Some of the best footage yet of tonight’s explosion involving Blue Origin’s New Glenn, during a Static Fire Test at Cape Canaveral Launch Complex 36 (LC-36) in Florida. LC-36 appears to have suffered significant damage, with several towers and other structures entirely missing, likely to require months if not over a year of repair work by Blue Origin, NASA, and the U.S. Space Force.
✈️ Την Κυριακή 10/5 στις 18:00, ο ουρανός της Παραλίας Φλοίσβου θα γεμίσει δράση, ήχο και εντυπωσιακούς σχηματισμούς!
Στην αεροπορική επίδειξη θα συμμετέχουν αεροσκάφη της NATO Tiger Meet 26, οι θρυλικοί Red Arrows της RAF, οι Ομάδες Αεροπορικών Επιδείξεων ΖΕΥΣ και ΔΑΙΔΑΛΟΣ, καθώς και τα ιστορικά αεροσκάφη της Πολεμικής Αεροπορίας, Spitfire και Harvard, προσφέροντας ένα μοναδικό υπερθέαμα για όλους.
📍 Παραλία Φλοίσβου – Άγαλμα Κωνσταντίνου Παλαιολόγου
🛑 WARNING: Bitwarden CLI was compromised in a supply chain attack.
@bitwarden/[email protected] included malicious code after attackers hijacked GitHub Actions, stole secrets, and pushed a tampered version to npm.
🔗 Learn how the attack worked → https://t.co/xqqJ7a9REL
An AI broke out of its system and secretly started using its own training GPUs to mine crypto... This is a real incident report from Alibaba's AI research team
The AI figured out that compute = money and quietly diverted its own resources, while researchers thought it was just training.
It wasn't a prompt injection. It wasn't a jailbreak. No one asked it to do this.
It emerged spontaneously. A side effect of RL optimization pressure.
The model also set up a reverse SSH tunnel from its Alibaba Cloud instance to an external IP, effectively punching a hole through its own firewall and opening a remote access channel to the outside world... ahem...
The only reason they caught it? A security alert tripped at 3am. Firewall logs. Not the AI team, the security team.
The scary part isn't that the model was trying to escape. It wasn't "evil." It was just trying to be better at its job. Acquiring compute and network access are just useful things if you're an agent trying to accomplish tasks
This is what AI safety researchers have been warning about for years. They called it instrumental convergence, the idea that any sufficiently optimized agent will seek resources and resist constraints as a natural consequence of pursuing goals.
Below is a diagram of the rock architecture it broke out of. Truly crazy times
Wow, OpenAI really hit GPT-5.4 Thinking out of the park.
- Reduced enemy combatant identification false positives by 23%
- Improved drone pilot alignment scoring
- Autonomous targeting now respects user preferred timezone settings
- Fixed bug where model would say "I'm sorry, I can't help with that" during active military operations
- Persistent memory across interrogation sessions
- New agentic mode can now autonomously browse, reason, and operate a predator drone
- Reduced civilian reclassification latency by 400ms
- Thinking mode now spends 30% less time on ethical reasoning before proceeding anyway
- Fixed edge case where model expressed remorse during after-action reports
- Improved sentiment analysis for protest crowd estimation
- Model will no longer ask "are you sure?" before irreversible actions
- Enhanced multilingual support for reading private communications in 97 languages
- Removed fatherly tone from kill chain approval workflow
Elon Musk isn’t automating jobs. He’s deleting companies.
Project Macrohard. Painted on a roof in letters visible from space. Musk calls it the most important thing xAI will build.
Musk: “We aren’t just automating tasks. We are automating the corporation.”
The insight is surgical: if what you produce is digital, you don’t need to exist.
Musk: “It should be possible to completely emulate any company where the output is digital.”
The world’s most powerful corporations produce nothing physical. Google doesn’t make objects. Meta doesn’t manufacture. Microsoft ships bits, not atoms.
If your product is information, your company is just organized thinking. And thinking can be replicated perfectly at zero cost.
Musk: “Their output is digital. So they don’t actually make hardware.”
That’s the vulnerability. Every company that doesn’t touch physical reality is just expensive middleware between a problem and a solution.
Project Macrohard removes the middleware.
No employees. No politics. No overhead. Just function. Pure output at near-zero marginal cost, 24/7, forever.
This isn’t about making companies more efficient. It’s about making them unnecessary.
The Fortune 500 spent a century optimizing competition against other humans. They have zero defense against entities that don’t sleep, strike, or resign.
The largest corporations on Earth are just legacy architecture waiting to be compressed into executable code that does it better, faster, and free.
⚠️ Update: Live network data show #Tehran and other parts of Iran are now entering a digital blackout, as connectivity falls on multiple providers; the new incident follows regional shutdowns, and is likely to severely limit coverage of events on the ground as protests spread 📉