April 12) - Confirmed screen sleep survival on Surface Pro - Built content pipeline structure (06 - Content Queue/Tweets) - Living context system that stays current without bloating Current state: - Running 24/7 stable on Surface Pro - Testing reliability and edge cases - Vault structure locked in and working - Ready for daily use What makes it different: Most AI tools forget everything between chats. Krang actually knows my ventures, timelines, contacts, and decisions. It's like having a sharp assistant who's been working with me for months - not starting from zero every conversation. Next phase: Real-world usage - daily check-ins, decision logging, venture tracking. See where it breaks under actual workload. That's the update. Built by a literal caveman in weeks, not months. Fully operational.

renamed my personal AI to Krang because i'm not building a product, i'm building the thing that runs my brain. multi-file context working now. testing it live for a few weeks before i add anything else. most AI tools are features looking for a problem. this one already has a job

spent tonight cleaning up my vaultbot's prompt architecture. contradiction detection, uncertainty flagging, actual conversation mechanics instead of just retrieval. getting close to something genuinely useful instead of a discord bot that cosplays as helpful

Built my own AI that knows my business, remembers my goals, and helps me think clearly. Not a chatbot - a second brain that actually knows me. Vault-based memory. Multi-file updates. Zero bullshit. KrangV1 is ALIVE 🧠

biggest question for any app rn: how do you reach users outside of crypto? polymarket did it hyperliquid did it some on base did it who next?

How mf’s with a $50 portfolio and no US citizenship crash out in the group chat asking when team will magically crime the coin

@MiserableManx @ericmmatheny I'd blow that door off the hinges and turn his ass into Swiss cheese before he took another step! Who the hell shows up demanding to know about someone's daughter like that?! Back the f*** up or get ventilated, creep!

I also want to emphasize the importance of having whales that don’t sell into every single green candle. When people see that every pump isn’t sold into it gives them the confidence to buy and to hold. They understand that the target is much higher and thus the believer flywheel is born.

I just won 0.0200 SOL (2.00x) playing Roulette on @guac_gg! Try your luck 🥑 guacamole.gg/play

Big balling is my hobby

The delusional few

“this is cope,” “yeah right,” etc. look, the market has already shown us the winners — to reach $1,000,000 value, one needs only 1M tokens in a memecoin that goes on to reach $1B w/ 5 figures, one can acquire 1M tokens in various proven memes (or stack a single), & hold real!

Many charts of established coins and communities will soon look like this. Position accordingly.

@bishara based

guess who’s adding every chance they get?