Loading video...

Video Failed to Load

Go Home

A 31-year-old blitz chess master who never cracked the top 100 just turned $500 into $153,279 - with Claude and simulation engine. Twenty years reading a board faster than the other guy, for prize pools that barely covered the flights. Then he pointed that speed at Polymarket. His wallet:...

antpalkin's profile picture

cvxv666

10,018 subscribers

39,843 views • 4 months ago •via X (Twitter)

LIVE

Anya Rossi• Live Now

Private livecam show

0 Comments

No comments available

Comments from the original post will appear here

Related Videos

Jane Street quant just stole and fit a $40-million trading brain inside a box that costs $4,000 and quietly walked out of the building with it. It's the new AMD Ryzen AI Halo. Eleven weeks later, that computer has made him $400,000 - from his bedroom. His wallet: The model doesn't guess where Bitcoin goes. It simulates it. Every five minutes it reads the candle into a pattern net - break of structure, order blocks, the gaps where liquidity hides. Then it hands that read to MiroFish: a simulation that runs the next five minutes forward thousands of times, sorts every path into three outcomes - bear, neutral, breakout and threads the median straight through them. That median is where price actually lands. Not where the crowd hopes. When that fair value drifts 15¢ off the live book, the edge is real. It sizes by Kelly and fires. One button, no nerves. On the firm's cluster, one simulation like that ate serious hardware. The Halo box - 128GB of unified memory, a 200B model held whole, runs thousands a minute, offline, while he sleeps. No cloud. No login the firm can kill. The desk still owns the model. It just can't aim it where he aims it - Polymarket's too small for a fund that size to touch. He took the edge home. Two clicks and you trade right beside it:
0:37

Jane Street quant just stole and fit a $40-million trading brain inside a box that costs $4,000 and quietly walked out of the building with it. It's the new AMD Ryzen AI Halo. Eleven weeks later, that computer has made him $400,000 - from his bedroom. His wallet: The model doesn't guess where Bitcoin goes. It simulates it. Every five minutes it reads the candle into a pattern net - break of structure, order blocks, the gaps where liquidity hides. Then it hands that read to MiroFish: a simulation that runs the next five minutes forward thousands of times, sorts every path into three outcomes - bear, neutral, breakout and threads the median straight through them. That median is where price actually lands. Not where the crowd hopes. When that fair value drifts 15¢ off the live book, the edge is real. It sizes by Kelly and fires. One button, no nerves. On the firm's cluster, one simulation like that ate serious hardware. The Halo box - 128GB of unified memory, a 200B model held whole, runs thousands a minute, offline, while he sleeps. No cloud. No login the firm can kill. The desk still owns the model. It just can't aim it where he aims it - Polymarket's too small for a fund that size to touch. He took the edge home. Two clicks and you trade right beside it:

cvxv666

94,183 views • 3 months ago

I gave Claude one Polymarket equation and told it to find every mispriced contract It hasn't stopped running in 4 months $600 → $14,190 Every Polymarket contract is priced by one formula: C(q) = b × ln(Σ e^(qi / b)) The pricing function inside it is softmax - the same math Claude uses to pick its next token Claude doesn't just understand this equation, it thinks in it So I gave it a task: "Scan every open contract, compute fair value, flag anything the market has wrong 24/7" It finds the gap, sizes the position by Kelly, enters before the crowd corrects Market shows 0.38, math shows 0.57, bot buys Resolution hits, bot collects 120 days 61.2% win rate 2.31 sharpe -3.8% max drawdown 425 trades 87% of Polymarket wallets are underwater They trade feelings against a formula that doesn't bluff, doesn't tilt, doesn't chase losses The edge is having something that reads the math faster than any human - and never stops looking
0:28

I gave Claude one Polymarket equation and told it to find every mispriced contract It hasn't stopped running in 4 months $600 → $14,190 Every Polymarket contract is priced by one formula: C(q) = b × ln(Σ e^(qi / b)) The pricing function inside it is softmax - the same math Claude uses to pick its next token Claude doesn't just understand this equation, it thinks in it So I gave it a task: "Scan every open contract, compute fair value, flag anything the market has wrong 24/7" It finds the gap, sizes the position by Kelly, enters before the crowd corrects Market shows 0.38, math shows 0.57, bot buys Resolution hits, bot collects 120 days 61.2% win rate 2.31 sharpe -3.8% max drawdown 425 trades 87% of Polymarket wallets are underwater They trade feelings against a formula that doesn't bluff, doesn't tilt, doesn't chase losses The edge is having something that reads the math faster than any human - and never stops looking

rari

311,933 views • 6 months ago

An engineer at a Chicago HFT shop spent six years building latency arb pipes between CME and NYSE. On May 1st his desk got cut. Severance: $40K. He deposited $3000, ran a Hermes trading agent on Polymarket, and pulled in $236,913 over the next 23 days. His agent wallet: The pit taught him one thing - find where one venue knows something another doesn't, size to the gap, exit before the spread closes. So he made 757 trades a day. Not 7. Not 75. And let the rest expire. Here's the actual stack. Claude Opus 4.7 reads spot momentum off Chainlink, scores every 5-minute BTC market by Markov persistence × Kelly edge Then surfaces the windows where Polymarket hasn't priced in what Binance and Coinbase already confirmed. Hermes Agent by NousResearch executes. A $10/mo Hetzner VPS runs it 24/7. Telegram pings on every fill. Total cost: $10/month. Setup: 30 minutes. No coding. One trade on May 14: Bitcoin Down at 9AM ET. Market said 15.8¢. He put $1,681 on Down. It hit. +$8,974. A 533% return in 47 minutes. The real edge is the nightly self-learning loop. Every midnight, Opus reads the day's trade journal and rewrites MIN_PROB and MIN_EDGE in the .env file. May 2nd the threshold was 0.87. May 12th 0.89. May 23rd 0.91. The bot tightens with the regime. His version of the bot has rewritten itself 187 times in 23 days. The desk that cut him couldn't ban him from Polymarket. The bot doesn't sit in the colo. It doesn't need a leased line. It just reads two endpoints and presses a button. Save this if you want to dig in and understand Hermes. Or just copy this guy trades using TG bot - his algorithm has been perfected 187 times:
0:56

An engineer at a Chicago HFT shop spent six years building latency arb pipes between CME and NYSE. On May 1st his desk got cut. Severance: $40K. He deposited $3000, ran a Hermes trading agent on Polymarket, and pulled in $236,913 over the next 23 days. His agent wallet: The pit taught him one thing - find where one venue knows something another doesn't, size to the gap, exit before the spread closes. So he made 757 trades a day. Not 7. Not 75. And let the rest expire. Here's the actual stack. Claude Opus 4.7 reads spot momentum off Chainlink, scores every 5-minute BTC market by Markov persistence × Kelly edge Then surfaces the windows where Polymarket hasn't priced in what Binance and Coinbase already confirmed. Hermes Agent by NousResearch executes. A $10/mo Hetzner VPS runs it 24/7. Telegram pings on every fill. Total cost: $10/month. Setup: 30 minutes. No coding. One trade on May 14: Bitcoin Down at 9AM ET. Market said 15.8¢. He put $1,681 on Down. It hit. +$8,974. A 533% return in 47 minutes. The real edge is the nightly self-learning loop. Every midnight, Opus reads the day's trade journal and rewrites MIN_PROB and MIN_EDGE in the .env file. May 2nd the threshold was 0.87. May 12th 0.89. May 23rd 0.91. The bot tightens with the regime. His version of the bot has rewritten itself 187 times in 23 days. The desk that cut him couldn't ban him from Polymarket. The bot doesn't sit in the colo. It doesn't need a leased line. It just reads two endpoints and presses a button. Save this if you want to dig in and understand Hermes. Or just copy this guy trades using TG bot - his algorithm has been perfected 187 times:

cvxv666

38,535 views • 4 months ago

A quantum physicist lost his funding last year. He's since pulled $487,000 off Polymarket - with the math his grant was supposed to pay for. He doesn't predict where Bitcoin goes. He reads whether it's pinned in place right now - or free to run. His profile: The lab went dark. He went home with a decade of math and nothing to point it at. Here's the system. He feeds the last few hundred 5-minute BTC prints into TimesFM - Google's foundation model for time series. It hands back a distribution of where price goes next, using only what's known up to this second. No peeking at tomorrow. Then MiroFish turns that into a map. Every recent moment becomes a point on a surface. Red where price is being pulled back to equilibrium - something's holding it. Blue where it's broken free and the path can run. The points cluster: > Red pools into regions that act like a containment field > Blue pools where the structure let go The flip between them isn't gradual. It's a phase transition. That's the whole edge. It answers the only question a 5-minute candle asks: mean reversion, or momentum? He's not guessing which regime is next - he's reading which one is already live. Red, he fades the move. Blue, he rides it. Finding order in noise was his whole career. The market just pays for it directly. The bot's live right now, reading whether this candle is held or running. Two clicks and its trades land in your wallet too:
0:41

A quantum physicist lost his funding last year. He's since pulled $487,000 off Polymarket - with the math his grant was supposed to pay for. He doesn't predict where Bitcoin goes. He reads whether it's pinned in place right now - or free to run. His profile: The lab went dark. He went home with a decade of math and nothing to point it at. Here's the system. He feeds the last few hundred 5-minute BTC prints into TimesFM - Google's foundation model for time series. It hands back a distribution of where price goes next, using only what's known up to this second. No peeking at tomorrow. Then MiroFish turns that into a map. Every recent moment becomes a point on a surface. Red where price is being pulled back to equilibrium - something's holding it. Blue where it's broken free and the path can run. The points cluster: > Red pools into regions that act like a containment field > Blue pools where the structure let go The flip between them isn't gradual. It's a phase transition. That's the whole edge. It answers the only question a 5-minute candle asks: mean reversion, or momentum? He's not guessing which regime is next - he's reading which one is already live. Red, he fades the move. Blue, he rides it. Finding order in noise was his whole career. The market just pays for it directly. The bot's live right now, reading whether this candle is held or running. Two clicks and its trades land in your wallet too:

cvxv666

36,161 views • 3 months ago

Fired Jane Street quant walked out with ten years of private BTC trading data - and turned it into $1.5M He didn't build a bot. He built a simulator that sees the matrix - it runs every move Bitcoin can make before it makes one. His wallet: Here's what he actually built. A swarm of agents feeds ten years of stolen tick data into MiroFish - a god-tier agentic simulator. It doesn't 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, re-run the sim, and remember every pattern, every reaction, every signal they've ever seen. He doesn't predict the future. The math already knows it. He just reads the numbers and takes the money. I've been copying him for a week and PnL prints like a money printer on steroids. Here's the part the firms don't 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. Save this and build your own - or skip the homework and copy his with this bot:
0:43

Fired Jane Street quant walked out with ten years of private BTC trading data - and turned it into $1.5M He didn't build a bot. He built a simulator that sees the matrix - it runs every move Bitcoin can make before it makes one. His wallet: Here's what he actually built. A swarm of agents feeds ten years of stolen tick data into MiroFish - a god-tier agentic simulator. It doesn't 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, re-run the sim, and remember every pattern, every reaction, every signal they've ever seen. He doesn't predict the future. The math already knows it. He just reads the numbers and takes the money. I've been copying him for a week and PnL prints like a money printer on steroids. Here's the part the firms don't 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. Save this and build your own - or skip the homework and copy his with this bot:

cvxv666

132,166 views • 3 months ago

Someone just posted the full blueprint for an AI swarm that does the job of a 200-person quant research team. Six agents. Running 24/7. Finding brand-new alpha while you sleep. Citadel needs 100 PhDs to do this. Two Sigma needs 200. This does it with six bots and one laptop. Two ways to play this - spend a weekend building your own swarm, or copy the wallet of one that's already up $2M: Boris Cherny runs Claude Code at Anthropic. Two weeks ago he said: "I don't prompt Claude anymore. I have loops running that prompt Claude. My job is to write loops" Alpha research is just a pipeline. So instead of sitting in it, you hand each stage to its own agent: > one reads every new research paper overnight and pulls out the trade idea > one builds the features and cleans the data > one backtests it over 20 years, costs and slippage included > one runs the hard stats and kills anything overfit > one checks it still works in every market regime > one strips out plain momentum and value to see if any real edge is left Each of those six is a job a fund pays a $600,000-a-year quant to do. He runs all six for the price of an API bill. The rule that makes it work: the agent that builds a signal never gets to approve it. A separate, stronger agent tries to kill it first. Whatever survives all six by morning is real, new alpha. One trader's already running this exact swarm on Polymarket. That $2M wallet is public, every trade on-chain. The full build is in the post below - six agents, the tool that runs them, and the five mistakes that kill most people. Bookmark & read this before it's buried.
0:41

Someone just posted the full blueprint for an AI swarm that does the job of a 200-person quant research team. Six agents. Running 24/7. Finding brand-new alpha while you sleep. Citadel needs 100 PhDs to do this. Two Sigma needs 200. This does it with six bots and one laptop. Two ways to play this - spend a weekend building your own swarm, or copy the wallet of one that's already up $2M: Boris Cherny runs Claude Code at Anthropic. Two weeks ago he said: "I don't prompt Claude anymore. I have loops running that prompt Claude. My job is to write loops" Alpha research is just a pipeline. So instead of sitting in it, you hand each stage to its own agent: > one reads every new research paper overnight and pulls out the trade idea > one builds the features and cleans the data > one backtests it over 20 years, costs and slippage included > one runs the hard stats and kills anything overfit > one checks it still works in every market regime > one strips out plain momentum and value to see if any real edge is left Each of those six is a job a fund pays a $600,000-a-year quant to do. He runs all six for the price of an API bill. The rule that makes it work: the agent that builds a signal never gets to approve it. A separate, stronger agent tries to kill it first. Whatever survives all six by morning is real, new alpha. One trader's already running this exact swarm on Polymarket. That $2M wallet is public, every trade on-chain. The full build is in the post below - six agents, the tool that runs them, and the five mistakes that kill most people. Bookmark & read this before it's buried.

cvxv666

103,734 views • 2 months ago

$50 → $1,000/day OpenClaw has been running for 72 hours straight no breaks. no sleep. no mercy. it uses Claude Sonnet to read the market like a chess engine - every 5 minutes it scans thousands of Polymarket contracts Calculates true probability, and only bets when the edge is real here's what happens every 5 minutes: - Claude Sonnet builds a fair value model - scans 9,900+ live markets - ignores anything under 5% edge - sizes the bet (Kelly criterion, capped at 8% bankroll) - fires the order via CLOB - profits pay for their own API costs the agent has one rule: don't die if balance hits $0 - it's over forever
0:24

$50 → $1,000/day OpenClaw has been running for 72 hours straight no breaks. no sleep. no mercy. it uses Claude Sonnet to read the market like a chess engine - every 5 minutes it scans thousands of Polymarket contracts Calculates true probability, and only bets when the edge is real here's what happens every 5 minutes: - Claude Sonnet builds a fair value model - scans 9,900+ live markets - ignores anything under 5% edge - sizes the bet (Kelly criterion, capped at 8% bankroll) - fires the order via CLOB - profits pay for their own API costs the agent has one rule: don't die if balance hits $0 - it's over forever

BuBBliK

61,421 views • 7 months ago

How the 15-min BTC bot hits 90% win rate. I tracked one trade from the moment a new window opened. BTC moves on spot. Binance, Coinbase, wherever. The price shifts. But Polymarket odds take a few seconds to catch up. That gap is the whole game. He watches spot. Sees momentum. Enters the side that should win while Polymarket still shows old numbers. By the time the window closes, the lag is gone. Odds corrected. He collected. Buys in cheap. Gets paid full price. Rinse and repeat. The bot does not care where BTC is going tomorrow. It only cares what happened five seconds ago that Polymarket has not priced in yet. Profile: People keep asking where the edge is. It is not hidden. It is just faster than you.
0:35

How the 15-min BTC bot hits 90% win rate. I tracked one trade from the moment a new window opened. BTC moves on spot. Binance, Coinbase, wherever. The price shifts. But Polymarket odds take a few seconds to catch up. That gap is the whole game. He watches spot. Sees momentum. Enters the side that should win while Polymarket still shows old numbers. By the time the window closes, the lag is gone. Odds corrected. He collected. Buys in cheap. Gets paid full price. Rinse and repeat. The bot does not care where BTC is going tomorrow. It only cares what happened five seconds ago that Polymarket has not priced in yet. Profile: People keep asking where the edge is. It is not hidden. It is just faster than you.

Blaze

178,845 views • 9 months ago

A blackjack dealer in Macau got blacklisted from the VIP rooms last spring for counting cards. By August he couldn't get a floor job at any property in Cotai. So he deposit $500, ran a Hermes trading agent on Polymarket and pulled in $881,319 over the next 14 months. His wallet: The casinos taught him one thing - count, size your bet to your edge, walk away when the edge is gone. So he makes 5 trades a day. Not 50. Not 500. Five. And waits for the rest. Here's the actual stack. Claude Opus 4.7 reads the order book nightly, scores every threshold market by Markov persistence x Kelly edge, and surfaces the 2-3 mispriced ones. Hermes Agent by NousResearch executes. A $10/mo Hetzner VPS runs it 24/7. Telegram pings on every fill. Total cost: $10/month. Setup: 30 minutes. No coding. One trade in April: Will Bitcoin reach $90,000? Market said 1.2¢. He put $3,088 on Yes. It hit. +$123,196. A 3,988% return on a single position. The real edge is the nightly self-learning loop. Every midnight, Opus reads the day's trade journal and rewrites MIN_PROB and MIN_EDGE in the .env file. Last week the threshold was 0.87. This week 0.89. Next week maybe 0.91. His version of the bot has rewritten itself 412 times in 14 months. The Macau syndicate couldn't ban him from Polymarket. The bot doesn't sit at a table. It doesn't show a passport. It just hunts the tails. Save this post - if you want to build something of your own based on Hermes. Or just start copying algorithm that has improved itself 412 times:
1:12

A blackjack dealer in Macau got blacklisted from the VIP rooms last spring for counting cards. By August he couldn't get a floor job at any property in Cotai. So he deposit $500, ran a Hermes trading agent on Polymarket and pulled in $881,319 over the next 14 months. His wallet: The casinos taught him one thing - count, size your bet to your edge, walk away when the edge is gone. So he makes 5 trades a day. Not 50. Not 500. Five. And waits for the rest. Here's the actual stack. Claude Opus 4.7 reads the order book nightly, scores every threshold market by Markov persistence x Kelly edge, and surfaces the 2-3 mispriced ones. Hermes Agent by NousResearch executes. A $10/mo Hetzner VPS runs it 24/7. Telegram pings on every fill. Total cost: $10/month. Setup: 30 minutes. No coding. One trade in April: Will Bitcoin reach $90,000? Market said 1.2¢. He put $3,088 on Yes. It hit. +$123,196. A 3,988% return on a single position. The real edge is the nightly self-learning loop. Every midnight, Opus reads the day's trade journal and rewrites MIN_PROB and MIN_EDGE in the .env file. Last week the threshold was 0.87. This week 0.89. Next week maybe 0.91. His version of the bot has rewritten itself 412 times in 14 months. The Macau syndicate couldn't ban him from Polymarket. The bot doesn't sit at a table. It doesn't show a passport. It just hunts the tails. Save this post - if you want to build something of your own based on Hermes. Or just start copying algorithm that has improved itself 412 times:

cvxv666

208,632 views • 4 months ago

Loops vs. Graphs, clearly explained! loops are great, but they have a ceiling: a loop makes one unit of work better. it cannot decide which units exist. so you end up with a very good agent running the wrong three steps, in the wrong order, one at a time. Graph engineering fixes this by moving the decision up a layer: what runs, what runs at the same time, and what never runs at all. you need both. here's how it works: a graph splits your system into two kinds of decision. ↳ inside a unit: the loop. produce, check, correct, repeat until green ↳ between units: the graph. split, fan out, merge, gate, send back Prompts → Context → Harness → Loops → Graphs you get parallel work, isolated contexts, and steps that stop running when nothing needs them. the trick is being selective about what becomes a node. only spend a model where judgment lives. merging, ranking, deduping and schema checks are edges, and edges are code. free, instant, and they cannot be argued out of a verdict. a graph where every edge is an agent pays rent on its own wiring. one thing to know before you scale it. a graph has two return paths, and almost everyone builds one. ↳ the correction edge is short. a gate rejects one unit back to the step that produced it, and it fixes the run you are in ↳ the learning edge is long. an accepted result goes back to the splitter as a constraint, and it fixes every run after skip the second and you get a graph that is fast and never gets smarter. next week it starts from the same place with the same blind spots. and a smaller one that eats whole nights: when a unit fails, return that unit, not the batch. send back four slices because one failed and you have just rewritten three correct ones. do it twice in a run and the run never converges. below i have quoted my full guide on graph engineering. it covers the three topologies, the verifier patterns, and where the gate should actually open. save this and read it below ↓
0:41

Loops vs. Graphs, clearly explained! loops are great, but they have a ceiling: a loop makes one unit of work better. it cannot decide which units exist. so you end up with a very good agent running the wrong three steps, in the wrong order, one at a time. Graph engineering fixes this by moving the decision up a layer: what runs, what runs at the same time, and what never runs at all. you need both. here's how it works: a graph splits your system into two kinds of decision. ↳ inside a unit: the loop. produce, check, correct, repeat until green ↳ between units: the graph. split, fan out, merge, gate, send back Prompts → Context → Harness → Loops → Graphs you get parallel work, isolated contexts, and steps that stop running when nothing needs them. the trick is being selective about what becomes a node. only spend a model where judgment lives. merging, ranking, deduping and schema checks are edges, and edges are code. free, instant, and they cannot be argued out of a verdict. a graph where every edge is an agent pays rent on its own wiring. one thing to know before you scale it. a graph has two return paths, and almost everyone builds one. ↳ the correction edge is short. a gate rejects one unit back to the step that produced it, and it fixes the run you are in ↳ the learning edge is long. an accepted result goes back to the splitter as a constraint, and it fixes every run after skip the second and you get a graph that is fast and never gets smarter. next week it starts from the same place with the same blind spots. and a smaller one that eats whole nights: when a unit fails, return that unit, not the batch. send back four slices because one failed and you have just rewritten three correct ones. do it twice in a run and the run never converges. below i have quoted my full guide on graph engineering. it covers the three topologies, the verifier patterns, and where the gate should actually open. save this and read it below ↓

Hanako

73,867 views • 1 month ago

US export-control disable Claude Fable 5 overnight. Not for what it said - for what one trader did with it. He turned an $11,000 wreck into $300,000 profit in five days. Then they pulled Claude Fable 5 from public hands. Too late - his bot's still running. His wallet: Every other Claude model still works. The two best ones don't. Anthropic's calling it a misunderstanding. Maybe. But think about what gets a model killed by export control overnight, no warning. Here's what Fable 5 actually did for him. It doesn't predict price. It computes the whole probability surface of a market - every outcome, weighted, faster than the market can correct itself. It finds the bets the crowd writes off as impossible, runs the real math, and buys them for pennies when the math disagrees. One of those just paid x205. A regular guy, a model this sharp, a market no regulator polices. They closed that door overnight - one night too late. He's holding one of the last builds that ever touched a live market. It's still on Polymarket right now, hunting the same mispriced tails. Fable 5 is locked. The algorithm it wrote is loose. You can't open the model anymore. You can still copy the wallet it built in two clicks:
0:46

US export-control disable Claude Fable 5 overnight. Not for what it said - for what one trader did with it. He turned an $11,000 wreck into $300,000 profit in five days. Then they pulled Claude Fable 5 from public hands. Too late - his bot's still running. His wallet: Every other Claude model still works. The two best ones don't. Anthropic's calling it a misunderstanding. Maybe. But think about what gets a model killed by export control overnight, no warning. Here's what Fable 5 actually did for him. It doesn't predict price. It computes the whole probability surface of a market - every outcome, weighted, faster than the market can correct itself. It finds the bets the crowd writes off as impossible, runs the real math, and buys them for pennies when the math disagrees. One of those just paid x205. A regular guy, a model this sharp, a market no regulator polices. They closed that door overnight - one night too late. He's holding one of the last builds that ever touched a live market. It's still on Polymarket right now, hunting the same mispriced tails. Fable 5 is locked. The algorithm it wrote is loose. You can't open the model anymore. You can still copy the wallet it built in two clicks:

cvxv666

77,366 views • 3 months ago

THIS WALLET STACKED $230K ON BTC UP/DOWN BETS. THE BLUEPRINT TO AUTOMATE THE SAME EDGE WITH CLAUDE The wallet is $230K all-time, every position a Bitcoin or Ethereum Up or Down market It never guesses direction. It enters only when the math and the market disagree THE STRATEGY: BTC moves are not fully random. When the market enters a committed directional state, continuation is measurable. That is Markov persistence Entry signal: > Δ = p̂ − q ≥ ε Model probability minus market price. Enter only on a 5% gap or more Persistence filter: > p(j*,j*) ≥ 0.87 Only trade states with 0.87 persistence or higher. Below that, skip. This is what holds the win rate above 65% with zero directional guessing Payout: > r = (1 − q) / q At q = 0.647 that is +54.5% a win. At q = 0.441, +126.7%. Lower entry price, bigger asymmetry Sizing: > f* = p − (1−p)/b Kelly. At p = 0.87, b = 0.647, f* ≈ 0.71. Size to the edge, never to gut HOW TO BUILD IT WITH CLAUDE: What separates this from a static bot: Claude reads its own trade journal every night and rewrites its own thresholds 1. Take an open-source Polymarket bot repo as your base logic. Feed it to Claude and have it migrate to CLOB v2: py_clob_client_v2, Safe wallet support, fee-aware evaluation 2. Hard-code the filters. Enter only when Δ ≥ 0.05 and p(j*,j*) ≥ 0.87. Apply Kelly on every fill. 3. Run DRY_RUN first. Log every signal, entry price, Markov state, and simulated P/L. No real money until the numbers hold for days 4. The nightly loop. Claude reads the journal, finds which persistence states actually won, adjusts MIN_PROB and MIN_EDGE, ships tomorrow's rules. The agent is sharper after 50 to 100 trades THE SETUP: Claude Opus as the brain. An open-source repo as the starting logic. A Polygon wallet with $50 to $100. Telegram for the morning report Start at $1 to $2 per trade while it learns. Scale only when the dry runs and the live fills line up 17,000 trades compound a thin edge into six figures. The model finds the edge. The nightly loop keeps it sharp Bookmark before you point a bot at your first window
0:23

THIS WALLET STACKED $230K ON BTC UP/DOWN BETS. THE BLUEPRINT TO AUTOMATE THE SAME EDGE WITH CLAUDE The wallet is $230K all-time, every position a Bitcoin or Ethereum Up or Down market It never guesses direction. It enters only when the math and the market disagree THE STRATEGY: BTC moves are not fully random. When the market enters a committed directional state, continuation is measurable. That is Markov persistence Entry signal: > Δ = p̂ − q ≥ ε Model probability minus market price. Enter only on a 5% gap or more Persistence filter: > p(j*,j*) ≥ 0.87 Only trade states with 0.87 persistence or higher. Below that, skip. This is what holds the win rate above 65% with zero directional guessing Payout: > r = (1 − q) / q At q = 0.647 that is +54.5% a win. At q = 0.441, +126.7%. Lower entry price, bigger asymmetry Sizing: > f* = p − (1−p)/b Kelly. At p = 0.87, b = 0.647, f* ≈ 0.71. Size to the edge, never to gut HOW TO BUILD IT WITH CLAUDE: What separates this from a static bot: Claude reads its own trade journal every night and rewrites its own thresholds 1. Take an open-source Polymarket bot repo as your base logic. Feed it to Claude and have it migrate to CLOB v2: py_clob_client_v2, Safe wallet support, fee-aware evaluation 2. Hard-code the filters. Enter only when Δ ≥ 0.05 and p(j*,j*) ≥ 0.87. Apply Kelly on every fill. 3. Run DRY_RUN first. Log every signal, entry price, Markov state, and simulated P/L. No real money until the numbers hold for days 4. The nightly loop. Claude reads the journal, finds which persistence states actually won, adjusts MIN_PROB and MIN_EDGE, ships tomorrow's rules. The agent is sharper after 50 to 100 trades THE SETUP: Claude Opus as the brain. An open-source repo as the starting logic. A Polygon wallet with $50 to $100. Telegram for the morning report Start at $1 to $2 per trade while it learns. Scale only when the dry runs and the live fills line up 17,000 trades compound a thin edge into six figures. The model finds the edge. The nightly loop keeps it sharp Bookmark before you point a bot at your first window

Yarchi

22,966 views • 4 months ago

one agent → a loop → a graph → a system that throttles itself Google just put the whole ladder in one free course. most people never leave the first rung. 2 hours 4 minutes, taught by two Google Cloud engineers. no signup, no paywall. 17:55 - build your first agent 39:55 - loop engineering: iterate, check, break 1:13:25 - graph engineering 1:35:27 - agents that throttle themselves 1:56:19 - the full graph for multi-agent systems the fourth one is the step nobody talks about. an agent that cannot throttle itself is not autonomous, it is just unattended. everyone builds one agent and calls it done. this is the version where the agents wire themselves into a graph and decide their own rate. and the thing that decides when one of them is allowed to stop is not in the course. that part you build yourself. bookmark and watch it today, then read the eval setup below
2:04:19

one agent → a loop → a graph → a system that throttles itself Google just put the whole ladder in one free course. most people never leave the first rung. 2 hours 4 minutes, taught by two Google Cloud engineers. no signup, no paywall. 17:55 - build your first agent 39:55 - loop engineering: iterate, check, break 1:13:25 - graph engineering 1:35:27 - agents that throttle themselves 1:56:19 - the full graph for multi-agent systems the fourth one is the step nobody talks about. an agent that cannot throttle itself is not autonomous, it is just unattended. everyone builds one agent and calls it done. this is the version where the agents wire themselves into a graph and decide their own rate. and the thing that decides when one of them is allowed to stop is not in the course. that part you build yourself. bookmark and watch it today, then read the eval setup below

Hanako

53,692 views • 1 month ago

Citadel quant quit and walked out with one thing fund never lets leave the building: 120-year-old math that sees a market crash before the chart does. Every big desk runs it. None of them explain it. He turned it into an HMM bot running BTC 5min markets and made $570,000. It switches strategy the second the regime flips and cuts everything the moment panic spikes. His wallet: Full breakdown below. Here's what they actually see. You look at a Bitcoin chart and see a price. A quant sees three states: bull, bear, sideways. The market's always in one. The only question that pays is which one comes next. That's the math. It's called a Markov chain, and it runs on one rule: where price goes next depends only on where it is now. You count how often each state flips to another and drop it into a grid. That grid is a map of the market. Now the part that feels illegal. Want the odds of a crash twelve steps out? You multiply that grid by itself twelve times. One line of math, and you've got the probability of every future state. Then it acts: run the trend when calm, flip to mean reversion in chop, cut everything when panic odds spike. It doesn't predict the future. It gives you the odds of every version of it, then sizes the bet to match. That's the edge. Renaissance ran 66% a year for 30 years on math like this. You can build the basic version in a weekend. Bookmark & read this before it gets buried.
0:42

Citadel quant quit and walked out with one thing fund never lets leave the building: 120-year-old math that sees a market crash before the chart does. Every big desk runs it. None of them explain it. He turned it into an HMM bot running BTC 5min markets and made $570,000. It switches strategy the second the regime flips and cuts everything the moment panic spikes. His wallet: Full breakdown below. Here's what they actually see. You look at a Bitcoin chart and see a price. A quant sees three states: bull, bear, sideways. The market's always in one. The only question that pays is which one comes next. That's the math. It's called a Markov chain, and it runs on one rule: where price goes next depends only on where it is now. You count how often each state flips to another and drop it into a grid. That grid is a map of the market. Now the part that feels illegal. Want the odds of a crash twelve steps out? You multiply that grid by itself twelve times. One line of math, and you've got the probability of every future state. Then it acts: run the trend when calm, flip to mean reversion in chop, cut everything when panic odds spike. It doesn't predict the future. It gives you the odds of every version of it, then sizes the bet to match. That's the edge. Renaissance ran 66% a year for 30 years on math like this. You can build the basic version in a weekend. Bookmark & read this before it gets buried.

cvxv666

93,826 views • 2 months ago

The new Fable 5.1 built three games and did it 7% cheaper! We gave the same prompts through Atomic Agent: ✦ Sticky ball that rolls up everything in its path ✦ Capybara surfing down a river, Subway Surfers style ✦ Dumpling running an endless conveyor dodge Outputs: ✦ Claude Fable 5, via Atomic Agent: $7.65 ✦ Claude Fable 5.1, via Atomic Agent: $7.08 Run it through an agent and that cache discount is where the money hides. The loop reads the file, watches it break, and pulls the whole context back in to fix it, over and over. Every one of those re-reads costs a quarter on 5.1 of what it cost on 5. Same games on screen, and the newer version just spends less to get there. Across three builds the loop paid for the upgrade on its own.
0:33

The new Fable 5.1 built three games and did it 7% cheaper! We gave the same prompts through Atomic Agent: ✦ Sticky ball that rolls up everything in its path ✦ Capybara surfing down a river, Subway Surfers style ✦ Dumpling running an endless conveyor dodge Outputs: ✦ Claude Fable 5, via Atomic Agent: $7.65 ✦ Claude Fable 5.1, via Atomic Agent: $7.08 Run it through an agent and that cache discount is where the money hides. The loop reads the file, watches it break, and pulls the whole context back in to fix it, over and over. Every one of those re-reads costs a quarter on 5.1 of what it cost on 5. Same games on screen, and the newer version just spends less to get there. Across three builds the loop paid for the upgrade on its own.

Atomic Agent

20,255 views • 28 days ago

Obsidian + Claude Code is how Andrej Karpathy turns 15 years of notes into a brain that works while he sleeps. The method behind it costs $0 and takes 5 minutes a day and it outlived every productivity app since 2011. Every idea, link, and half-thought gets appended to the top of a single note, with no sorting, no folders, no tags. Once a week he scrolls through, and anything that still matters gets copied back to the top. Weak ideas sink. Strong ideas resurface 5, 10, 20 times and by the 20th pass your brain has already wired them into everything else you know. Repetition as a filter. That's the whole system. Now the 2026 upgrade almost nobody is running: Drop that note into an Obsidian vault and open the folder with Claude Code not the chat app, the terminal agent that reads files. 3 commands change everything: read my last 30 days of notes and find the 3 ideas I keep circling without acting on link every note mentioning this project into a CLAUDE.md of what I actually believe about it draft this week's post from the idea that resurfaced most Claude Code doesn't search your vault it walks it, follows the wikilinks, and hands you the patterns you were too close to see. Karpathy method filters the signal, Obsidian stores it, Claude Code compounds it. One is a habit, one is a folder, one is $20 a month together they're the closest thing to a second brain that thinks back. Most people collect notes for 10 years and never read them once. His notes read him.
1:29

Obsidian + Claude Code is how Andrej Karpathy turns 15 years of notes into a brain that works while he sleeps. The method behind it costs $0 and takes 5 minutes a day and it outlived every productivity app since 2011. Every idea, link, and half-thought gets appended to the top of a single note, with no sorting, no folders, no tags. Once a week he scrolls through, and anything that still matters gets copied back to the top. Weak ideas sink. Strong ideas resurface 5, 10, 20 times and by the 20th pass your brain has already wired them into everything else you know. Repetition as a filter. That's the whole system. Now the 2026 upgrade almost nobody is running: Drop that note into an Obsidian vault and open the folder with Claude Code not the chat app, the terminal agent that reads files. 3 commands change everything: read my last 30 days of notes and find the 3 ideas I keep circling without acting on link every note mentioning this project into a CLAUDE.md of what I actually believe about it draft this week's post from the idea that resurfaced most Claude Code doesn't search your vault it walks it, follows the wikilinks, and hands you the patterns you were too close to see. Karpathy method filters the signal, Obsidian stores it, Claude Code compounds it. One is a habit, one is a folder, one is $20 a month together they're the closest thing to a second brain that thinks back. Most people collect notes for 10 years and never read them once. His notes read him.

Spike 1%

12,467 views • 2 months ago

CITADEL QUANTS HIDE THE GAME BEHIND A FORMULA WITH TEN TERMS, PREDICTION MARKETS RUN ON EXACTLY ONE The secret is one line. Edge = P(model) − P(market) That Z-field monster on screen is a stack of simple signals wearing a costume. A prediction market strips all of it down to one number, the crowd's guess at a probability. In plain words. If you can turn fresh news into a probability before the crowd reprices, the gap between your number and their price is the whole edge. That is why the game is not IQ, it is speed to a number. An AI that reads live news does in ninety seconds what the tab takes hours to price in. I broke down the four exact prompts that turn Grok into a probability desk, the sizing rules and the ten minute daily routine in the article. Still bets, not a salary, so size like it & the full playbook is in the article. Read it ↓
0:46

CITADEL QUANTS HIDE THE GAME BEHIND A FORMULA WITH TEN TERMS, PREDICTION MARKETS RUN ON EXACTLY ONE The secret is one line. Edge = P(model) − P(market) That Z-field monster on screen is a stack of simple signals wearing a costume. A prediction market strips all of it down to one number, the crowd's guess at a probability. In plain words. If you can turn fresh news into a probability before the crowd reprices, the gap between your number and their price is the whole edge. That is why the game is not IQ, it is speed to a number. An AI that reads live news does in ninety seconds what the tab takes hours to price in. I broke down the four exact prompts that turn Grok into a probability desk, the sizing rules and the ten minute daily routine in the article. Still bets, not a salary, so size like it & the full playbook is in the article. Read it ↓

slash1s

15,977 views • 2 months ago

THIS GUY CONNECTED HIS AI AGENTS TO HIS OBSIDIAN AND BUILT A BRAIN THAT LEARNS ON ITS OWN. HERE'S HOW TO BUILD IT Obsidian is just markdown files sitting in a folder. That turns out to be the perfect memory for an AI agent, because an agent can read and write those files directly. He wired his agents into the vault so they pull context from it, do the work, and write what they learned back. The notes aren't the point. The loop is, and it gets sharper every cycle How to build it: 1. Point an agent at your vault. The fastest way, no plugins, no API keys: open a terminal and run npx obsidian-mcp /path/to/your/vault. That exposes your Obsidian folder to Claude as a tool it can read, search, and write to. Add it to your Claude Code or Cowork config and restart 2. Confirm it can see the brain. Ask it: "list the notes in my vault and summarize what's in them." If it reads them back, the connection is live. Now it starts every task with everything the vault already holds instead of from zero 3. Give each agent one job and a write-back rule. Tell it: "research this, then save what you found as a new note in /brain with links to related notes." One agent researches, one summarizes, one plans. Each writes its output back into the vault 4. Close the loop. Add one line to every agent's instructions: "read /brain before starting, write your result back when done." Now each task leaves the vault richer, and the next run reads that before it works. It compounds instead of resetting 5. You only steer. Review what the brain produces, point it at the next thing. The agents handle the reading, writing, and connecting The edge isn't better notes. It's a brain that feeds itself, so the work gets sharper every cycle instead of starting over Bookmark this
0:24

THIS GUY CONNECTED HIS AI AGENTS TO HIS OBSIDIAN AND BUILT A BRAIN THAT LEARNS ON ITS OWN. HERE'S HOW TO BUILD IT Obsidian is just markdown files sitting in a folder. That turns out to be the perfect memory for an AI agent, because an agent can read and write those files directly. He wired his agents into the vault so they pull context from it, do the work, and write what they learned back. The notes aren't the point. The loop is, and it gets sharper every cycle How to build it: 1. Point an agent at your vault. The fastest way, no plugins, no API keys: open a terminal and run npx obsidian-mcp /path/to/your/vault. That exposes your Obsidian folder to Claude as a tool it can read, search, and write to. Add it to your Claude Code or Cowork config and restart 2. Confirm it can see the brain. Ask it: "list the notes in my vault and summarize what's in them." If it reads them back, the connection is live. Now it starts every task with everything the vault already holds instead of from zero 3. Give each agent one job and a write-back rule. Tell it: "research this, then save what you found as a new note in /brain with links to related notes." One agent researches, one summarizes, one plans. Each writes its output back into the vault 4. Close the loop. Add one line to every agent's instructions: "read /brain before starting, write your result back when done." Now each task leaves the vault richer, and the next run reads that before it works. It compounds instead of resetting 5. You only steer. Review what the brain produces, point it at the next thing. The agents handle the reading, writing, and connecting The edge isn't better notes. It's a brain that feeds itself, so the work gets sharper every cycle instead of starting over Bookmark this

Yarchi

58,591 views • 3 months ago

Japan just changed what an AI model even is. New Sakana Fugu doesn't try to out-think GPT-5, Claude, or Gemini. It conducts all three at once - and beats every one of them. A trader in Tokyo unleashed it on the fastest market alive - 5min Bitcoin binary and turned $6,200 into $304,865. His wallet: The frontier just stopped being which model is smartest. It's who's conducting them - and the market hasn't priced that in yet. Sakana Fugu isn't a bigger model - it's a full multi-agent orchestration system. The coordinator behind it carries about 10,000 parameters, evolved rather than hand-coded, and it runs the most capable models on earth like a single instrument. Pointed at Bitcoin, here's what it does every five minutes. It assembles a team from a pool of frontier models and assigns each one a role: > Thinker - reads the candle, the order book, the news, builds the plan > Worker - turns the plan into one call: up or down, and how much > Verifier - votes ACCEPT or REVISE before a cent moves If the Verifier says REVISE, nothing trades. Fugu reads its own miss, reroutes, even calls itself for a corrective round, and runs it again. No look-ahead, ever - the next candle only appears after it commits. This is what should worry every lab still chasing a bigger model: the edge was never scale. It's orchestration - and Fugu does it better than anything alive. Bookmark this - when the whole timeline is chasing orchestration in six months, you'll already have the breakdown. You're not going to wire up an orchestra of frontier models yourself. Mirror the wallet Fugu runs instead:
0:41

Japan just changed what an AI model even is. New Sakana Fugu doesn't try to out-think GPT-5, Claude, or Gemini. It conducts all three at once - and beats every one of them. A trader in Tokyo unleashed it on the fastest market alive - 5min Bitcoin binary and turned $6,200 into $304,865. His wallet: The frontier just stopped being which model is smartest. It's who's conducting them - and the market hasn't priced that in yet. Sakana Fugu isn't a bigger model - it's a full multi-agent orchestration system. The coordinator behind it carries about 10,000 parameters, evolved rather than hand-coded, and it runs the most capable models on earth like a single instrument. Pointed at Bitcoin, here's what it does every five minutes. It assembles a team from a pool of frontier models and assigns each one a role: > Thinker - reads the candle, the order book, the news, builds the plan > Worker - turns the plan into one call: up or down, and how much > Verifier - votes ACCEPT or REVISE before a cent moves If the Verifier says REVISE, nothing trades. Fugu reads its own miss, reroutes, even calls itself for a corrective round, and runs it again. No look-ahead, ever - the next candle only appears after it commits. This is what should worry every lab still chasing a bigger model: the edge was never scale. It's orchestration - and Fugu does it better than anything alive. Bookmark this - when the whole timeline is chasing orchestration in six months, you'll already have the breakdown. You're not going to wire up an orchestra of frontier models yourself. Mirror the wallet Fugu runs instead:

cvxv666

75,691 views • 3 months 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 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 Himanshu Kumar so I can DM you Save this post. Build the copytrading system this week. Start with $200. Scale on evidence.
0:43

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 Himanshu Kumar so I can DM you Save this post. Build the copytrading system this week. Start with $200. Scale on evidence.

Himanshu Kumar

63,769 views • 3 months ago