Kawaki Smith

🔴 TRUCO: Convierte CUALQUIER repositorio en un prompt para COPIARLO. Si eres fan de la INGENIERÍA INVERSA, esto te interesa:

Anthropic published the official guide to building a knowledge graph with Claude You build one when the answer lives in no single document. Who works with the people who worked on project X. Which vendors touch this incident. Retrieval gives you the closest chunks and the chain is in none of them. This used to mean training an entity recognizer on your domain, training a relation classifier, and writing merge rules by hand. Three things to maintain, all three going stale as your data shifts. In the guide each one is a prompt. The cheap model does extraction. You describe the output shape once and every document comes back as typed entities and triples that fit it. It also writes one sentence about each entity, which looks like a throwaway until the next step. Then the smart model merges duplicates, and this is where the old stack failed. Edit distance catches NASA against National Aeronautics and Space Administration. It does nothing for Edwin Aldrin and Buzz Aldrin, which share no characters and are the same man. The model reads those one line descriptions and merges on meaning. Queries run the other way. Take the neighborhood around one node, hand Claude the raw triples, and every claim in the answer points at the edge it came from. It ships with a gold set and a scorer, so you change one line of the extraction prompt, rerun, and watch precision move. Full breakdown in the article below Bookmark this

🚨 ANTHROPIC JUST KILLED THE DEMO AGENT ERA. Their Agents team showed exactly what production grade looks like. Not theory. Not a tutorial. A four layer framework for multi agent systems built to actually work in the real world. 30 minutes. This is the video I wish existed 6 months ago.

AN ENGINEER SOLVED A PROBLEM NOBODY IN THE COMPANY COULD FULLY MAP ON THEIR OWN The problem wasn't a missing tool, it was that every process touched three other processes nobody had written down. He wired dozens of agents into one shared graph instead of documenting each department's workflow by hand. Each agent owned one process, reading its own logs while writing every dependency it found straight into the shared structure. Within days, the graph surfaced a connection between two departments that had been quietly blocking each other for over a year. The company adopted the graph as its live map of operations, and it hasn't stopped finding new connections since. See what the graph found in the first 48 hours below👇

We cut our AI bill by 30%. Cheaper Inference sells access to most models cheaper than the list price. Same models. Same API format. Smaller bill. You can use OpenAI, Anthropic, and Google models through one OpenAI-compatible API. Just switch your endpoint

50 amazing websites that look ‘illegal’ but are not. 1. cobalt.tools — Download any social media video 2. photopea.com — Free Photoshop 3. temp-mail.org — Temporary email with one click 4. tinywow.com — Over 100 free tools in one site 5. archive.org — Access any old webpage 6. libgen.li — Millions of free textbooks 7. sci-hub.al — Free research articles 8. alternativeto.net — Find free alternatives to apps 9. justwatch.com — Locate where to stream any content 10. gutenberg.org — 70,000 free classic books 11. pdfdrive.com — Free PDF downloads 12. openculture.com — Free courses from elite universities 13. wolframalpha.com — Solve any math problem instantly 14. remove.bg — Remove background with one click 15. cleanup.pictures — Erase objects from photos 16. unscreen.com — Remove video backgrounds for free 17. squoosh.app — Compress any image for free 18. excalidraw.com — Free hand-drawn graphics 19. carbon.now.sh — Turn code into artwork 20. ray.so — Impressive code screenshots 21. flightradar24.com — Track any flight in real time 22. camelcamelcamel.com — Track Amazon price history 23. haveibeenpwned.com — Check if you've been hacked 24. virustotal.com — Scan malware in any file 25. privnote.com — Send self-destructing messages 26. file.io — Share files that auto-delete 27. archive.ph — Archive any webpage forever 28. accountkiller.com — Erase your presence from any website 29. radio.garden — Listen to any radio station in the world 30. tunefind.com — Identify songs from any show 31. musicforprogramming.net — Music to focus 32. mynoise.net — Custom soundscapes to focus 33. annasarchive.org — Search every book ever written 34. elicit.com — AI assistant for research articles 35. consensus.app — Search scientific consensus 36. connectedpapers.com — Visually map research 37. semanticscholar.org — Free academic search 38. scispace.com — Understand any research article 39. summarize.tech — Summarize any YouTube video 40. phind.com — AI search for developers 41. regex101.com — Test any regular expression instantly 42. codebeautify.org — Format any code clearly 43. explainshell.com — Understand terminal commands 44. tldraw.com — Infinite whiteboard in the browser 45. downdetector.com — Check if a website is down 46. tineye.com — Reverse image search 47. fast.com — Test your internet speed 48. smallpdf.com — Edit PDFs for free 49. ilovepdf.com — Merge and split PDFs 50. 10minutemail.com — Temporary email in seconds All free. Save it before you forget.🔖

If you want to become world-class with LLMs, read these 7 books (save now):

The biggest opportunities right now: 1. build for solving loneliness (the more AI floods everything, the more people crave real human connection, IRL and small social) 2. build for agents that need to spend money (they're getting virtual cards and budgets, someone builds the spend controls, fraud protection, receipts) 3. build for people drowning in AI output (everyone generates infinite drafts now, the bottleneck moved to reviewing and choosing, build the judgment layer) 4. build for the burnout economy (everyone is expected to always be on and always optimizing, and the backlash toward rest, slowness, and enough is building) 5. build for verifying humans (deepfakes broke trust, every dating app, marketplace, and video call needs proof-of-human within 2 years) 6. build for the physical world (the trades, hardware, robots that AI is finally reaching) 7. build for the agent that answers the phone (every local business misses calls after 5pm, a voice agent that books the job is worth thousands a month) 8. build for the aging (70M+ boomers who want to stay healthy, sharp, and connected) 9. build for the LLM-search land grab (being the cited answer is the new SEO) 10. build for the newly automated (the paralegal, the analyst, the marketer whose job just changed and needs to reskill fast) 11. build for the seat-pricing collapse (software repricing from $50/seat to per-outcome, whoever nails outcome billing wins a category) 12. build for AI enablement (95% of businesses use nothing beyond ChatGPT, someone has to onboard the other 95%) 13. build for the agency everyone resents (businesses pay $1k/mo to agencies they hate, an agent that does 80% of it undercuts the model) 14. build for verticals on 2011 software (dentists, HOAs, contractors, all overdue for an AI-native rebuild) 15. build for reviving dead software (thousands of abandoned apps with real users, agents can maintain what a team couldn't, buy and revive) 16. build for markets too small to matter before (500 lobster fishermen was never worth a team, now it's a weekend and a real business) 17. build for agents hiring agents (a shadow economy is forming, it needs escrow, reputation, and dispute resolution for machines) 18. build for the anti-AI premium (as everything gets generated, human-made and analog become status symbols people pay up for) 19. build for distribution-first (anyone can build the product now, so the audience is the moat, media company first, product second) 20. build for the reinvention of college (what does an MBA even mean anymore) 21. build for a world with more free time than it knows what to do with (if AI takes the busywork, the question becomes what people do with the hours, and that's a civilization sized market) note: more trends/ideas @ideabrowser (free to sign up) 22. build for the return to the physical (screens fill with slop, people crave the real world, the hands-on, the local, the analog) 23. build for the caregiving wave. The population is aging fast, tens of millions are caring for parents, and the whole burden is landing on families with no support. 24. build for the longevity shift (people want to live to 100 healthy, and a whole industry is forming around actively managing your own biology) 25. build for the housing and rootlessness problem (people can't afford to settle down, and the whole idea of a stable home base is up for grabs) 26. build for spiritual hunger (as institutions hollow out, the need for meaning, ritual, and belonging is exploding into new forms) KEEP BUILDING

🤯 YOUR AGENT NOW READS THE ENTIRE INTERNET WITHOUT A SINGLE API KEY someone open-sourced this repo. it replaces $300/mo in paid APIs Agent Reach. #1 Repository of the Day on GitHub trending. 63,165 stars install it - and your agent reads: - any web page → clean markdown - YouTube → captions and search, plus 1800+ video sites - GitHub → repos, issues, search - X → timeline, search, full posts - Reddit → threads and comments - LinkedIn, Instagram, Facebook - RSS and semantic search across the whole web no keys. no subscriptions. important: your cookies stay on your machine. they never leave your agent has been guessing about the internet this whole time. save this before your next research task ↓

Karpathy didn't just "hook AI up to a vault." He published a pattern: notes aren't an archive - they're a living wiki that a neural net writes and maintains. Most Obsidian vaults are graveyards. Hundreds of notes, zero links, zero return visits. Karpathy's method breaks that. Three commands, that's it: 1. Ingest. Drop in an article, a podcast, a PDF - Claude splits it into atomic notes and wires the wikilinks between them itself. 2. Query. Ask a question - get an answer assembled from your own notes, with exact citations to where it came from. 3. Lint. Once a week, Claude walks the entire vault: finds contradictions, kills stale claims, and pulls orphan notes back into the graph. No vector database. Just markdown files that the AI reads and rewrites directly. Karpathy's own vault grew to ~100 articles and 400,000 words this way not in a week, but over months of discipline. His advice: don't chase scale on day one. Start with ten sources, get ingest/query/lint to feel automatic - then expand. The gap between people who accumulate knowledge and people who lose it isn't talent. It's the habit of keeping the graph alive.

@Love_Cinemaa They r cooking for it, wait till next elections.

I want that vintage single seat Janasena back

@KTRBRS Uneducated kii politicians background verification cheyali whoever doing political entries n scrape the all

“Criminal waste” is what the cheap minister says about our engineering Graduates’ abilities What are his qualification to be saying so? Did he ever write any competitive exam in his entire life? Did he ever appear for an interview for a job? Did he ever work anywhere In a professional environment? For those who don’t know, he was a real estate broker turned blackmailer who hustled his way to buying the CM post in bankrupt Congress party

Don't prompt: "make the UI look good" > Install this: npx skills add anthropics/claude-code --skill frontend-design Don't prompt: "why is this code broken?" > Install this: npx skills add obra/superpowers --skill systematic-debugging Don't prompt: "read this PDF and make me a spreadsheet" > Add the marketplace: /plugin marketplace add anthropics/skills > Then install: /plugin install document-skills@anthropic-agent-skills Don't prompt: "is my app secure?" > Run this: /security-review Don't prompt: "make my site load faster" > Install this: npx skills add addyosmani/web-quality-skills Don't prompt: "remember what we did last session" > Add the marketplace: /plugin marketplace add anthropics/skills > Then install: /plugin install document-skills@anthropic-agent-skills Don't prompt: "review my code" > Add the marketplace: /plugin marketplace add anthropics/claude-plugins-official > Then install: /plugin install code-review@claude-plugins-official Don't prompt: "this got too complicated, clean it up" > Install this: npx skills add anthropics/claude-code --skill simplify Don't prompt: "build me an interactive demo" > Install this: npx skills add anthropics/claude-code --skill artifacts-builder Bookmark this

Types of U.I design you should know 🤯

@TeluguScribe Politicians r illiterates they don't know what they are talking 🤡🤡🤡🤡🤡🤡just running circus shows.

ఇంజనీరింగ్ స్టూడెంట్స్ క్రిమినల్ వేస్ట్ వారికి కనీసం జాబ్ అప్లికేషన్ నింపడం కూడా రాదు సంవత్సరానికి లక్షా 10 వేల మంది ఇంజనీరింగ్ గ్రాడ్యుయేట్ సర్టిఫికేట్ తీసుకుంటున్నారు - సీఎం రేవంత్ రెడ్డి

@RVarsR @BJP4Telangana Good visualization

@BJP4Telangana Gallo lekkalu bale estharu ra miru

E20 వల్ల లాభం ఉంటే పెట్రోల్ రేట్లు ఎందుకు తగ్గలేదు? దీని వెనుక ఉన్న అసలు వాస్తవం ఇది! ప్రపంచవ్యాప్తంగా ముడి చమురు ధరలు భగ్గుమంటున్నా, ఆ కల్లోలం మన దేశ ప్రజల మీద పడకుండా మోదీ ప్రభుత్వం E20 బ్లెండింగ్ ద్వారా అడ్డుకట్ట వేసింది. 2022-26 మధ్య అంతర్జాతీయంగా పెట్రోల్ ధరలు 20% నుండి 40% వరకు పెరిగితే, మన దగ్గర కేవలం 5.6% మాత్రమే పెరిగాయి. దీనివల్ల ఏకంగా ₹1.90 లక్షల కోట్ల విదేశీ మారకద్రవ్యం దేశానికి ఆదా అయ్యింది. ఇదీ శ్రీ నరేంద్ర మోదీ గారి ముందుచూపు!

🚨EL WEB SCRAPING HA MUERTO🚨 Han creado PixelRAG. Un sistema open source que pasa de parsear HTML. En vez de convertir una web en texto... hace una captura de pantalla. Y luego un modelo vision-language lee la respuesta directamente desde los píxeles. Brutal. Porque el scraping tradicional rompe muchísima información: → tablas → gráficos → layouts → columnas → botones → PDFs → páginas mal estructuradas Todo eso se pierde cuando conviertes una web en texto plano. PixelRAG hace algo mucho más simple: indexa la página tal y como la ve una persona. El equipo ha creado un índice visual de Wikipedia entero: +30 millones de screenshots. Y aun así supera al mejor RAG basado en texto por un 18,1% en QA. Tiene también un plugin para Claude Code. Básicamente le da ojos a Claude. Le pasas una URL, un paper de arXiv o tu web local… y Claude no scrapea el DOM. La mira. Cómo funciona: → renderiza webs, PDFs e imágenes en tiles → los embebe con Qwen3-VL-Embedding → construye un índice FAISS → busca visualmente sobre píxeles Todo open source bajo Apache-2.0. Link al repo en los comentarios 👇🏻

@NeeshamX Andhaaaa bhakths r crying in comments 😂😂😂

@6Item36813 They r pretending they r speaking language, as they looting money n bluffing the people with random attention seeking words always want to be headlines.

Anna ki rendu lev😭🙏🏻