MrQ @OnlyQuez

A guy named Shin just open-sourced Palantir for free. It is called OpenPlanter. Palantir helps governments and giant companies connect scattered data about people, businesses, contracts, and money. OpenPlanter could give that same power to ordinary people. Feed it corporate records, political donations, lobbying disclosures, government contracts, or other public documents. Its AI agents search through everything, identify the same people and companies across different datasets, and turn the connections into a live knowledge graph. You can click an entity, follow its relationships, and inspect the original sources behind each finding. The desktop app already works on macOS, Windows, and Linux. It supports OpenAI, Anthropic, OpenRouter, and local models through Ollama. But OpenPlanter is still an early-stage project. It is nowhere close to having the data, resources, or maturity of Palantir. Still, the idea is huge. Powerful institutions have spent years using data to investigate the public. Now the public may finally get a tool for investigating them.

@baker_girlie If this was a black dude he’d have been shot 30 seconds in!

Dropping straight knowledge! Get on this people.

x.com/i/article/2068…

Most LLM routers decide where to send a prompt by… calling another model. Wayfinder doesn’t. It reads the prompts structure - deterministic, offline, microseconds - to keep cheap prompts on your local model and send hard ones to the frontier ones. Self-hosted, bring your own keys, OpenAI-compatible

@AlfieJCarter Stack

Scientists have identified a reversal of the long-standing Flynn effect—the roughly 200-year trend of rising average intelligence (measured via IQ and cognitive tests) across generations. For the first time in modern recorded history, Generation Z (born roughly 1997–2012) shows lower performance than previous generations in key cognitive domains, including attention, memory, literacy, numeracy, executive function, problem-solving, and general IQ—despite spending more years in formal education than ever before. Neuroscientist and educator Dr. Jared Cooney Horvath, PhD, MEd, testified before the U.S. Senate Committee on Commerce, Science, and Transportation on January 15, 2026, highlighting this shift. In his written testimony, he stated that cognitive development in children across much of the developed world has stalled or reversed over the past two decades, with declines evident in international assessments (e.g., PISA, TIMSS) and other large-scale data starting around the mid-2000s and accelerating post-2010. Horvath attributes the primary driver not to reduced schooling, but to the widespread integration of digital screens and educational technology (EdTech) in classrooms. He argues that human brains evolved for deep, focused learning through face-to-face interaction and sustained attention, not fragmented skimming or constant task-switching encouraged by devices. Key points from his testimony include: - Teens now spend over half their waking hours on screens, with significant portions in school involving computers or tablets—often leading to off-task behavior and shallower processing. - Evidence from meta-analyses and national/international studies shows a consistent pattern: higher classroom screen exposure correlates with weaker outcomes in reading, math, science, and higher-order reasoning. - Digital tools may aid narrow, repetitive skill practice in controlled settings, but in core academic contexts, they tend to reduce depth of understanding, retention, and critical thinking. Horvath describes this as a "structural mismatch" between human cognition and how digital platforms are designed (to capture and fragment attention), warning that unchecked EdTech adoption risks long-term harm to workforce skills, innovation, and societal reasoning. [Horvath, J. C. (2026). Written testimony before the U.S. Senate Committee on Commerce, Science, and Transportation. U.S. Senate]

In a significant breakthrough for surgical robotics, researchers at Johns Hopkins University have developed an AI system that enabled a robot to successfully perform gallbladder removal on a deceased pig with almost complete autonomy and no direct human control during the procedure. The dual-layer AI model, trained on 17 hours of real human surgical video footage, interpreted live visual input and converted it into precise verbal-style commands (such as “clip the second duct”), which were then executed by the robot. Across eight separate procedures involving 17 distinct surgical tasks, the robot achieved a remarkable 100% success rate, demonstrating advanced real-time self-correction capabilities and only minimal human assistance (primarily for tool changes or occasional adjustments). While fully autonomous surgery on living humans is still years away, this milestone represents a major step forward in the field. The robot’s ability to detect and fix its own mistakes highlights the growing potential of AI to enhance surgical precision and safety. Researchers plan to advance testing to live animal models next, with the long-term vision of developing systems that can reduce complication rates, improve consistency, and help alleviate the workload on surgical teams. [Kim, J. W., et al. (2026). SRT-H: A hierarchical framework for autonomous surgery using language-guided imitation learning. Science Robotics, 10(104), eadt5254. DOI: 10.1126/scirobotics.adt5254]

x.com/i/article/2067…

@JulianGoldieSEO Agent OS

Google's former CEO just said what everyone in AI already knows Building wealth is getting easier if you actually learn the tools Not by scrolling AI threads By understanding agents, Codex Code, prompts, memory, skills, MCP, and routines Save this before it disappears from your feed Everything below is free: ChatGPT basics help.openai.com/en/articles/92… OpenAI Academy academy.openai.com Prompt engineering developers.openai.com/api/docs/guide… GPT-5.5 prompting guide developers.openai.com/api/docs/guide… OpenAI API docs developers.openai.com/api/docs Responses API developers.openai.com/api/docs/guide… Agents SDK developers.openai.com/api/docs/guide… Agents SDK quickstart developers.openai.com/api/docs/guide… Tools + function calling developers.openai.com/api/docs/guide… Structured outputs developers.openai.com/api/docs/guide… Conversation state developers.openai.com/api/docs/guide… ChatGPT memory help.openai.com/en/articles/85… ChatGPT projects help.openai.com/en/articles/10… Custom GPTs help.openai.com/en/articles/85… Tasks in ChatGPT help.openai.com/en/articles/10… Codex overview developers.openai.com/codex Codex quickstart developers.openai.com/codex/quicksta… Codex CLI developers.openai.com/codex/cli Codex GitHub repo github.com/openai/codex Codex best practices developers.openai.com/codex/learn/be… AGENTS.md developers.openai.com/codex/guides/a… Codex skills developers.openai.com/codex/skills Codex MCP developers.openai.com/codex/mcp Codex subagents developers.openai.com/codex/subagents ChatGPT Apps SDK developers.openai.com/apps-sdk Apps SDK quickstart developers.openai.com/apps-sdk/quick… ChatGPT developer mode + MCP developers.openai.com/api/docs/guide… OpenAI Cookbook github.com/openai/openai-… All of this costs $0 Most people will keep asking AI one question at a time Save this and start learning the stack

rari @0xwhrrari

a month ago

x.com/i/article/2065…

54 188 1K 3.3M 4K

x.com/i/article/2066…

Important AI context to keep in mind when using these tools

A fired Jane Street quant walked out with 10 years of private BTC trading data. Turned it into $1.5M. He did not build a bot. He built a simulator that runs every move Bitcoin can make before it makes one. I found his wallet. Been copying him for a week. PnL prints like clockwork. Here is what he actually built. A swarm of agents feeds 10 years of stolen tick data into MiroFish. A god-tier agentic simulator. It does not forecast the next candle. It spins up a virtual market and plays Bitcoin forward through thousands of scenarios at once. Six agents each validate their own call. A trade only fires when they converge. They collect data 24/7, rerun the sim, and remember every pattern, every reaction, every signal they have ever seen. He does not predict the future. The math already knows it. He just reads the numbers and takes the money. Here is the part firms do not want public: MiroFish just broke algo trading. The desks are quietly building their own simulators right now. The window where one solo wallet can run this is still open. Barely. I rebuilt his approach using Claude. One prompt. Fed it the same framework. Let it run. The agent monitors his wallet 24/7. Copies every position in real-time. No delay. No guessing. Just mirror and profit. You only need Claude + device + 1 hour to deploy. Giving this free for 24 hours. To get it: 1. Comment the word "QUANT" 2. Like and retweet this post 3. Follow me @codewithimanshu so I can DM you Save this post. Build the copytrading system this week. Start with $200. Scale on evidence.

This Chinese quant built a simulation of how SPX price reacts to any global event. Made over $100K with full blockchain proof. He knows exactly where price will go. I found his wallet. Reverse-engineered the entire system. Then used Claude new Fable 5 rebuild it in under an 1hr. Here is what he actually did. Loaded 40+ years of SPX trading history into MiroFish simulator. 18K stars on GitHub. Open source. Agentic framework. AI analyzed every single moment in that trading history. Now he has a fully functional SPX price prediction system. Dozens of successful SPX trades. Hundreds of tests across other stock markets. Here is exactly what the stack looks like: → Market data APIs for SPX price (Alpha Vantage or Quandl) → Data pipeline built in Python → Feature engineering for output signals like RSI and MACD → Seed dataset for MiroFish (convert data into structured context) → Multi-agent simulation (macro strategist, earnings analyst, sentiment analyst agents) → Probability forecast running different scenarios → Trading decision model (SPX futures ES, SPY ETF) The agents do not predict. They simulate. They run every possible market reaction to news, earnings, Fed moves, geopolitical events. Then they converge on the highest probability outcome. Trade only fires when all agents agree. I rebuilt this exact pipeline with Claude Fable 5 (Mythos) Fed it the framework. Connected the data sources. Let it simulate. The agent monitors SPX patterns 24/7. Runs scenarios in real-time. Alerts me when convergence hits. You can feed the whole thing to Claude and build your first simulation model together. Even a small one. You only need Claude + device + 1 hour to deploy. Giving this free for 24 hours. To get it: 1. Comment the word "Claude" 2. Like and retweet this 3. Follow me @codewithimanshu so I can DM you Save this post. Build the simulation system this week. Start with paper trading. Scale on evidence.

keyboard_arrow_left Previous keyboard_arrow_right Next

Himanshu Kumar @codewithimanshu

a month ago

A fired Jane Street quant walked out with 10 years of private BTC trading data. Turned it into $1.5M. He did not build a bot. He built a simulator that runs every move Bitcoin can make before it makes one. I found his wallet. Been copying him for a week. PnL prints like

188 146 316 63K 292

Also if you want to try building your own bot, here I wrote a step-by-step guide on how I personally did it.

AdiiX @adiix_official

a month ago

x.com/i/article/2064…

14 32 421 437K 2K

CHINESE GIRL WITH CLAUDE 5.0 JUST DROPPED THE FULL 31-MIN TRADING BOT BUILD GUIDE (Build Apps & Automations) bookmark it and watch when you've got 31 quiet minutes, you will forget what losing manual trades are forever.

AdiiX @adiix_official

a month ago

Claude 5.0 built a Chinese girl a trading bot. skip to 0:08 look at her journal, bot easily earns your monthly salary in a couple of days. how it works: the bot runs mean reversion on s&p 500 and nasdaq on 15-min candles, catching the small overextensions indices make every

193 369 6K 7.9M 18K

KARPATHY JUST HANDED EVERY DEVELOPER THE EXACT FILE CLAUDE CODE NEEDED FROM DAY ONE. 65 lines. 110K stars. the cheat code for every broken workflow you've been blaming on the model. if I had this a year ago, I would've shipped twice as fast. make sure to bookmark it before it gets lost in your feed. I was losing 2 hours a day to Claude rewriting code I didn't ask it to touch. then I found CLAUDE. md. 90 seconds to set up. changed everything. Karpathy identified 4 failure patterns Claude Code repeats constantly, in his own words: → silent assumptions: Claude makes decisions without checking with you → code bloat: 1000 lines written when 100 would do → collateral damage: Claude edits code unrelated to the task → no success criteria: Claude loops with no finish line these aren't model failures. they're missing instructions. CLAUDE. md gives Claude the 4 rules it needed from day one: → think before coding, state assumptions. ask before assuming. → simplicity first, minimum code. nothing speculative. → surgical changes, touch only what is required. nothing adjacent. → goal-driven execution, define success before starting. loop until verified. 65 lines. no build step. no framework. no dependencies. just the 4 principles every developer already knew, but needed Karpathy to write down. (Link to the REPO in the comment below) the guide on how to build a second brain with CLAUDE is in the article below.

Hamza Khalid @humzaakhalid

a month ago

x.com/i/article/2062…

22 66 469 1.3M 2K

A TEAM OF AI RESEARCHERS JUST OPEN-SOURCED THE BLOOMBERG TERMINAL FOR QUANT FINANCE. A Bloomberg Terminal costs $25,000 per year per seat. Banks pay for thousands of them. This thing reads every quant paper, every financial blog, every SEC filing, every arXiv preprint, and turns it into a searchable knowledge base. For free. It's called QuantMind. It just got accepted to the NeurIPS 2025 GenAI in Finance Workshop. Here's what it actually does: → Ingests arXiv quant papers, financial news, blogs, and reports automatically → Parses PDFs, HTML, tables, and figures into structured knowledge → Tags every paper by research area and topic → Builds a semantic knowledge graph you can query in plain English → Plugs into DeepResearch, RAG, and MCP for multi-hop reasoning → Two-stage architecture: extract once, retrieve forever Here's the wildest part: The financial research industry publishes around 500 new papers and reports every single day. Hedge funds pay six-figure salaries to junior analysts whose entire job is reading them. QuantMind reads all of it. Tags it. Embeds it. Lets you ask it questions. 154 stars. 22 forks. 173 commits. MIT license. Python. One honest note: this is a framework, not a magic alpha machine. You still need to know what to ask. But the "I haven't read that paper yet" excuse is officially dead. The thing Wall Street charges $25,000 a year for is sitting on GitHub. Free. Link in the comments.

Nobel Prize winner Demis Hassabis just accidentally revealed who survives the next 5 years and who doesn't. "One person who understands AI will outperform an entire startup team" Most founders heard that and thought: "Oh no, I need to learn prompt engineering" Wrong. That's not what "understands AI" means anymore. It means: building workflows. Chaining systems. Automating entire departments. Not typing better questions into ChatGPT. The split is brutal: > 90% of people = still using AI like a calculator > 10% of people = treating it like infrastructure In 5 years, the 10% will run everything with half the headcount. The 90%? Replaceable. Which group are you in? Watch the full breakdown. This is the only skill gap that actually matters right now. Bookmark this. You'll want to reference it.

rari @0xwhrrari

a month ago

x.com/i/article/2062…

28 95 668 2.1M 3K