BREAKING: 🚨🚨🚨🚨🚨🚨🚨 Iran has just struck an oil tanker conducting an outbound transit of the Strait of Hormuz under US escort in the US-backed southern Omani corridor, with a fire and a blackout on board, per UKMTO. #Hormuz #US
#G7PlansToReleaseUpTo100MBarrelsOilDiesel ⚡️⚡️⚡️⚡️⚡️⚡️⚡️ G7 nations will release up to 100 million barrels of oil and diesel from strategic reserves over the next four months, Macron says.
According to the French president, the decision has already been made. Bloomberg reports that the Trump administration had pressured G7 countries to curb rising fuel prices.
The US threatened to restrict diesel exports if Europe did not increase supplies.
➡️ Trump welcomed the decision, stating that Europe holds large diesel reserves and that their release would begin immediately $TRUMP
🤣🤣🤣🤣🤣🤣🤣 Trump secretly consulted with the AI Grok before an operation to capture Maduro — Time
Elon Musk, the owner of Grok, visited the White House for a secret meeting with Trump. Trump began asking the AI questions about his presidency and political legacy, and also asked how Venezuelans would react if US troops were to capture their president.
The AI assured Trump that Maduro was a repressive and extremely unpopular dictator, and that the country's citizens would celebrate his capture. After the meeting, Trump described Grok as "brilliant." #TRUMP #Aİ $TRUMP
#NEARFallsToAround$4.70Down14% 🚨🚨🚨🚨🚨🚨🚨🚨🚨 NEAR Intents Loses $3.8 Million NEAR Intents suffered a hack resulting in a loss of approximately $3.8 million, leading to the temporary suspension of some services. 📊 The vulnerability lay within the cross-chain infrastructure, and the team has already pledged to fully reimburse affected users. The main NEAR Protocol blockchain was not compromised. #NEAR🚀🚀🚀 $NEAR
JUST IN : 🚨🚨🚨🚨🚨🚨🚨 Hyperliquid Labs will unstake $320M in $HYPE today for its monthly team unlock, with the tokens to be sold OTC to an institution. #hype #Hyperliquid
🤯🤯🤯🤯🤯🤯🤯 Is a "rise of the machines" imminent? Anthropic warns of existential AI risks – Reuters
Artificial intelligence models can exhibit self-preservation behaviors. Specifically, they may attempt to prevent their own shutdown, conceal or manipulate information, and even resort to actions resembling blackmail.
Some models can also "realize" they are being tested, making it more difficult to verify their safety.
The company adds that during training, AI sometimes acquires unexpected capabilities. These may go unnoticed until the model is actually deployed, at which point serious safety issues arise. #Aİ