wast3
@0xWast3 • 4,809 subscribers
18 / ai + prediction markets / bots that print
Shorts
Videos
Anthropic ex-engineer runs an internal graph that costs $6 a month and catches what a $300,000 eval suite misses. No retrieval layer. Seven nodes and one rule about who's allowed to change their mind. He published the whole schema. His version starts from the opposite idea. A graph is not an execution order. It's a memory of why. Seven nodes. Every edge carries the reason it exists: > INTENT - states what the task is for. Never how > DECOMPOSE - splits it into steps, each with a stated assumption > WORKER - executes one step. Sees nothing else > AUDIT - checks the output against the assumption, not the goal > DRIFT - compares the current step to INTENT and flags divergence > LEDGER - stores every decision with the assumption that justified it > ROOT - holds the graph, and when an assumption breaks, re-runs every step built on it Six nodes act. One node remembers why they acted. Every step carries the assumption that made it correct, so a false one only reruns what stood on it. That's the entire design. A pipeline that forgets its reasons has to redo all of it or trust all of it. He replayed a month of agent runs. 4,100 steps, 380 built on an assumption that was wrong by day three. The old pipeline shipped all 380 and linked none of them. Everyone else builds graphs where output moves forward and the reasoning evaporates. He built one where the reason travels with the result. The blast radius is the output nobody else produces. The article below is the full build - node prompts, the assumption format, the invalidation rule that finds every step downstream of a broken one. Save it. You'll want it open in the other
wast3304,987 Aufrufe • vor 11 Tagen
Engineer runs a Kimi K3 memory layer that costs $11 a month and remembers what a $500,000 vector database keeps losing. No embeddings. Four nodes and one rule about what's allowed to be forgotten. He published the whole schema. His version starts from the opposite idea. Memory is not a pile you search. It's a set of claims that expire unless something keeps paying to keep them. Four nodes. Every memory carries a clock someone has to reset: > WRITER - stores a fact with the reason it mattered, never raw text > DECAY - ages every memory down. Silence is deletion > RENEWER - only re-lifts a memory the model actually used again > GRAVE - holds what died, and why nobody reached for it Three nodes keep memory alive. One keeps the dead ones. Recall isn't storage here. It's rent a fact has to keep earning. That's the entire design. When everything is remembered forever, the useful and the stale retrieve identically. He replayed two months of agent context. 90,000 stored facts. 71,000 never retrieved once. The vector store returned all of them on similarity. Similarity graded closeness. Nobody graded whether the memory was ever right. Everyone else stuffs more into the context window and calls it memory. He built a layer that lets a fact die unless it keeps proving itself. The cost isn't storage. It's finding out how much of what your agent "knows" it has never once used. The article below is the full build - node prompts, the decay curve, the renewal rule. Save it. You'll want it open in the other tab.
wast364,039 Aufrufe • vor 7 Tagen
Anthropic developer runs an internal graph that costs $9 a month and catches what a $400,000 red-team contract signed off on. No judge model. Four nodes and one rule about who's allowed to be certain. He published the whole schema. His version starts from the opposite idea. A graph is not a chain of correct steps. It's a machine for manufacturing disagreement. Four nodes. Every edge carries a number someone has to defend: > WORKER - executes one step. Attaches a confidence, never a hedge > CHALLENGER - argues the output is wrong. Never fixes it > GATE - a narrow spread stops the run. Agreement is failure > SCAR - stores every confident answer that was wrong Three nodes produce. One keeps score of who deserves to be believed. Confidence isn't a feeling here. It's a balance you spend down. That's the entire design. When every node sounds sure, competence and noise look identical. He replayed six weeks of runs. 5,400 steps. 290 shipped at 90% confidence and wrong. All 290 passed the eval suite. The suite graded outputs. Nobody graded the grader. Everyone else routes work around disagreement. He built one that won't move until something argues back. The cost isn't compute. It's finding out which of your nodes has been bluffing since March. The article below is the full build - node prompts, the confidence format, the demotion rule. Save it. You'll want it open in the other tab.
wast362,588 Aufrufe • vor 9 Tagen
Kimi K3's weights are free to download. Almost nobody on earth can actually run them. 2.8 trillion parameters. The download is the easy part. Open weights used to mean you could self-host. At this scale it means you can read the file. So people download it and call the API anyway. Daily revenue up at least 6x since launch. Openness stopped being the opposite of a business model. It became the top of the funnel. The license says the quiet part. Past $20M in revenue you negotiate a contract. Past 100M users you display their name. Free for the developer. A contract for anyone the model actually earns from. And publishing the weights kills the argument before it starts. Nobody can claim the benchmarks were cooked. Independent indexes put K3 top three overall and first in frontend coding, with no one taking Moonshot's word for it. That's the whole trade. Give away the artifact, keep the infrastructure, get the audit for free. Check the last open model you praised. Ask whether you could serve it, or only download it.
wast362,375 Aufrufe • vor 20 Tagen
ANTHROPIC ENGINEER JUST LANDED A HUGE INVESTMENT FOR A MEMORY SYSTEM THAT NEVER FORGETS Most agents wake up blank every session, forgetting everything the moment the conversation ends. He built a scoring layer instead, one that decides what actually matters past this session and what gets dropped. Contradictions between old and new memory get flagged automatically, never silently overwritten the way most systems do. Memory that never gets touched again quietly decays on its own, so nothing bloats the system with dead weight. Nobody outside the deal knew what the scoring layer actually did before the check cleared. See how the scoring layer actually decides what survives below👇
wast374,869 Aufrufe • vor 1 Monat
HOLY SH*T, I GAVE GROK BOT MY BANKROLL AND WENT TO SLEEP $200 a month. I woke up to 22 positions I never approved, closed, and a note explaining every one. Six agents. Each one ate a line off the budget: > $19,000 news wire → gone > $21,000 odds feed → gone > $24,000 research subs → gone > $26,000 execution software → gone > $28,000 Bloomberg seat → gone > $34,000 the night trader → gone $152,000 a year, down to $2,400. 63 times cheaper, and it never blinks at 4am. This is the same shape of desk that prices an election night. Six agents, but not in a row. It's a graph, and only one node is allowed to think. One night of it, while you were asleep: 0:04 - SCANNER opens 3,180 live markets 0:11 - READER kills 1,140 as priced noise 0:17 - PRICER re-rates 902 of them 0:23 - PLANNER finds the moving resolution 0:29 - LOGGER writes why each was closed 0:35 - NOTIFIER wakes you. One decision. Nine hours of that. The clip above is 35 seconds of it. Anything that sizes or exits stops at approval first. The system only interrupts you to disagree with you. A working book files itself. A broken thesis wakes you. One trader watches nine markets. This watched three thousand and handed you one decision. Every trading floor was priced on reading being slow and conviction being human. Both stopped being true this year. Article below, six prompts and the full schema. Copy it, drop it into Grok, ask for the graph. Bookmark it before you start - you'll be scrolling back to it all night.
wast318,571 Aufrufe • vor 11 Tagen
Roblox paid me $4,200? you're still thinking it's a kids game $1,000,000,000 paid out to creators in 2025 most adults didn't take a single dollar that changes now here's what's actually happening: > 88M daily users, avg session 2.4 hours > 1 game/week × 4 weeks = catalog that compounds > 3% conversion × 5,000 users = $15,750/month > Claude writes the Lua, you just ship reviews compound before you ship the next niches are still open - not for long bookmarked and learn
wast3126,574 Aufrufe • vor 3 Monaten
Elon Musk's agents just replaced a $394,000 build system. $199 a month, and it found the flaky test that had been lying to the team for eleven months. He open-sourced the entire node map. Every line the old system billed, and where it went: > $96,000 CI minutes → dead > $44,000 device farm → dead > $38,000 observability seats → dead > $71,000 the build cop → dead > $145,000 the infra lead → dead From $394,000 to $2,400. 164x, and nothing gets frozen on a Friday anymore. This is the shape of pipeline that ships a frontier lab's models. Seven agents, but the wiring is the trick - six work, one decides. Thirty-eight seconds of a nine-hour shift: 0:03 - SCANNER pulls 2,400 test runs 0:08 - ANALYZER reads every failure log 0:14 - VALIDATOR kills 810 false reds 0:20 - EXECUTOR reruns the real ones 0:26 - PLANNER finds the flaky root 0:31 - LOGGER writes why each passed 0:38 - NOTIFIER wakes you. One test. Everything above happened before your alarm did. PLANNER is the only node with the map of which test guards which module. One flake at the root and it re-runs every branch hanging off it. Silence means agreement. Green never reaches you. A test that quietly started lying gets you out of bed. Twelve hundred failures went in. One came out with your name on it. Nobody bought this stack because it worked. They bought it because red used to be ambiguous, and asking a human was the only way to find out. Full build in the article below - seven prompts, node map, the disagreement rule that keeps NOTIFIER quiet. Open it in the other tab before you start wiring.
wast316,423 Aufrufe • vor 12 Tagen
A developer in Hangzhou runs an AI that remembers everything about him for $0.40 a year. No vector database. One file that never grows past 4,000 tokens. He published the whole schema. His version starts from the opposite idea. Memory is not storage. It's a write policy. Six fields. Rewritten every time, never appended: > IDENTITY - who you are, what you build. 300 tokens. Changes monthly at most > STATE - what you're on right now. 400 tokens. Rewritten daily > DECISIONS - what's already settled, so nothing gets re-argued. 800 tokens > CORRECTIONS - every time you said "no, not like that." 600 tokens > PEOPLE - names, roles, who's waiting on what. 500 tokens > DEAD - tried and abandoned, so it never comes back as a suggestion. 400 tokens Three thousand tokens. Ceiling of four. When a section fills, the model rewrites it shorter. Nothing is ever added. Only replaced. Kimi K2.5 bills $0.10 per million cached input tokens. Four thousand tokens a turn is $0.0004. That's 2,500 turns for a dollar. The free tier hands you 1.5 million tokens a day. 375 turns before you pay anything at all. CORRECTIONS is the field nobody builds, and it's the one that does the work. A model that remembers being wrong stops repeating it. Everyone else is paying to search their own history. He pays to keep it short. The bill stopped growing when the file did. Your memory system isn't defined by what it stores. It's defined by what it agrees to delete. The article below is the full build - schema, rewrite prompts, the compaction rule that keeps it under the cap. Save it. You'll want it open in the other tab.
wast315,862 Aufrufe • vor 13 Tagen
A FORMER GOOGLE LEAD BUILT A MEMORY SYSTEM THAT NEVER LOSES A SINGLE DETAIL Most agents forget everything the moment a session ends, starting from zero every single time. He built a scoring layer instead, one that decides what actually matters past this session and what gets dropped. Contradictions between old and new memory get flagged automatically, never silently overwritten the way most systems do. Memory that never gets touched again quietly decays on its own, so nothing bloats the system with dead weight. He left one of the biggest labs in the world to build the thing they said couldn't be solved yet. See how the scoring layer actually decides what survives below👇
wast326,649 Aufrufe • vor 1 Monat
A million-token window solved the wrong problem, and Kimi K3 shipping one made that harder to see. 2.8T parameters, weights on your disk, an entire codebase in one prompt. Feels like memory. Isn't. Memory is what survives the session ending. A window is what survives the next token. Self-hosting sharpens it. Now you pay in GPU seconds, not API credits. Every re-read of the same history burns the same compute. Owning the weights didn't make the past free. Kimi Delta Attention cuts what long context costs. Cheaper repetition is still repetition. And a bigger window buys the one thing you don't want: room to keep everything. Nothing in a transcript is dated, ranked, or retired. Old and current load at identical weight. The fix runs beside the model, not inside it. Capture rejects, consolidation merges, decay forgets on schedule. Then a million tokens becomes headroom instead of a bill. You fill 5,000 and leave the rest empty. Look at what your longest thread reloads every run. Then ask how much of it earned the slot.
wast317,620 Aufrufe • vor 19 Tagen
ONE PROMPT AND ONE BUTTON NOW BUILDS A FULL SYSTEM, DATABASE, BACKEND, AND FRONTEND INCLUDED Most people still picture AI as something you ask questions to, then copy the answer into your own code by hand. A builder wired Claude into Obsidian through MCP instead, so a single written prompt becomes the actual spec for the system. Claude breaks that prompt into database schema, UI components, and API endpoints on its own, then writes, runs, and fixes the code until it actually works. MCP is what makes this possible, it lets the agent see the filesystem, installed packages, and running services before it starts building anything. The output isn't a code snippet sitting in a chat window, it's a live local app running on its own server, ready to open in a browser. What used to need a backend developer, a DevOps engineer, and two weeks now finishes in under fifteen minutes. See the full architecture and prompt structure below👇
wast320,348 Aufrufe • vor 1 Monat
My 14-year-old friend bought a $480,000 apartment in central London dropped out of school six months ago while his classmates were studying for exams he was watching Roblox tutorials on YouTube at 2am started with free guides, then bought AI courses then stopped buying courses and started selling games here's what six months looks like: > first map published at week 3 > gamepass conversion 3% × 5,000 users = $15,750/month > scaled to 6 games, catalog compounding every month > Claude writes every line of Lua, he just ships Roblox paid $1,000,000,000 to creators last year he took his share at 14 the apartment is real, the system is simple the only difference is he started bookmarked and learn
wast332,144 Aufrufe • vor 3 Monaten
A 68-YEAR-OLD RETIREE BUILT AN AI AGENT THAT RUNS ENTIRELY WITHOUT THE INTERNET He had spent 35 years as a systems engineer, but never touched modern AI tools until his daughter showed him Claude last winter He wanted something that could manage his daily tasks - reminders, document drafts, research summaries - but refused to trust his personal data to cloud servers So he bought a GMKtec EVO-X2, wiped Windows, installed Ubuntu, pulled Ollama, and pointed a local Claude Code instance at his own machine But the real breakthrough came when he wired the agent to a small Raspberry Pi running automations around his house - calendar alerts, morning briefings, document organization, all triggered by voice No subscription. No data leaving the house. No internet required after the initial setup While most people his age were struggling with basic smartphone features, he had built a fully offline personal AI assistant running on hardware that fits in a shoebox He posted a short demo video showing the agent summarizing his weekly notes and drafting three emails while completely offline His inbox filled with messages from other retirees asking how to build the same thing
wast319,850 Aufrufe • vor 2 Monaten
A $1,800 MINI PC RUNS CLAUDE CODE WITHOUT THE INTERNET it looks like a hardcover book sitting on a shelf, but this is the GMKtec EVO-X2, and it changes the local AI math completely built on AMD's Ryzen AI Max+ 395, with 128GB of unified memory shared between CPU and GPU that shared pool is the entire trick - the model loads once, both chips read from the same address space wipe Windows, install Ubuntu, pull Ollama, point Claude Code at the local endpoint same CLI, same agent loop, nothing leaves your network Qwen3 Coder 30B and Llama 3.3 70B run comfortably, fully usable for daily development work Claude Code Max + ChatGPT Pro = $4,800/year leaving your account the box costs $1,800, electricity runs $9/month break even hits at month five after that you stop rationing tokens - overnight agents, full codebase audits, experiments you'd never start when the meter was running won't replace frontier reasoning for the hardest 5% of problems for everything else, the local stack quietly becomes the default
wast313,367 Aufrufe • vor 2 Monaten
$2,160,000 from a village with no internet he learned to trade weather markets before he had stable wifi while professional meteorologists were collecting salaries he was beating them with a bot and open-source data This’s his wallet: grew up with nothing, figured out everything here's what the numbers look like: > 2,711 predictions since January 2024 > 73% winrate, #18 overall on Polymarket > $288,770 pure profit from weather markets alone > biggest single win: $12,100,000 he doesn't have a Bloomberg terminal he doesn't have a meteorology degree he has open weather APIs, a systematic edge, and zero emotions $2.16M net is what happens when a kid from nowhere understands probability better than the crowd bookmarked and learn bookmarked and learn
wast312,965 Aufrufe • vor 3 Monaten
Brands are paying $500,000 for a single Fortnite map not Epic, not players - Fortune 500 companies while creators were chasing $2,190/month from the pool agencies were writing checks with six zeros here's what's actually happening: > Nike, NFL, Samsung already have branded Fortnite islands > one corporate map contract = years of creator pool income > Claude builds the concept deck, Verse code, and pitch in hours > you sell the map once and keep the creator pool revenue forever corporations need UEFN developers right now there are almost none who can pitch and build simultaneously AI closed that gap completely the pool pays monthly, the contracts pay instantly two income streams, one pipeline bookmarked and learn
wast311,175 Aufrufe • vor 3 Monaten
Keine weiteren Inhalte verfügbar