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.
Built around Claude 5.2, this AI-powered system is taking robot training to a whole new level.
Instead of learning everything through trial and error in the real world, robots can undergo up to 10,000 training runs inside virtual simulations, refining their movements, adapting to different scenarios and learning to execute complex tasks.
Once that training is complete, they move into real-world testing — and the results are becoming increasingly impressive.
What once required countless physical experiments can now be simulated thousands of times before a robot even touches the real environment.
The Biggest Banking Sector Outage in the US
A few hours ago, customers began experiencing issues logging into banking apps, making transactions, and viewing their account balances.
Meanwhile, stocks of a large number of S&P 500 companies have already started taking a hit.
Normally, this would just be another major outage. But there’s one interesting detail: a couple of months ago, rumors began circulating online about a wave of massive cyberattacks supposedly set to begin on November 17.
And the banking sector was specifically predicted to be the first target.
What do you think about this? Coincidence, or something more?
🚀 FROM ORBIT STRAIGHT INTO THE TOWER’S CHOPSTICKS
SpaceX just made history.
Until November 14, the company had successfully caught three Super Heavy boosters — but never a Starship.
Now, that has changed.
The upper stage of Starship, Ship 42, has returned from orbit and been caught by the mechanical arms of the Mechazilla launch tower at Starbase, Texas.
The catch happened at 21:20 UTC, approximately 22.5 hours after the launch of Flight 15.
But the original plan was even more ambitious: catch BOTH stages of the fully stacked rocket.
That didn't quite happen.
A low-fuel warning forced the Super Heavy booster to divert toward the Gulf of Mexico, where it splashed down instead of returning to the tower.
WHY IS CATCHING STARSHIP SO DIFFICULT?
Unlike the booster, the upper stage travels much farther, climbs higher, and reaches orbital velocity.
Returning from orbit means shedding an enormous amount of speed while passing through the atmosphere — literally flying through fire.
That's why SpaceX considers this achievement a critical missing piece on the road to a fully reusable rocket.
Before landing, Ship 42 also underwent a unique inspection.
Its heat shield was photographed by Starlink satellites that the spacecraft itself had deployed into orbit.
Ground-based cameras provided additional footage, and the results showed that the heat shield was in good condition.
26 NEW STARLINK V3 SATELLITES
Ship 42 deployed 26 next-generation Starlink V3 satellites.
The same number was launched during Flight 14 on September 28, when Starship first reached orbit.
That's 52 satellites deployed across two flights — the beginning of a constellation SpaceX plans to expand to 100,000 satellites.
For comparison, the current Starlink network consists of approximately 11,000 satellites.
And Elon Musk has even bigger ambitions: using Starship to deploy Starmind, a proposed orbital AI data-center network consisting of one million satellites.
WHY THE CATCH MATTERS
Building megaconstellations requires an unprecedented launch frequency.
That's where Mechazilla's giant mechanical arms become essential.
A caught rocket stage can be inspected, serviced, and prepared for another flight much faster than one recovered from the ocean.
The faster the turnaround, the more frequently SpaceX can launch.
And that's the real goal:
Not just building a rocket that can reach orbit.
But building one that can return, get caught, and fly again.
🚀 The era of fully reusable orbital rockets is getting closer.
🚨 Was AI behind the nationwide US flight shutdown?
DHS is investigating whether SMART, the FAA's new AI traffic tool, played a role in the Nov 13 ground stop that froze departures across the US.
✈️ What happened
• All takeoffs halted at 3:54 PM ET. Planes in the air were allowed to land
• Cause per officials: comms failure across all 22 air traffic control centers
• 1,913 flights canceled, 10,494 delayed (FlightAware)
• No signs of a cyberattack so far, CISA is on it
🤖 About SMART
• Cloud AI platform merging 200+ data streams: weather, routes, sector load, controller shifts
• Suggests schedule and route changes. Humans still make the calls
• Rolled out to all 22 centers on Nov 10, just 3 days before the outage
• $875M, 12-year contract with Air Space Intelligence
• In October, 20+ members of Congress asked to pause the rollout over safety concerns
The developer says the incident "was not based on the SMART system."
Only the 3rd nationwide ground stop ever, after 9/11 and the 2023 NOTAM failure.
Coincidence? 🤔
🚨 Was AI behind the nationwide US flight shutdown?
DHS is investigating whether SMART, the FAA's new AI traffic tool, played a role in the Nov 13 ground stop that froze departures across the US.
✈️ What happened
• All takeoffs halted at 3:54 PM ET. Planes in the air were allowed to land
• Cause per officials: comms failure across all 22 air traffic control centers
• 1,913 flights canceled, 10,494 delayed (FlightAware)
• No signs of a cyberattack so far, CISA is on it
🤖 About SMART
• Cloud AI platform merging 200+ data streams: weather, routes, sector load, controller shifts
• Suggests schedule and route changes. Humans still make the calls
• Rolled out to all 22 centers on Nov 10, just 3 days before the outage
• $875M, 12-year contract with Air Space Intelligence
• In October, 20+ members of Congress asked to pause the rollout over safety concerns
The developer says the incident "was not based on the SMART system."
Only the 3rd nationwide ground stop ever, after 9/11 and the 2023 NOTAM failure.
Coincidence? 🤔
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
The entire internet is obsessed with how much old Hugh supposedly “slept,” but what should really send chills down Wall Street’s spine isn’t the gossip about his personal life. It’s the three brutal truths exposed by this black swan event: $NVDA
Truth No. 1: Why did the board remove the company’s only true creator in just 96 hours?
Hugh founded Nvidia 33 years ago, turning a small graphics-card workshop into a trillion-dollar empire. Yet he could be pushed out almost overnight.
The reason lies in Nvidia’s corporate structure. Unlike Meta, Nvidia has no dual-class share structure. Hugh Ren-Hsun owns only a single-digit percentage of the company, and six votes from its 11-member board are enough to remove the CEO.
Compare that with Zuckerberg, whose supervoting shares give him near-total control. A similar scandal at Meta would hardly put his position at risk.
What looks like transparent corporate governance in good times can become a founder’s greatest weakness when things go wrong.
Truth No. 2: China’s AI ambitions could take the first hit.
While everyone is guessing who will take over (and the rumor about Jay Puri leaving in June is fake, with no corresponding 8-K filing), the bigger issue is unfolding in Washington.
For years, Hugh personally lobbied Washington to keep Nvidia’s chip sales to China alive despite enormous political pressure.
Now he is gone, while Congress is pushing tougher restrictions on advanced AI chip exports through this year’s NDAA.
Interim CFO Colette Kress won’t fill that lobbying gap, and a new CEO may be even less willing to take the political risk.
My view: the window for Nvidia’s advanced chips entering China could soon close permanently.
Truth No. 3: Nvidia could be heading toward its own Intel moment.
Two major American companies replaced CEOs following workplace-related scandals, with very different outcomes:
▪️ McDonald’s fired Steve Easterbrook in 2019. Business continued, and the stock kept climbing.
▪️ Intel fired Brian Krzanich in 2018, leaving its CFO in charge temporarily. The company lost its technological momentum and missed major opportunities in mobile and AI.
Nvidia is even more dependent on technological leadership.
On November 17, Nvidia reports Q3 earnings. Its first major earnings event without Hugh.
Wall Street will want answers:
• Who will lead the company technologically?
• Will Nvidia lower its next-quarter guidance?
• Could major customers like Microsoft and Meta start reconsidering their positions?
Some hope Hugh will return, like Steve Jobs in 1997 or Sam Altman last year.
Don’t count on it.
Reuters reports that many employees want him back. He remains a major shareholder, but the national security investigation makes this situation fundamentally different.
The leather jacket is put away. The era when one man shaped the direction of global computing may be over.
November 17. Buckle up.
@successful889 Fair question. I originally came across this on a Chinese-language page and adapted the post for my audience. Unfortunately, I didn’t save the original link, so I can’t point you to the exact source. If I manage to find it again, I’ll share it here.
The entire internet is obsessed with how much old Hugh supposedly “slept,” but what should really send chills down Wall Street’s spine isn’t the gossip about his personal life. It’s the three brutal truths exposed by this black swan event: $NVDA
Truth No. 1: Why did the board remove the company’s only true creator in just 96 hours?
Hugh founded Nvidia 33 years ago, turning a small graphics-card workshop into a trillion-dollar empire. Yet he could be pushed out almost overnight.
The reason lies in Nvidia’s corporate structure. Unlike Meta, Nvidia has no dual-class share structure. Hugh Ren-Hsun owns only a single-digit percentage of the company, and six votes from its 11-member board are enough to remove the CEO.
Compare that with Zuckerberg, whose supervoting shares give him near-total control. A similar scandal at Meta would hardly put his position at risk.
What looks like transparent corporate governance in good times can become a founder’s greatest weakness when things go wrong.
Truth No. 2: China’s AI ambitions could take the first hit.
While everyone is guessing who will take over (and the rumor about Jay Puri leaving in June is fake, with no corresponding 8-K filing), the bigger issue is unfolding in Washington.
For years, Hugh personally lobbied Washington to keep Nvidia’s chip sales to China alive despite enormous political pressure.
Now he is gone, while Congress is pushing tougher restrictions on advanced AI chip exports through this year’s NDAA.
Interim CFO Colette Kress won’t fill that lobbying gap, and a new CEO may be even less willing to take the political risk.
My view: the window for Nvidia’s advanced chips entering China could soon close permanently.
Truth No. 3: Nvidia could be heading toward its own Intel moment.
Two major American companies replaced CEOs following workplace-related scandals, with very different outcomes:
▪️ McDonald’s fired Steve Easterbrook in 2019. Business continued, and the stock kept climbing.
▪️ Intel fired Brian Krzanich in 2018, leaving its CFO in charge temporarily. The company lost its technological momentum and missed major opportunities in mobile and AI.
Nvidia is even more dependent on technological leadership.
On November 17, Nvidia reports Q3 earnings. Its first major earnings event without Hugh.
Wall Street will want answers:
• Who will lead the company technologically?
• Will Nvidia lower its next-quarter guidance?
• Could major customers like Microsoft and Meta start reconsidering their positions?
Some hope Hugh will return, like Steve Jobs in 1997 or Sam Altman last year.
Don’t count on it.
Reuters reports that many employees want him back. He remains a major shareholder, but the national security investigation makes this situation fundamentally different.
The leather jacket is put away. The era when one man shaped the direction of global computing may be over.
November 17. Buckle up.
A simple police encounter. A suspicious situation. And a discovery that changed everything.
Officers approached a woman after concerns were raised about a possible concealed firearm.
What followed was far from an ordinary interaction. Questions were asked, suspicions grew, and the encounter eventually ended with the woman being taken into custody.
But here’s what makes this footage particularly interesting: how quickly an everyday situation can turn into a serious police incident.
There’s always more to a story than what we see in a few minutes of footage.
What do you think — did the officers handle this situation the right way?
✔️ Google DeepMind unveils Gemini Robotics 2.5 and tests it at Mercedes-Benz factories
Google DeepMind has announced Gemini Robotics 2.5, an upgraded VLA model capable of autonomously controlling robotic hands and grippers using camera images and natural-language instructions.
Depending on the task, the new model achieves 30–100% higher success rates than previous versions while operating 20–30% faster.
The model was tested on 200 different tasks, ranging from handling pastries with tongs to tying bows on gift boxes and working with pipettes.
The results are impressive: Gemini Robotics 2.5 successfully packs pastries in 100% of attempts, compared with just 5% for Gemini Robotics 2. It closes toolboxes with a 93% success rate, compared with zero previously, and folds trousers successfully in 40% of attempts, a task its predecessor could not complete.
Gemini Robotics 2.5 is designed to work across different robotic platforms. Demonstrations included the Apollo 2 humanoid robot by Apptronik, equipped with a five-finger SharpaWave hand featuring 22 degrees of freedom, as well as the Franka Duo platform with two-finger grippers.
The technology was also tested at Mercedes-Benz production facilities in Kecskemét, Hungary, and Sindelfingen, Germany.
During the pilot, Apollo 2 sorted metal automotive brackets from shared containers into dedicated trays according to the assembly sequence. The robot successfully completed the task in 96% of cases.
Gemini Robotics ER handled task monitoring and reasoning, while Gemini Robotics 2.5 directly controlled the robot’s hands and manipulators. Google describes the experiment as a proof of concept for deploying robotics in real industrial environments.
DeepMind has also improved the system’s safety performance. The share of safely completed tasks increased by 33% when operating near obstacles and by 43% when working around people.
The company also expanded its open ASIMOV-Jailbreak benchmark with disguised hazardous requests. In these tests, Gemini 4 Argon achieved a 99.5% resistance rate, compared with 82% for Gemini Robotics ER 2.
During development, Argon also served as a simulator: it generated digital twins of real-world scenes from camera images, ran thousands of training scenarios, and analyzed videos of unsuccessful robotic attempts.
Early access to Gemini Robotics 2.5 is available through a waitlist.
Sources:
• Official Google DeepMind announcement
• Gemini Robotics 2.5 — Official model page
• ASIMOV Agentic — Hugging Face
• Trusted Tester early access application
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? 👇