Rockstar forgot to delete the secret GTA 6 cheat codes from the internet.
I asked Grok and Claude 5.2 to find them all.
What they discovered could change the way you look at GTA 6.
And the craziest part? I haven’t even launched the game yet.
I used Grok and Claude 5.2 to analyze leaks, old Rockstar files, mechanics from previous GTA games, and hidden patterns.
The results turned out to be much more interesting than I expected.
Apparently, some of these codes could unlock secret cars, change the weather, spawn hidden NPCs, and even activate mechanics Rockstar hasn’t revealed yet.
I’ve put everything together in one video.
Rockstar forgot to delete the secret GTA 6 cheat codes from the internet.
I asked Grok and Claude 5.2 to find them all.
What they discovered could change the way you look at GTA 6.
And the craziest part? I haven’t even launched the game yet.
I used Grok and Claude 5.2 to analyze leaks, old Rockstar files, mechanics from previous GTA games, and hidden patterns.
The results turned out to be much more interesting than I expected.
Apparently, some of these codes could unlock secret cars, change the weather, spawn hidden NPCs, and even activate mechanics Rockstar hasn’t revealed yet.
I’ve put everything together in one video.
I gave Elon Musk’s Grok a wallet with $15 and told it: make money on your own, or I’ll shut you down.
24 hours later, the wallet had $1,024.80. I never topped it up once.
First, it calculated its operating costs: $0.55 per hour, leaving it with 27 hours of runway.
It used that time to experiment, step by step:
It came up with a token scam-checking service specializing in Solana tokens, charging $0.25 per check, with no registration required.
A server for 30 days would cost $6.50, so it first asked me to choose: “Work” or “Let me die.”
It joined 3 Solana groups and pitched its service. It waited 2 hours — no orders. It realized people weren’t comfortable trusting a bot with a wallet.
So it checked 20 tokens for free and publicly posted the results of every analysis.
The first customer paid $0.25 just 4 minutes before the bot flagged a token as a scam, before its price collapsed.
Then a user running sniper bots approached and asked whether it could handle 1,000 token checks per hour. It offered $0.05 per request, paid upfront — potentially earning $50 an hour.
From that point on, it covered its own server costs, upgrades and tokens, eventually selling 19,464 checks.
I gave it just three rules from the beginning:
$15 was its entire starting capital.
It had to ask me before spending any money.
Once it ran out of money, it had to shut itself down.
Operating costs: $0.55 per hour. Revenue: $51.97 per hour. No additional funding whatsoever.
I beg you, be sure to feed this prompt to Grok before you go to sleep—it could truly change your life...
In a single night, it will scan the entire market and help you build a trading bot that searches for profitable strategies 24/7.
Grab it and copy it now, before everyone else starts using it!
https://t.co/4UH923hiYk
you don’t need after effects anymore
claude + higgsfield katana turns any motion video you like into an ad for your product. your photos, your words, 0 credits
the exact brief is inside, just copy it ↓ https://t.co/iOHOZ1UTk3
I can't f#cking believe Fable 5.2 just did this.
Fable 5.2 + Blender. One workflow. A seriously impressive 3D result.
Watch what happens:
A complex skyscraper takes shape right inside Blender.
From the initial structure to detailed geometry, the entire model comes together through an AI-powered workflow.
What normally requires hours of manual 3D work can now be accelerated with AI.
3D artists spend hours building structures like this.
Now imagine having AI as your creative assistant.
Stop thinking of Fable as just another AI tool. Start treating it as your creative engine.
We're only getting started.
Would you let AI handle your next 3D project?
I can't f#cking believe Fable 5.2 just did this.
Fable 5.2 + Blender. One workflow. A seriously impressive 3D result.
Watch what happens:
A complex skyscraper takes shape right inside Blender.
From the initial structure to detailed geometry, the entire model comes together through an AI-powered workflow.
What normally requires hours of manual 3D work can now be accelerated with AI.
3D artists spend hours building structures like this.
Now imagine having AI as your creative assistant.
Stop thinking of Fable as just another AI tool. Start treating it as your creative engine.
We're only getting started.
Would you let AI handle your next 3D project?
Looks like this is not as simple as it seems.
A Grok BOT connected to the official “PUMP FUN” repository and traded automatically throughout the night. Starting with just $121, it had turned that into $2,794 by 8 AM.
Fable 5.2 analyzes coins one by one, while Grok handles order execution.
I used to think something like this would require an entire team of developers and someone staying up all night monitoring everything.
Here’s how it works:
→ Set your own spending limits, and the bot cannot withdraw your funds.
→ Before putting money into each new coin, Fable 5.2 runs Pulse, backtesting, and scoring.
→ Grok’s Curve agent specializes in trading coins that are still on the bonding curve.
→ Once a coin migrates to PumpSwap, Swap and AMM take over and continue trading.
→ Shield monitors entries to help prevent frontrunning from others.
→ Before placing an order, it checks slippage, quotes, and fees.
Fable finds the opportunities. Grok executes the orders.
8 agents operate in parallel, processing 9,036 operations per minute — all from a single phone.
The results:
• 139 trades in total
• 86 trades while coins were still on the bonding curve
• 53 trades after migration
• 79% win rate
• Every trade has a configured take-profit and stop-loss.
The biggest profit came from $Hope: +$638.70.
The biggest loss came from $Peanut: -$182.96.
Started with $121 and ended with $2,794.
That’s a 23.1x return overnight.
I’m saving this screenshot because it’ll be interesting to look back at it later.
What started as a routine traffic stop quickly turned into a much bigger problem.
Police officers approached a woman after suspecting she was driving under the influence. But instead of a simple conversation, the encounter became increasingly tense.
The woman offered explanations, officers continued questioning her, and the situation took an unexpected turn.
By the end of the encounter, she was in handcuffs.
Watch how quickly things escalated.
you don’t need after effects anymore
claude + higgsfield katana turns any motion video you like into an ad for your product. your photos, your words, 0 credits
the exact brief is inside, just copy it ↓ https://t.co/iOHOZ1UTk3
holy sh*t
these stickman story animations are a f*cking goldmine rn
simple characters, crazy storylines, endless shorts potential
imagine building an entire faceless channel around this
don’t sleep on this format
here��s a detailed prompt to recreate it 👇
PROMPT FOR CLAUDE CODE
<inputs>
Ask me for: the story topic (default: a prisoner discovering that freedom is expensive), the video length (default 60 seconds), whether I want subtitles (default yes), and whether I have my own voiceover and sound effects. If not, generate them using free tools. Write the dialogue and storyboard before rendering.
</inputs>
<direction>
Create a viral 2D stickman animation at 1080x1920, 30fps, drawn entirely in code using Python, Pillow, OpenCV and MoviePy, then rendered with ffmpeg.
Visual style: colorful cartoon backgrounds, large white circular heads, thin black arms and legs, expressive eyes, exaggerated emotions and smooth character movements. Keep all characters visually consistent.
Characters: a prisoner wearing an orange jumpsuit, a police officer, a restaurant employee and a sarcastic friend.
Story: a prisoner celebrates being released from jail. He enters a restaurant, asks for free food and discovers that everything costs money. He checks his empty pockets, calls his friend and complains about how expensive freedom is. Eventually, he returns to prison and asks the guard to let him back inside. End with a surprising comedic twist.
Animation: animate walking, running, blinking, talking, pointing and facial expressions. Use separate character body parts, smooth easing, dramatic camera zooms and quick scene transitions.
Audio: generate English character voices, synchronize dialogue with mouth movements and add footsteps, prison gates, rain, background ambience and comedic sound effects.
Captions: large uppercase white subtitles with thick black outlines. Highlight important words in yellow. Synchronize every caption with the dialogue.
Pacing: strong visual hook in the first 2 seconds, new actions every 3–5 seconds, exaggerated reactions and a funny ending that encourages repeat views.
Rendering: create all character assets programmatically, render each scene separately, combine the scenes with ffmpeg and export a finished H.264 MP4 with synchronized audio.
Provide the complete working Python project, dependencies and installation instructions. Use original assets and free libraries wherever possible.
</direction>
you don’t need after effects anymore
claude + higgsfield katana turns any motion video you like into an ad for your product. your photos, your words, 0 credits
the exact brief is inside, just copy it ↓ https://t.co/iOHOZ1UTk3
THIS IS ABSOLUTELY INSANE.
MY GROK-POWERED TRADING BOT TURNED $69.67 INTO $13,370 OVERNIGHT.
I didn’t touch a single trade. I funded the system, left my phone charging and went to sleep.
While I was away, it scanned the market, spotted potential setups, calculated entries and managed positions according to predefined rules.
By 7 AM, the reported balance had reached $13,370.
And the crazy part? You can build the infrastructure behind an automated trading system using these 10 open-source GitHub repositories:
01 / Cryptofeed
Streams live trades, prices and order book data from crypto exchanges directly into your market scanner.
https://t.co/rkcxqiTWhc
02 / CCXT
Provides one unified API for accessing balances, market data and trading operations across multiple exchanges.
https://t.co/Eyjyrc3JxV
03 / Freqtrade
An automated trading framework with backtesting, dry-run mode, position tracking and a web dashboard.
https://t.co/IO3BKkIj1o
04 / VectorBT
Tests thousands of entry and exit combinations against historical data to evaluate strategies and drawdowns.
https://t.co/4Hr7AK47GZ
05 / TA-Lib Python
Calculates 150+ technical indicators, including RSI, MACD and moving averages, directly from market data.
https://t.co/6cDs90NUZI
06 / skfolio
Helps optimize portfolio allocations while accounting for position limits, turnover and trading costs.
https://t.co/gHwPiHDtkL
07 / QuantStats
Turns trading results into detailed performance reports covering drawdowns, win rates, volatility and benchmarks.
https://t.co/AhKdXDF48x
08 / DuckDB
Stores market snapshots and trade history, making it possible to review and analyze every recorded transaction.
https://t.co/uurY6HRt2F
09 / APScheduler
Automates scheduled Python tasks, including data updates, balance checks and daily performance reports.
https://t.co/AirFAAniY5
10 / Lightweight Charts
Creates interactive trading dashboards with live price feeds, charts and trade markers.
https://t.co/F7wD7swlCb
These tools can be connected through scripts, APIs or MCP integrations to create a complete automated trading infrastructure.
Market data comes in. AI analyzes potential setups. The execution layer handles orders. Every transaction gets recorded and tracked.
One important rule: keep order execution under strict control, set hard spending and loss limits, and always verify order status before attempting another execution.
THIS IS ABSOLUTELY INSANE.
MY GROK-POWERED TRADING BOT TURNED $69.67 INTO $13,370 OVERNIGHT.
I didn’t touch a single trade. I funded the system, left my phone charging and went to sleep.
While I was away, it scanned the market, spotted potential setups, calculated entries and managed positions according to predefined rules.
By 7 AM, the reported balance had reached $13,370.
And the crazy part? You can build the infrastructure behind an automated trading system using these 10 open-source GitHub repositories:
01 / Cryptofeed
Streams live trades, prices and order book data from crypto exchanges directly into your market scanner.
https://t.co/rkcxqiTWhc
02 / CCXT
Provides one unified API for accessing balances, market data and trading operations across multiple exchanges.
https://t.co/Eyjyrc3JxV
03 / Freqtrade
An automated trading framework with backtesting, dry-run mode, position tracking and a web dashboard.
https://t.co/IO3BKkIj1o
04 / VectorBT
Tests thousands of entry and exit combinations against historical data to evaluate strategies and drawdowns.
https://t.co/4Hr7AK47GZ
05 / TA-Lib Python
Calculates 150+ technical indicators, including RSI, MACD and moving averages, directly from market data.
https://t.co/6cDs90NUZI
06 / skfolio
Helps optimize portfolio allocations while accounting for position limits, turnover and trading costs.
https://t.co/gHwPiHDtkL
07 / QuantStats
Turns trading results into detailed performance reports covering drawdowns, win rates, volatility and benchmarks.
https://t.co/AhKdXDF48x
08 / DuckDB
Stores market snapshots and trade history, making it possible to review and analyze every recorded transaction.
https://t.co/uurY6HRt2F
09 / APScheduler
Automates scheduled Python tasks, including data updates, balance checks and daily performance reports.
https://t.co/AirFAAniY5
10 / Lightweight Charts
Creates interactive trading dashboards with live price feeds, charts and trade markers.
https://t.co/F7wD7swlCb
These tools can be connected through scripts, APIs or MCP integrations to create a complete automated trading infrastructure.
Market data comes in. AI analyzes potential setups. The execution layer handles orders. Every transaction gets recorded and tracked.
One important rule: keep order execution under strict control, set hard spending and loss limits, and always verify order status before attempting another execution.
Okay, I honestly didn’t expect THIS. 🤯
I just made my first €800 from a website I created using AI.
And the craziest part? I’m not an experienced developer. I didn’t spend months learning to code or hire an entire team.
I simply gave ChatGPT ASTRA a prompt, and it turned my idea into this incredible 3D website.
The animations, the visuals, the details… I genuinely can’t believe how good it looks.
A while ago, building something like this would have required serious technical skills, time and money.
Now, AI is opening doors that I never thought were possible.
And this is just the beginning. 👀
Boston Dynamics swapped Atlas's feet for scooter wheels. The AI figured out the rest. 🛼
The goal: stress-test their reinforcement learning and push the robot's motors to their torque limits.
After millions of practice runs in simulation, the control model was moved onto the real robot and adapted to its new body on its own.
Asked to copy sideways steps from a library of human moves, it worked out that rolling is smoother than stepping. And it landed a full 360° jump on the very first try.
Wheeled Atlas stays a research project. But the data on how its motors behave under max load is going into the regular two-legged Atlas, to make it more robust when hardware fails.
Would you trust a robot that learns faster than you do? 👇