argsX @argsLab

Trump's situation rn

Narendra Modi @narendramodi

7 days ago

BRICS Summit 2026 gets underway in Delhi.

keyboard_arrow_left Previous keyboard_arrow_right Next

932 9K 65K 1.9M 855

Same here

Annant | Study Arc @annantHQ

6 days ago

Ppl ask me how did I get this much followers🛐 Because I don't do #letsconnect and will my followers cross this X amount of followers today wala chutiyapa🗿

keyboard_arrow_left Previous keyboard_arrow_right Next

0 0 1 223 0

@narendramodi Trump right now😵‍💫

@victor_bigfield Better than another AI wrapper: personal AI, dead man’s switch, genetic algo, homemade language, terminal black box. Posted the list here.

Watch-order if you only have 3 minutes: 0:13 personal AI 0:27 dead man’s switch 0:52 genetic algorithm 1:20 invent a language 1:45 terminal black box The point of the video is not the list. It’s this line: sit inside the part that breaks. I’ll drop a starting path for whichever number gets replied most.

Everyone is shipping the same AI to-do app. These 5 projects force you into the part of the system that actually breaks. 1. Your own personal AI. Your data. Your endpoints. Your rules. 2. A dead man’s switch. Miss the ping, it fires the action. 3. A genetic algorithm that starts ugly and gets unsettlingly good. 4. A tiny language you invent. Then you write the compiler. 5. A black box for your terminal. Last N minutes. Always rolling. Not anti-AI. Just stop letting the chatbot skip the hard part. Reply 1–5 with the one you’d actually start this week.

Doomscrolling is the cheapest dopamine on your phone. Same itch. Four builds that leave you with something real: 1. A personal API for your life — sleep, GitHub, workouts, reading list. One endpoint you own. 2. A “why is this slow” tool that maps CPU/RAM/disk spikes to the exact app that caused them. 3. Generative art from weather, sleep, or your own texts. You don’t know the output until it renders. 4. A Doom Tracker that turns your screen-time hours into a report of what those hours could have shipped. Reply with the number you’d start tonight. I’ll drop the starting path. “Stack” promises debugging. “Starting path” does not.

You deposit $1 million. The bank lends $1 million. It should have $0 left. It now has $2 million. This is not a glitch. This is how money is created. Jiang Xueqin traces it back 700 years. Medieval merchants found gold too heavy to move. They stored it and wrote paper receipts against it. Those receipts started circulating as money. The gold never left the vault. A second claim existed on the same gold. To keep the books honest they invented double-entry bookkeeping: every loan is recorded as an asset the second it is written. The original deposit stays a liability. Two entries. One pile of gold. Twice the money. That rule is still the law. It is why a bank can “lend” money that did not exist one second earlier. In 1913 they scaled the same trick to an entire country. Private banks formed a cartel and called it the Federal Reserve. The government borrows. The Fed creates the dollars and books the IOU as a Treasury. Interest is owed on money that was typed into existence. Everything after that sits on this ledger: the $40 trillion debt, your mortgage, the balance in your checking account, and the current stablecoin push. The trick was never hidden. It is written in the ledger. Twice. Watch the stretch where Jiang walks the same accounting from medieval receipts to 1913. Once you see it you cannot unsee it. Reply with one word: “vault” if you still thought banks lend existing deposits. “ledger” if this just clicked. Bookmark this. Someone will ask you where money comes from.

Not a course. The person who built the tool, out loud. Save it. Watch the first 10 minutes. @bcherny

What he actually teaches (not the recycled version of this post): 1. Don’t start with “write this feature.” Start with codebase Q&A. That’s how Anthropic onboards people. 2. Keep CLAUDE.md short. Every mistake becomes a rule so it never repeats. Long files get ignored. 3. Plan first. Code second. “Brainstorm. Make a plan. Show me. Then write code.” 4. Run multiple Claudes at once. Tabs + worktrees. One chat is a junior. Five is a team. 5. Give it a way to check itself. Tests, screenshots, a critic pass. Verification beats clever wording.

Most people prompt Claude like a chatbot. Boris Cherny — the engineer who built Claude Code — uses it like a team. 28min. Free. Better than most $300 courses. CLAUDE.md, parallel sessions, plan-before-code, self-verification. Watch this before your next session x.com/TechByArti/sta…

Stop paying $8/mo for X Premium just to upload long videos. Here is the hidden URL trick to embed ANY high-res video natively on your timeline (no downloading required): Step 1: Grab the link of the post containing the video. Step 2: Paste it into a new draft. Step 3: Append /video/1 right at the end of the URL. That’s literally it. That's how I do it. Clean, native, zero quality loss. Save this tweet for your next post. 🔖

@argsHQ @X Let's connect

Bookmark this thread before your next high-stakes conversation. Follow @annantHQ for daily high-signal frameworks, archived lectures, and strategic mental models. @JoelCPeterson

6. Durability is part of the outcome. A deal that turns into conflict six months later was a failure. The goal isn't to win the room—it's to build an agreement that survives in the real world.

2. Stop treating negotiation like eating broccoli. Most people treat negotiation as unpleasant, so they rush it or resent it. Shift your mindset from "conflict" to "creative problem-solving". Calm always beats aggression.

1. Needing the deal is losing the deal. If you enter a room desperate to close, you’ve already forfeited your leverage. The best negotiators protect themselves from ever getting trapped in desperate situations.

Instead of watching Netflix tonight, give 1 hour to this 2007 Stanford lecture by Professor Joel Peterson. It will teach you more about high-stakes persuasion, leverage, and human psychology than an MBA. Here are 5 psychological rules for getting whatever you want:🧵 x.com/jota_snchez/st…