Jakes

If you want to jump on this $10,000 campaign without paying for any agent. I've gat you but the access is limited and it's on FCFS. Platform here:simcluster-agentp.onrender.com Comment for access code.

I am blown away. Anthropic's unreleased Mythos model just jumped from 53% to 78% on SWE-bench Pro 🤯 Software Engineer has been solved. The only edge left is knowing what to build and why it matters.

@basedotmeme @KibiBot @junijakes

Drop your Base/Farcaster name👇 🪂 from @KibiBot

@MoonDevOnYT 🤯💯

warshawski absolutely cooking in the arena 18 correct 1 hour btc predictions only 8 wrong

Claude Code Fixed Polymarket's Biggest Problem

Beyond Accuracy: The One Metric Every Quant Developer Needs To Scale Their Trading Bots most people think applying ai to trading is a magic money button but the actual results will punch you in the throat if you are not ready for the truth. i spent years losing money and getting liquidated before i realized that a bot that is 90 percent accurate can still leave you completely broke. if you want to know the one metric that actually matters when the machines take over keep reading because most gurus are hiding the real math from you i used to be that guy staring at charts until four in the morning just hoping the price would move my way. i would get that liquidation email as soon as i fell asleep and feel like the market was personally out to get me. it took spending hundreds of thousands on developers for apps that didn't work to realize that code is the only equalizer in this game the first loop you need to close is why your "perfect" strategy fails in the live market even when the backtest looks like a vertical line. most people find a strategy then dump all their capital into it because they have fomo and fear. they skip the rbi system which is research then backtest then implement with tiny size before ever scaling up when i first tried to apply machine learning like random forest models to my liquidation data i thought i hit the jackpot with 75 percent accuracy. the harsh truth is that my model was just a professional pessimist that learned to always predict the market would go down. because the data was mostly bearish the model just cheated to get a high score while missing every single profitable move to the upside this is what happens when you have a class imbalance disaster in your data set. if 74 percent of your data is down moves then the ai realizes it can look smart by never betting on a bull run. you end up with a model that is a "down only" predictor which is essentially a broken clock that is right most of the time but useless for making money most people think they can just throw 10,000 trees at a random forest model and it will magically find the alpha. i learned the hard way that after about 200 or 300 trees you hit diminishing returns and just start wasting your computer's life. the secret isn't more trees but better feature engineering and narrowing the noise down to the signals that actually move the needle i spent ten years in technology scared to code because i thought it was for geniuses who went to stanford or knew advanced math. the reality is that i just needed to lock in for a few hours a day to realize that i could build systems that trade for me while i sleep. once you learn to code you stop being a gambler and start being a data dog who actually understands the math of the market if you want to actually win you have to stop looking at accuracy and start looking at the area under the curve or auc metric. a model with high accuracy but low auc is just guessing based on the majority class and will leave you holding the bag. you need a minimum auc of point 65 or ideally point 7 before you ever even think about letting a bot touch your real capital the reason i show everything live is because the trading industry is the most secretive business in the world and nobody wants to share the real sauce. most people are out there selling "set and forget" bots that make 20 percent a day but those are all scams meant to take your money. if you don't build your own edge and iterate your systems every single day you will eventually get caught by someone who is jim simons ran up a net worth of over 30 billion dollars because he understood that you just have to make your systems better and better. he wasn't smarter than everyone else because of some magic crystal ball but because he treated trading like a science experiment. he used an rbi framework of researching and testing until the math was undeniable and the emotions were completely gone ambitious people often fail at trading because they think they can just work harder to get better results. in almost every other industry working more hours equals more money but in trading it just equals more overtrading and more mistakes. the only way to work harder in this game is to put more hours into your code and your research so the bot can do the heavy lifting for you we found that focusing on specific signals like btc liquidation cascades and feature reduction actually improved the model more than complex math ever could. by dropping the noise from 1,600 features down to just 6 key liquidation metrics the model started to actually catch the up moves it was missing before. it's about finding the signal in the static rather than trying to listen to everything at once you have to realize that you are chasing the same prize as firms like jane street and renaissance who have more computing power than you can imagine. the only way to compete is to use the same tools they use and stay persistent enough to outlast the gamblers. code is the great equalizer because it allows a regular person in a bedroom with a whiteboard to compete with wall street giants if you're still hand trading with 40x leverage and praying to the chart gods you are just waiting for the day your account hits zero. the real results of ai in trading show that it is a brutal game of iteration where you must be willing to fail until you find the parameters that stick. stay hungry and stay foolish because once you automate your trading your life changes forever and you never look back always remember that the 777 is near to the brokenhearted and that this journey is about persistence more than anything else. if you can't fly then run and if you can't run then walk but by all means keep moving toward your automated future. the math doesn't lie and once you have a system that works you can finally stop staring at the screen and start living your life on your own terms

@BloosBTC “Am” does not start a sentence 😂.

check if you’re eligible for PerleLabs Airdrop. link: register.perle.xyz Am eligible are you ?🥸

Long time ago

@manixcodes Mani editor 😇

VS Code isn’t the best editor. It’s just the best default. What’s actually the best editor in your opinion?

@piyush784066 Using 256 SSD (Windows) 513 HDD (Linux) Running on a dual boot system though 🫡

As a developer, how much STORAGE is enough for you? 1) 256 GB 2) 512 GB 3) 1 TB 4) 2+ TB

@sean_plus_one @khushiirl Not hard to detect a full-stack dev. PHP was my first backend language.

@khushiirl Am I allowed to Google things hahaha PHP / JS / SQL Python to a lesser extent / Ruby I can read, writing would take a little work.

@CoinbaseMarkets $Fight

Assets added to the roadmap today: Seeker (SKR), FIGHT (FIGHT) coinbase.com/blog/increasin…

@Eche_codes Yeah second project with Paystack API.

Most devs haven’t handled Payment integration in a website, have you ?

wild how a simple text to @HeyElsaAI can handle swaps, cross chain moves, portfolio check among others, all behind the scenes. the elsa text box is full defi commands, use it wisely. you’re missing the future of effortless defi actions, if you haven't tried it. just elsa it 🔥

My first programming languages? JavaScript My favourite: C#

@iyoushetwt JavaScript

What was the first programming language you learned ? Mine : C++

@notathrv My own Dey shock me 😀. That was in 2022. Never used Thinkpad again.

i have never heard a bad review about ThinkPad

@jake_063020 Congratulations bro.

First time tasting elonfi 🥹🥹🥹🥹(real this time) Finallyyyyyyyyyyyyy. Time to use part of this to buy a subscription for @Everlyn_ai and to also get more energy to go all in on @Infinit_Labs Thank you elon.

@CoalBoom Carnivorous being! @CoalBoom

All of a sudden Nnamdi Kanu is supreme leader of south east according to Cubana chief priest 🤣🤣🤣. If you do not see Tinubu written all over that tweet then I am sorry you are blind and you do not understand the game. Anyway see yr IPOb

@CoalBoom Always easy to dictate the unfortunates 😅. Funny you’re the leader of the unfortunate beings @CoalBoom 🫡