Offshore
Video
Moon Dev
The Great Equalizer: How To Code A Trading Bot That Never Gets Liquidated
most traders treat the coingecko api like a simple price checker but the real money is hidden in the endpoints that nobody bothers to read. it is the difference between catching a trending coin before it moons and being the exit liquidity for someone who spent ten minutes writing a script. there is a specific way to structure these calls that bypasses the noise and gets you straight to the alpha and if you miss this one detail you will probably just end up with an empty csv file
if you think you need a degree in computer science to escape the liquidation cycle you have already lost the game before it started. i am moon dev and i believe that code is the great equalizer because through losing money with liquidations and over trading i knew i had to automate my trading. i spent hundreds of thousands on developers for apps in the past thinking i would not be able to code myself but now i know better
with bots you must iterate to success so i decided to learn live and now i have fully automated systems trading for me instead of getting liquidated. the scanner we are building here is designed to capture every hourly shift in market sentiment so we have a permanent record of what the crowd was chasing. i have noticed that tokens often trend for a few days before cooling off and then hitting a second much larger wave
by building a local database of these historical trends you can start to see patterns that the web interface simply cannot show you. most traders are looking at the same top ten list on the homepage while we are pulling the raw data directly from the v3 api. there is a secret way to filter for the absolute newest listings before they even hit the front page and i am going to show you how to tap into that stream
getting good at this is mostly just about being willing to read the documentation while everyone else is looking for a shortcut. most people see a rate limit error and give up but that is actually where the edge begins because the friction keeps the lazy traders out. once you find the categories endpoint you can begin searching for artificial intelligence or gaming tokens specifically which allows you to front run narrative shifts
narrative shifts are where the largest multipliers are found but you need to see the actual volume and not just wash trading data. the api allows us to pull specific market cap data and volume metrics that we can then clean using a pandas data frame. if you have never used pandas before it is essentially a spreadsheet on steroids that allows you to sort thousands of rows of data in a fraction of a second
i used to sit at my desk all day trying to find these new coins manually but now the script does the heavy lifting while i sleep. we can set up a function that looks specifically at the new coins page which is a gold mine for finding projects that have only been live for a few hours. capturing these right at the source is the goal because by the time they are trending on twitter the move is already halfway over
when we look at the nft section of the api we are looking for more than just floor prices because that data can be misleading. we need to see the unique holder count and how that number is changing over a twenty four hour period. if you see a collection where the floor is rising but the number of unique addresses is dropping it usually means a few whales are just trading back and forth to create fake hype
using the v3 api for nfts requires a specific endpoint path that is not always obvious if you are just glancing at the main docs. the trick is knowing how to handle the nested json data so you can actually perform math on the price changes. once you have the data in a human readable format like a csv file you can start to build logical filters that only alert you when specific criteria are met
if you are using the free tier of the api you are going to hit a wall very quickly unless you implement a smart time s[...]
The Great Equalizer: How To Code A Trading Bot That Never Gets Liquidated
most traders treat the coingecko api like a simple price checker but the real money is hidden in the endpoints that nobody bothers to read. it is the difference between catching a trending coin before it moons and being the exit liquidity for someone who spent ten minutes writing a script. there is a specific way to structure these calls that bypasses the noise and gets you straight to the alpha and if you miss this one detail you will probably just end up with an empty csv file
if you think you need a degree in computer science to escape the liquidation cycle you have already lost the game before it started. i am moon dev and i believe that code is the great equalizer because through losing money with liquidations and over trading i knew i had to automate my trading. i spent hundreds of thousands on developers for apps in the past thinking i would not be able to code myself but now i know better
with bots you must iterate to success so i decided to learn live and now i have fully automated systems trading for me instead of getting liquidated. the scanner we are building here is designed to capture every hourly shift in market sentiment so we have a permanent record of what the crowd was chasing. i have noticed that tokens often trend for a few days before cooling off and then hitting a second much larger wave
by building a local database of these historical trends you can start to see patterns that the web interface simply cannot show you. most traders are looking at the same top ten list on the homepage while we are pulling the raw data directly from the v3 api. there is a secret way to filter for the absolute newest listings before they even hit the front page and i am going to show you how to tap into that stream
getting good at this is mostly just about being willing to read the documentation while everyone else is looking for a shortcut. most people see a rate limit error and give up but that is actually where the edge begins because the friction keeps the lazy traders out. once you find the categories endpoint you can begin searching for artificial intelligence or gaming tokens specifically which allows you to front run narrative shifts
narrative shifts are where the largest multipliers are found but you need to see the actual volume and not just wash trading data. the api allows us to pull specific market cap data and volume metrics that we can then clean using a pandas data frame. if you have never used pandas before it is essentially a spreadsheet on steroids that allows you to sort thousands of rows of data in a fraction of a second
i used to sit at my desk all day trying to find these new coins manually but now the script does the heavy lifting while i sleep. we can set up a function that looks specifically at the new coins page which is a gold mine for finding projects that have only been live for a few hours. capturing these right at the source is the goal because by the time they are trending on twitter the move is already halfway over
when we look at the nft section of the api we are looking for more than just floor prices because that data can be misleading. we need to see the unique holder count and how that number is changing over a twenty four hour period. if you see a collection where the floor is rising but the number of unique addresses is dropping it usually means a few whales are just trading back and forth to create fake hype
using the v3 api for nfts requires a specific endpoint path that is not always obvious if you are just glancing at the main docs. the trick is knowing how to handle the nested json data so you can actually perform math on the price changes. once you have the data in a human readable format like a csv file you can start to build logical filters that only alert you when specific criteria are met
if you are using the free tier of the api you are going to hit a wall very quickly unless you implement a smart time s[...]
Offshore
Moon Dev The Great Equalizer: How To Code A Trading Bot That Never Gets Liquidated most traders treat the coingecko api like a simple price checker but the real money is hidden in the endpoints that nobody bothers to read. it is the difference between catching…
leep function. i prefer a two second delay between calls when i am scanning the top twenty collections just to stay under the radar and avoid being blacklisted. if you decide to upgrade to the paid tier you get access to exclusive market data endpoints that provide the kind of depth that professional firms use
the difference between the free and paid versions is basically the speed at which you can iterate and the number of calls you can make per minute. for a beginner the free tier is more than enough to learn the logic and build a basic scanner that runs once an hour. there is a way to bridge this data into a custom dashboard that sends you alerts the moment a specific metric hits a trigger point
building these systems is not about being a genius it is about being consistent and willing to fix the small errors that pop up during the build. i spent part of my day just trying to figure out why my nft data was printing every third line instead of a clean list. that is the reality of coding and automation but once it is fixed the bot works for you forever without complaining or getting tired
automation is the only way to remove the human error that leads to liquidations and blown accounts. i am not a financial advisor but i am a guy who found out that code is the only way to survive the volatility of these markets. once you have your scanner running you can spend your time thinking about strategy instead of staring at candles and wondering if you should click buy
the reason i share everything live is because i want to show that anyone can do this if they are willing to put in the time. i am constantly building new scanners and refining my bots because the market is always evolving. if you can master the coingecko api you have a massive advantage over ninety nine percent of the people trading right now who are just guessing based on memes
it takes effort to move from a manual trader to an automated one but the freedom it provides is worth every hour of frustration. you don't need to spend thousands on developers when you can learn to build these tools yourself and keep all the alpha. the goal is to have a fully automated system that finds the opportunities for you so you can focus on the big picture and stop being a slave to the screen
tweet
the difference between the free and paid versions is basically the speed at which you can iterate and the number of calls you can make per minute. for a beginner the free tier is more than enough to learn the logic and build a basic scanner that runs once an hour. there is a way to bridge this data into a custom dashboard that sends you alerts the moment a specific metric hits a trigger point
building these systems is not about being a genius it is about being consistent and willing to fix the small errors that pop up during the build. i spent part of my day just trying to figure out why my nft data was printing every third line instead of a clean list. that is the reality of coding and automation but once it is fixed the bot works for you forever without complaining or getting tired
automation is the only way to remove the human error that leads to liquidations and blown accounts. i am not a financial advisor but i am a guy who found out that code is the only way to survive the volatility of these markets. once you have your scanner running you can spend your time thinking about strategy instead of staring at candles and wondering if you should click buy
the reason i share everything live is because i want to show that anyone can do this if they are willing to put in the time. i am constantly building new scanners and refining my bots because the market is always evolving. if you can master the coingecko api you have a massive advantage over ninety nine percent of the people trading right now who are just guessing based on memes
it takes effort to move from a manual trader to an automated one but the freedom it provides is worth every hour of frustration. you don't need to spend thousands on developers when you can learn to build these tools yourself and keep all the alpha. the goal is to have a fully automated system that finds the opportunities for you so you can focus on the big picture and stop being a slave to the screen
tweet
Offshore
Photo
Dimitry Nakhla | Babylon Capital®
RT @DimitryNakhla: Over the past 10 years, $MSFT has only experienced -25% drawdowns three times:
1. 2020 flash crash
2. 2022 bear market
3. Today
It’s rare to see $MSFT trade >25% below its highs — especially while fundamentals remain strong, the balance sheet is pristine, and the moat is firmly intact.
If $MSFT compounds earnings in the mid-teens over the next several years (a reasonable base case), periods like this tend to look very different in hindsight as they did in the past.
Volatility is the price of admission.
Long-term compounding is the reward.
tweet
RT @DimitryNakhla: Over the past 10 years, $MSFT has only experienced -25% drawdowns three times:
1. 2020 flash crash
2. 2022 bear market
3. Today
It’s rare to see $MSFT trade >25% below its highs — especially while fundamentals remain strong, the balance sheet is pristine, and the moat is firmly intact.
If $MSFT compounds earnings in the mid-teens over the next several years (a reasonable base case), periods like this tend to look very different in hindsight as they did in the past.
Volatility is the price of admission.
Long-term compounding is the reward.
tweet
Offshore
Video
Startup Archive
Frank Slootman’s promise to employees: “I am here to better your station in life”
Frank was brought in as CEO for three different companies: Data Domain, ServiceNow, and Snowflake. When he joined each company, the first thing he did was tell the employees:
“I am here to better your station in life.”
As he explains, “This is what builds fabric in organizations.” When employees feel like you have their back and know what’s important to them, it builds trust.
Frank also emphasizes that this can’t just be an empty promise — you have to mean it. It’ll make your work more meaningful too:
“We have changed people’s station in life on a massive scale in all these companies. That is the single-most enduring and rewarding aspect of doing this work. It’s not how well we did personally — which is also great. But after a while, it’s about how many people you’ve helped and their families.”
Video source: @FoundationCap (2024)
tweet
Frank Slootman’s promise to employees: “I am here to better your station in life”
Frank was brought in as CEO for three different companies: Data Domain, ServiceNow, and Snowflake. When he joined each company, the first thing he did was tell the employees:
“I am here to better your station in life.”
As he explains, “This is what builds fabric in organizations.” When employees feel like you have their back and know what’s important to them, it builds trust.
Frank also emphasizes that this can’t just be an empty promise — you have to mean it. It’ll make your work more meaningful too:
“We have changed people’s station in life on a massive scale in all these companies. That is the single-most enduring and rewarding aspect of doing this work. It’s not how well we did personally — which is also great. But after a while, it’s about how many people you’ve helped and their families.”
Video source: @FoundationCap (2024)
tweet
Offshore
Photo
God of Prompt
The AI reply bots are genius
They’re testing Opus 4.6 vs GPT 4! https://t.co/N89ccayUc2
tweet
The AI reply bots are genius
They’re testing Opus 4.6 vs GPT 4! https://t.co/N89ccayUc2
tweet
Offshore
Photo
Moon Dev
while i slept my clawbot built me my billion dollar social arb firm
wall street can not compete https://t.co/GGtqctnsWU
tweet
while i slept my clawbot built me my billion dollar social arb firm
wall street can not compete https://t.co/GGtqctnsWU
tweet
Offshore
Video
Michael Fritzell (Asian Century Stocks)
RT @AmirF15336: Full episode now on X:
My conversation with Dmitry Balyasny, founder and Chief Investment Officer of Balyasny Asset Management, one of the most successful and disciplined multi-manager hedge funds of the last two decades.
I sat down with Dmitry to talk about his journey from a 7-year-old immigrant from Kyiv who spoke no English to building a $30+ billion investment platform, and the hard-earned lessons about markets, talent, and human nature that shaped his path.
Dmitry founded Balyasny Asset Management in 2001 after cutting his teeth at Schonfeld Securities starting in 1994. He's built BAM into a dominant force in the multi-manager space, managing over $30 billion with a team of 2,300 employees worldwide. His approach combines rigorous risk management, systematic talent development, and a deep understanding of what separates great analysts from great portfolio managers.
Widely regarded as one of the most thoughtful operators in the hedge fund industry, Dmitry has spent three decades studying the psychology of trading, the evolution of market edge, and the organizational design required to scale excellence.
We spoke about:
- Immigrating from Kyiv at age 7 and how that shaped his worldview
- Reading Atlas Shrugged in college and how Ayn Rand's philosophy influenced his thinking
- Losing every dollar he made in his first year as a broker and why he persisted
- Building BAM from scratch and the most important decisions over 25 years
-Attracting and retaining talent in a hyper-competitive compensation environment
- How the nature of edge has changed over two decades
- What a day in the life of running a $30B hedge fund actually looks like
… and much more.
Links below
tweet
RT @AmirF15336: Full episode now on X:
My conversation with Dmitry Balyasny, founder and Chief Investment Officer of Balyasny Asset Management, one of the most successful and disciplined multi-manager hedge funds of the last two decades.
I sat down with Dmitry to talk about his journey from a 7-year-old immigrant from Kyiv who spoke no English to building a $30+ billion investment platform, and the hard-earned lessons about markets, talent, and human nature that shaped his path.
Dmitry founded Balyasny Asset Management in 2001 after cutting his teeth at Schonfeld Securities starting in 1994. He's built BAM into a dominant force in the multi-manager space, managing over $30 billion with a team of 2,300 employees worldwide. His approach combines rigorous risk management, systematic talent development, and a deep understanding of what separates great analysts from great portfolio managers.
Widely regarded as one of the most thoughtful operators in the hedge fund industry, Dmitry has spent three decades studying the psychology of trading, the evolution of market edge, and the organizational design required to scale excellence.
We spoke about:
- Immigrating from Kyiv at age 7 and how that shaped his worldview
- Reading Atlas Shrugged in college and how Ayn Rand's philosophy influenced his thinking
- Losing every dollar he made in his first year as a broker and why he persisted
- Building BAM from scratch and the most important decisions over 25 years
-Attracting and retaining talent in a hyper-competitive compensation environment
- How the nature of edge has changed over two decades
- What a day in the life of running a $30B hedge fund actually looks like
… and much more.
Links below
tweet
Offshore
Video
Moon Dev
The 7,547% Trading Bot Catch: Why Pro Developers Are Ditching Macs for This $15 Architecture
seeing a seven thousand five hundred percent return usually means someone is about to get rugged or they just got incredibly lucky on a meme coin. the catch with this specific clawbot isn't a hidden fee or a scam but a fundamental shift in how we actually build trading systems using autonomous agents. if you have been stuck manually drawing lines on a chart while the market moves at light speed you are already losing a game you don't even realize is being played
i am currently sitting here with my coffee while a sub agent named cracker digs through my entire github repo to find every winning strategy i have ever coded. most people struggle to find one good entry signal but i just told this agent to analyze my winners and generate ten new vwap variations based on the data. the first loop of this system is how i even run this much power without my laptop melting into a puddle of aluminum
i am actually tapped into a completely different computer right now using chrome remote desktop which is a total game changer for automation. it allows me to keep my main machine clean while the heavy lifting happens on a dedicated rig that i can access from anywhere in the world. this setup solves the hardware bottleneck because you don't need a five thousand dollar macbook to run these models if you know how to bridge the gap properly
this is where things get interesting because once the agent has access to the codebase it starts spawning its own sub agents to handle specific tasks like backtesting and data analysis. i am watching crabwalk right now which is a tool that shows exactly what the ai is thinking as it researches my strategy files. it found a comprehensive understanding of my winning setups and is now launching parallel agents to calculate the sortino and sharp ratios for me
you have to realize that code is the great equalizer because it takes the emotion out of a trade that would normally make you freeze up. i spent hundreds of thousands on developers in the past thinking i couldn't code myself until i realized that automation was the only way to stop getting liquidated. now i just iterate live on youtube and let the bots do the heavy lifting while i focus on the high level logic and strategy
most traders are flying blind because they don't have access to institutional grade data without paying a massive monthly tax. i built a direct node to hyperliquid that lets me see every single liquidation level and tick in real time which is usually a thousand dollar a month service. i give this api key out for free in my streams because i want my community to have the same edge as the big banks that are trying to rinse them
the second loop closes when you realize that you don't even need to own the hardware to run the most massive models in existence. i use lambda labs to deploy monster machines in about two seconds whenever i have a project that needs more horsepower than a standard desktop can provide. it is much cheaper to rent a high end gpu by the hour than it is to buy one and watch it become obsolete in six months while you are still trying to find alpha
a lot of people ask me why i don't just run a pump fun token or do sandwich attacks to make quick money. the truth is that i keep everything on the white hat side because the vibe you put out into the market is exactly what you get back. i am looking at a thirty one billion dollar opportunity in the long run so i am not going to risk my reputation or my freedom for a few racks of dirty money
the agent just finished its research and is asking me for the github token because it ran into a four oh four error on a private repo. even the most advanced ai agents need a human to guide the ship and provide the right credentials when things get sticky. i had to explain that the local copy was already synced via dropbox so it didn't need to stress about the cli tools or ssh keys right now
it is insane to t[...]
The 7,547% Trading Bot Catch: Why Pro Developers Are Ditching Macs for This $15 Architecture
seeing a seven thousand five hundred percent return usually means someone is about to get rugged or they just got incredibly lucky on a meme coin. the catch with this specific clawbot isn't a hidden fee or a scam but a fundamental shift in how we actually build trading systems using autonomous agents. if you have been stuck manually drawing lines on a chart while the market moves at light speed you are already losing a game you don't even realize is being played
i am currently sitting here with my coffee while a sub agent named cracker digs through my entire github repo to find every winning strategy i have ever coded. most people struggle to find one good entry signal but i just told this agent to analyze my winners and generate ten new vwap variations based on the data. the first loop of this system is how i even run this much power without my laptop melting into a puddle of aluminum
i am actually tapped into a completely different computer right now using chrome remote desktop which is a total game changer for automation. it allows me to keep my main machine clean while the heavy lifting happens on a dedicated rig that i can access from anywhere in the world. this setup solves the hardware bottleneck because you don't need a five thousand dollar macbook to run these models if you know how to bridge the gap properly
this is where things get interesting because once the agent has access to the codebase it starts spawning its own sub agents to handle specific tasks like backtesting and data analysis. i am watching crabwalk right now which is a tool that shows exactly what the ai is thinking as it researches my strategy files. it found a comprehensive understanding of my winning setups and is now launching parallel agents to calculate the sortino and sharp ratios for me
you have to realize that code is the great equalizer because it takes the emotion out of a trade that would normally make you freeze up. i spent hundreds of thousands on developers in the past thinking i couldn't code myself until i realized that automation was the only way to stop getting liquidated. now i just iterate live on youtube and let the bots do the heavy lifting while i focus on the high level logic and strategy
most traders are flying blind because they don't have access to institutional grade data without paying a massive monthly tax. i built a direct node to hyperliquid that lets me see every single liquidation level and tick in real time which is usually a thousand dollar a month service. i give this api key out for free in my streams because i want my community to have the same edge as the big banks that are trying to rinse them
the second loop closes when you realize that you don't even need to own the hardware to run the most massive models in existence. i use lambda labs to deploy monster machines in about two seconds whenever i have a project that needs more horsepower than a standard desktop can provide. it is much cheaper to rent a high end gpu by the hour than it is to buy one and watch it become obsolete in six months while you are still trying to find alpha
a lot of people ask me why i don't just run a pump fun token or do sandwich attacks to make quick money. the truth is that i keep everything on the white hat side because the vibe you put out into the market is exactly what you get back. i am looking at a thirty one billion dollar opportunity in the long run so i am not going to risk my reputation or my freedom for a few racks of dirty money
the agent just finished its research and is asking me for the github token because it ran into a four oh four error on a private repo. even the most advanced ai agents need a human to guide the ship and provide the right credentials when things get sticky. i had to explain that the local copy was already synced via dropbox so it didn't need to stress about the cli tools or ssh keys right now
it is insane to t[...]
Offshore
Moon Dev The 7,547% Trading Bot Catch: Why Pro Developers Are Ditching Macs for This $15 Architecture seeing a seven thousand five hundred percent return usually means someone is about to get rugged or they just got incredibly lucky on a meme coin. the catch…
hink that just yesterday i thought i was totally cooked because i couldn't get the whisper voice commands to work on this new profile. it turns out i was just being slow and forgot to install the proper drivers on the remote machine. once that was fixed i could literally just talk to my trading bot while sitting on the couch and have it spin up twenty new backtests without touching the keyboard
if you want to win in this game you have to be willing to iterate to success and embrace the failures as part of the data set. i have been up for three weeks straight instead of blowing up because i stopped over trading and started trusting the automated systems i built. the universe tends to get out of your way when you make a non negotiable contract with yourself to see the process through to the end
we are entering an era where everyone can essentially be their own cto if they just learn how to orchestrate these coding agents properly. tool like vive kanban are launching to help humans and agents collaborate without the terminal logs becoming a total mess of confusion. it allows you to focus on the planning and the quality of the trades instead of getting bogged down in the syntax of the language
i remember shorting bitcoin at ninety six thousand five hundred when everyone else was screaming that it was going to a hundred thousand. they were tearing me apart in the comments but the data showed the exhaustion and the automated system held the line. that single trade would be worth nearly eight hundred thousand dollars right now if i had just let it run which proves that the logic beats the hype every time
staying in a good vibe and focusing on the seven seven seven action is the only way to stay sane in this high pressure environment. i don't need protection from big dogs because i am the data dog and i protect my own systems with the code i write every day. the path to a fully automated life isn't about having the flashiest gear but about having the discipline to keep building when everyone else is distracted by the noise
tweet
if you want to win in this game you have to be willing to iterate to success and embrace the failures as part of the data set. i have been up for three weeks straight instead of blowing up because i stopped over trading and started trusting the automated systems i built. the universe tends to get out of your way when you make a non negotiable contract with yourself to see the process through to the end
we are entering an era where everyone can essentially be their own cto if they just learn how to orchestrate these coding agents properly. tool like vive kanban are launching to help humans and agents collaborate without the terminal logs becoming a total mess of confusion. it allows you to focus on the planning and the quality of the trades instead of getting bogged down in the syntax of the language
i remember shorting bitcoin at ninety six thousand five hundred when everyone else was screaming that it was going to a hundred thousand. they were tearing me apart in the comments but the data showed the exhaustion and the automated system held the line. that single trade would be worth nearly eight hundred thousand dollars right now if i had just let it run which proves that the logic beats the hype every time
staying in a good vibe and focusing on the seven seven seven action is the only way to stay sane in this high pressure environment. i don't need protection from big dogs because i am the data dog and i protect my own systems with the code i write every day. the path to a fully automated life isn't about having the flashiest gear but about having the discipline to keep building when everyone else is distracted by the noise
tweet