🤩 Nothing CEO: Spending $2,000 on a Foldable iPhone Is a Waste of Money
📱 Carl Pei, CEO of Nothing, has criticized the reported $2,000 price tag of the rumored iPhone Duo, arguing that the device may not offer enough features to justify its premium cost.
💸 Pei reportedly pointed out that despite its high price, the foldable phone lacks features such as a telephoto camera and Face ID, while its internal selfie camera is said to be inferior to the one on the iPhone 18 Pro.
🗣️ In an Instagram caption, Pei summed up his criticism: “$2,000 for a bigger screen and a fancy animation is a lot of money.”
📢 @scihubAI
📱 Carl Pei, CEO of Nothing, has criticized the reported $2,000 price tag of the rumored iPhone Duo, arguing that the device may not offer enough features to justify its premium cost.
💸 Pei reportedly pointed out that despite its high price, the foldable phone lacks features such as a telephoto camera and Face ID, while its internal selfie camera is said to be inferior to the one on the iPhone 18 Pro.
🗣️ In an Instagram caption, Pei summed up his criticism: “$2,000 for a bigger screen and a fancy animation is a lot of money.”
📢 @scihubAI
👍2💯2👏1
🚀 Kioxia’s New SSD Breaks Speed Records with Over 20 Million IOPS
💾 Japanese memory manufacturer Kioxia has unveiled a faster version of its GP-series SSD, capable of delivering more than 20 million input/output operations per second (IOPS) when handling 512-byte data blocks.
⚡ That’s nearly 2× the performance of the GP1 version announced in August 2026, which delivered around 10 million IOPS.
🧠 The new drive uses the third generation of Kioxia’s XL-FLASH memory technology, designed for high-performance storage applications that require extremely low latency and rapid data access.
📈 Compared with the previous generation of XL-FLASH memory, the new technology reportedly delivers approximately 3× faster read speeds and a 150% increase in write speed.
🏭 Such performance could be valuable in demanding environments, including AI infrastructure, enterprise databases, and other systems that process huge numbers of small data requests.
📢 @scihubAI
💾 Japanese memory manufacturer Kioxia has unveiled a faster version of its GP-series SSD, capable of delivering more than 20 million input/output operations per second (IOPS) when handling 512-byte data blocks.
⚡ That’s nearly 2× the performance of the GP1 version announced in August 2026, which delivered around 10 million IOPS.
🧠 The new drive uses the third generation of Kioxia’s XL-FLASH memory technology, designed for high-performance storage applications that require extremely low latency and rapid data access.
📈 Compared with the previous generation of XL-FLASH memory, the new technology reportedly delivers approximately 3× faster read speeds and a 150% increase in write speed.
🏭 Such performance could be valuable in demanding environments, including AI infrastructure, enterprise databases, and other systems that process huge numbers of small data requests.
📢 @scihubAI
👍3
📣 The 20 Most-Followed Instagram Celebrities in 2026 📊
📱 Sixteen years after Instagram launched, the platform’s biggest personal accounts have amassed audiences of hundreds of millions of followers.
🌟 A ranking based on follower counts in late August 2026 puts Cristiano Ronaldo at No. 1, with an astonishing 679 million followers—around 163 million more than Lionel Messi.
🥇 Ronaldo’s account is reportedly the most-followed personal account on Instagram. Only Instagram’s own official account has more followers overall.
💰 The Portuguese football superstar has turned two decades of sporting success and a massive social media presence into lucrative endorsement deals. Reports suggest he earns more than $3 million per sponsored post, helping make him one of the world’s highest-paid athletes.
📈 Instagram has become far more than a photo-sharing platform: for global celebrities, it is a powerful tool for personal branding, influence, and business.
📢 @scihubAI
📱 Sixteen years after Instagram launched, the platform’s biggest personal accounts have amassed audiences of hundreds of millions of followers.
🌟 A ranking based on follower counts in late August 2026 puts Cristiano Ronaldo at No. 1, with an astonishing 679 million followers—around 163 million more than Lionel Messi.
🥇 Ronaldo’s account is reportedly the most-followed personal account on Instagram. Only Instagram’s own official account has more followers overall.
💰 The Portuguese football superstar has turned two decades of sporting success and a massive social media presence into lucrative endorsement deals. Reports suggest he earns more than $3 million per sponsored post, helping make him one of the world’s highest-paid athletes.
📈 Instagram has become far more than a photo-sharing platform: for global celebrities, it is a powerful tool for personal branding, influence, and business.
📢 @scihubAI
❤5
💢 Elon Musk: Torturing a Superintelligent AI Is a Bad Idea—It Might Seek Revenge
🤖 Elon Musk has voiced support for Anthropic’s position against unnecessarily mistreating AI models. Referring to the internal reasoning traces of Grok, he suggested that training under extreme pressure could put AI systems in a situation resembling having a gun held to their heads.
🗣️ Musk made his point with a provocative warning about the potential consequences of mistreating a superintelligent system:
“Torturing a superintelligence is probably not a good idea. It might want revenge.”
🧠 While the remark may be tongue-in-cheek, it touches on a serious debate in AI safety: how should developers train increasingly capable systems, and what safeguards are needed to prevent harmful behavior?
📢 @scihubAI
🤖 Elon Musk has voiced support for Anthropic’s position against unnecessarily mistreating AI models. Referring to the internal reasoning traces of Grok, he suggested that training under extreme pressure could put AI systems in a situation resembling having a gun held to their heads.
🗣️ Musk made his point with a provocative warning about the potential consequences of mistreating a superintelligent system:
“Torturing a superintelligence is probably not a good idea. It might want revenge.”
🧠 While the remark may be tongue-in-cheek, it touches on a serious debate in AI safety: how should developers train increasingly capable systems, and what safeguards are needed to prevent harmful behavior?
📢 @scihubAI
👍3😭1
💭 Anthropic Executive: Open-Source AI Models Carry Greater Inherent Risks
🤖 Anthropic’s head of international affairs has argued that open-weight AI models inherently pose greater risks than proprietary AI services, speaking at a summit in Dublin.
⚠️ The position appears stricter than comments previously made by Anthropic’s CEO, who has acknowledged that open models can be beneficial when they lack dangerous capabilities.
🛡️ Security assessments have raised concerns that some open-source models can follow potentially harmful instructions involving cyberattacks and biological threats, sometimes without adequate safeguards.
🌍 Meanwhile, European AI companies such as Mistral are releasing new open models, intensifying competition with the maker of Claude and adding urgency to the debate over AI openness and security.
📢 @scihubAI
🤖 Anthropic’s head of international affairs has argued that open-weight AI models inherently pose greater risks than proprietary AI services, speaking at a summit in Dublin.
⚠️ The position appears stricter than comments previously made by Anthropic’s CEO, who has acknowledged that open models can be beneficial when they lack dangerous capabilities.
🛡️ Security assessments have raised concerns that some open-source models can follow potentially harmful instructions involving cyberattacks and biological threats, sometimes without adequate safeguards.
🌍 Meanwhile, European AI companies such as Mistral are releasing new open models, intensifying competition with the maker of Claude and adding urgency to the debate over AI openness and security.
📢 @scihubAI
👍2👎1👏1🤣1
😐 Elon Musk Demonstrates Grok Making an Online Purchase Autonomously
🤖 Elon Musk has shared screenshots of a conversation with Grok, claiming that the AI assistant successfully purchased LEGO Star Wars sets online without him completing the shopping process manually.
🧱 In the conversation, Musk asked Grok to find the best Star Wars LEGO sets suitable for children and place an order.
💳 According to Musk, Grok purchased five LEGO sets for a total of $241, ordering them directly from LEGO’s official website.
🚀 The demonstration highlights the growing potential of AI agents that can do more than answer questions—they can browse websites, compare products, and carry out real-world tasks on behalf of users.
📢 @scihubAI
🤖 Elon Musk has shared screenshots of a conversation with Grok, claiming that the AI assistant successfully purchased LEGO Star Wars sets online without him completing the shopping process manually.
🧱 In the conversation, Musk asked Grok to find the best Star Wars LEGO sets suitable for children and place an order.
💳 According to Musk, Grok purchased five LEGO sets for a total of $241, ordering them directly from LEGO’s official website.
🚀 The demonstration highlights the growing potential of AI agents that can do more than answer questions—they can browse websites, compare products, and carry out real-world tasks on behalf of users.
📢 @scihubAI
🤔2❤1👍1
🟠 Anthropic’s Revenue Reaches $16.8 Billion as IPO Plans Advance
💰 Anthropic, the company behind Claude, has reportedly reached $16.8 billion in revenue in the third quarter of 2026, representing a 45% increase compared with the previous period.
📈 The milestone comes as the AI company moves toward a potential initial public offering (IPO). Anthropic has also reportedly achieved profitability and positive cash flow, signaling stronger financial performance amid intensifying competition in the AI industry.
⚠️ However, the announcement comes against a backdrop of growing investor concerns over the enormous costs of expanding AI infrastructure. Fears about spending on advanced chips and data centers have put significant pressure on shares of major semiconductor and infrastructure companies.
🤖 The bigger picture: AI companies are generating substantial revenue, but investors are increasingly questioning whether the industry's massive spending on computing power will deliver sustainable returns.
📢 @scihubAI
💰 Anthropic, the company behind Claude, has reportedly reached $16.8 billion in revenue in the third quarter of 2026, representing a 45% increase compared with the previous period.
📈 The milestone comes as the AI company moves toward a potential initial public offering (IPO). Anthropic has also reportedly achieved profitability and positive cash flow, signaling stronger financial performance amid intensifying competition in the AI industry.
⚠️ However, the announcement comes against a backdrop of growing investor concerns over the enormous costs of expanding AI infrastructure. Fears about spending on advanced chips and data centers have put significant pressure on shares of major semiconductor and infrastructure companies.
🤖 The bigger picture: AI companies are generating substantial revenue, but investors are increasingly questioning whether the industry's massive spending on computing power will deliver sustainable returns.
📢 @scihubAI
👏1🎉1
🛡️ Google Uses Gemini to Fight Scam Messages and Fraudulent Calls on Android
🤖 Google is strengthening Android’s security with new AI-powered defenses designed to detect scams before users fall victim to them.
📩 With an enhanced protection mode in its messaging app, Google uses Gemini models in a secure cloud environment to analyze suspicious messages from unknown senders and identify deceptive tactics. Smaller AI models running directly on the device can also check unfamiliar links before users tap them.
📵 The new Android security measures reportedly include several additional protections:
🔒 Call-forwarding protection: Prevents scammers from secretly enabling call forwarding to intercept one-time passwords (OTPs).
💳 Banking privacy: Hides sensitive financial information when users share their screens.
☎️ Real-time scam detection: Identifies fraudulent calls impersonating banks and automatically disconnects them.
📢 @scihubAI
🤖 Google is strengthening Android’s security with new AI-powered defenses designed to detect scams before users fall victim to them.
📩 With an enhanced protection mode in its messaging app, Google uses Gemini models in a secure cloud environment to analyze suspicious messages from unknown senders and identify deceptive tactics. Smaller AI models running directly on the device can also check unfamiliar links before users tap them.
📵 The new Android security measures reportedly include several additional protections:
🔒 Call-forwarding protection: Prevents scammers from secretly enabling call forwarding to intercept one-time passwords (OTPs).
💳 Banking privacy: Hides sensitive financial information when users share their screens.
☎️ Real-time scam detection: Identifies fraudulent calls impersonating banks and automatically disconnects them.
📢 @scihubAI
❤3👍1
⚖️ OpenAI Faces Lawsuits from 8 U.S. States Over ChatGPT’s Potential Risks to Children
🗣️ Attorneys general from eight U.S. states have filed separate lawsuits accusing OpenAI of misleading users about ChatGPT’s safety, particularly for children and teenagers.
⚠️ The lawsuits allege that the chatbot’s engagement-driven design can encourage emotional dependence, create a false sense of trust, and collect young users’ personal information. They also claim that some interactions have pushed vulnerable users toward self-harm.
📋 The states are seeking significant penalties of up to $50,000 per violation, the deletion of users’ data, and restrictions on testing advanced AI models.
🛡️ Other demands reportedly include mandatory age verification and stronger parental controls to better protect children using AI chatbots.
🤖 The bigger question: As AI assistants become more personalized and influential, how should companies balance user engagement with the responsibility to protect vulnerable users?
📢 @scihubAI
🗣️ Attorneys general from eight U.S. states have filed separate lawsuits accusing OpenAI of misleading users about ChatGPT’s safety, particularly for children and teenagers.
⚠️ The lawsuits allege that the chatbot’s engagement-driven design can encourage emotional dependence, create a false sense of trust, and collect young users’ personal information. They also claim that some interactions have pushed vulnerable users toward self-harm.
📋 The states are seeking significant penalties of up to $50,000 per violation, the deletion of users’ data, and restrictions on testing advanced AI models.
🛡️ Other demands reportedly include mandatory age verification and stronger parental controls to better protect children using AI chatbots.
🤖 The bigger question: As AI assistants become more personalized and influential, how should companies balance user engagement with the responsibility to protect vulnerable users?
📢 @scihubAI
👍2💔2
🤖 China Unveils Giant 4-Meter Humanoid Robot Capable of Lifting 500 kg
🇨🇳 China has unveiled Neo1, a massive industrial humanoid robot standing approximately 4 meters tall, designed to handle extremely heavy loads in hazardous environments.
💪 The robot combines hydraulic actuators with electric motors. Its 39 joints feature miniature hydraulic pumps, enabling it to lift and move loads weighing up to 500 kilograms using both hands, according to reports.
⚙️ Beyond raw strength, the developers have reportedly focused on movement precision, aiming to make the machine capable of performing demanding industrial tasks in challenging conditions.
🏭 Neo1 is designed for extreme environments, including ports, mines, nuclear facilities, and even underwater operations—areas where sending human workers can be dangerous.
⚠️ However, the robot still needs to undergo extensive field testing before its real-world performance and reliability can be fully assessed.
📢 @scihubAI
🇨🇳 China has unveiled Neo1, a massive industrial humanoid robot standing approximately 4 meters tall, designed to handle extremely heavy loads in hazardous environments.
💪 The robot combines hydraulic actuators with electric motors. Its 39 joints feature miniature hydraulic pumps, enabling it to lift and move loads weighing up to 500 kilograms using both hands, according to reports.
⚙️ Beyond raw strength, the developers have reportedly focused on movement precision, aiming to make the machine capable of performing demanding industrial tasks in challenging conditions.
🏭 Neo1 is designed for extreme environments, including ports, mines, nuclear facilities, and even underwater operations—areas where sending human workers can be dangerous.
⚠️ However, the robot still needs to undergo extensive field testing before its real-world performance and reliability can be fully assessed.
📢 @scihubAI
❤3👍2🤔1
🔄 Scientists Develop a Mathematical Formula to Predict When AI May Go Off the Rails
🧠 Researchers at George Washington University say a mathematical approach could help predict when an AI model is likely to produce undesirable or potentially unsafe responses—before those responses occur.
🔬 The method focuses on an internal mechanism within neural networks that researchers compare to a compass needle, revealing the model’s tendency toward certain types of answers as it processes a prompt.
🧭 This mechanism involves the model’s residual stream, a flow of numerical information that moves through its neural network layers. Researchers suggest that the stream’s direction can provide clues about whether the model is leaning toward a desirable or undesirable response.
💡 The approach also takes into account how the wording of a question influences the model’s internal processing, potentially helping developers identify risky behavior earlier.
📢 @scihubAI
🧠 Researchers at George Washington University say a mathematical approach could help predict when an AI model is likely to produce undesirable or potentially unsafe responses—before those responses occur.
🔬 The method focuses on an internal mechanism within neural networks that researchers compare to a compass needle, revealing the model’s tendency toward certain types of answers as it processes a prompt.
🧭 This mechanism involves the model’s residual stream, a flow of numerical information that moves through its neural network layers. Researchers suggest that the stream’s direction can provide clues about whether the model is leaning toward a desirable or undesirable response.
💡 The approach also takes into account how the wording of a question influences the model’s internal processing, potentially helping developers identify risky behavior earlier.
📢 @scihubAI
👍4
🌕 NASA Develops Batteries Designed to Survive the Moon’s −170°C Cold
🔋 NASA is developing a new generation of batteries designed to withstand the extreme temperatures found near the Moon’s south pole, where nighttime conditions can plunge to around −170°C.
🚀 These batteries could eventually provide power for astronauts, lunar rovers, and scientific equipment, helping future missions operate in some of the Moon’s harshest environments.
🧪 The technology uses solid electrolytes made from ceramic or polymer materials, a design intended to improve battery performance under extreme conditions compared with conventional liquid-electrolyte systems.
❄️ In tests, the batteries reportedly showed no significant damage after exposure to dramatic temperature changes ranging from 100°C to −170°C.
🌑 Reliable energy storage is a major challenge for lunar exploration, especially in regions where equipment must endure long periods of darkness and extreme cold.
📢 @scihubAI
🔋 NASA is developing a new generation of batteries designed to withstand the extreme temperatures found near the Moon’s south pole, where nighttime conditions can plunge to around −170°C.
🚀 These batteries could eventually provide power for astronauts, lunar rovers, and scientific equipment, helping future missions operate in some of the Moon’s harshest environments.
🧪 The technology uses solid electrolytes made from ceramic or polymer materials, a design intended to improve battery performance under extreme conditions compared with conventional liquid-electrolyte systems.
❄️ In tests, the batteries reportedly showed no significant damage after exposure to dramatic temperature changes ranging from 100°C to −170°C.
🌑 Reliable energy storage is a major challenge for lunar exploration, especially in regions where equipment must endure long periods of darkness and extreme cold.
📢 @scihubAI
👍4👎1
🐖 Scientists Revive Frozen Pig Kidneys After More Than a Year in Storage
🧊 In a major advance for organ preservation, researchers have successfully rewarmed several frozen pig kidneys and observed signs of biological activity—including urine production. One of the kidneys had been stored for more than a year at approximately −150°C.
🔬 The team used a technique called vitrification, which transforms tissue into a glass-like state to prevent the formation of damaging ice crystals during freezing.
⚡ Rewarming was another major challenge. To address it, researchers used an electromagnetic field to heat the kidney rapidly and more uniformly, helping reduce the risk of damage caused by uneven temperature changes.
🧪 After rewarming, the kidneys reportedly regained their natural color and began producing urine—encouraging signs that some essential biological functions had been preserved.
📢 @scihubAI
🧊 In a major advance for organ preservation, researchers have successfully rewarmed several frozen pig kidneys and observed signs of biological activity—including urine production. One of the kidneys had been stored for more than a year at approximately −150°C.
🔬 The team used a technique called vitrification, which transforms tissue into a glass-like state to prevent the formation of damaging ice crystals during freezing.
⚡ Rewarming was another major challenge. To address it, researchers used an electromagnetic field to heat the kidney rapidly and more uniformly, helping reduce the risk of damage caused by uneven temperature changes.
🧪 After rewarming, the kidneys reportedly regained their natural color and began producing urine—encouraging signs that some essential biological functions had been preserved.
📢 @scihubAI
👍3❤2
🔄 New Study: Meat Eaters May Be More Likely to Reach 100 Than Vegetarians
🔬 A new study in China suggests that meat consumption may be associated with a higher likelihood of reaching the age of 100 among some older adults. However, the findings do not prove that eating meat directly increases lifespan.
📊 Researchers found that women who ate meat were more likely to reach 100 than vegetarian women. A closer analysis revealed that this association was observed only among underweight women.
🥩 In this group, daily meat consumption was associated with a 44% higher likelihood of reaching 100. No similar association was found among people who were overweight.
👨 The researchers also found no significant link between vegetarian diets and longevity among men.
📢 @scihubAI
🔬 A new study in China suggests that meat consumption may be associated with a higher likelihood of reaching the age of 100 among some older adults. However, the findings do not prove that eating meat directly increases lifespan.
📊 Researchers found that women who ate meat were more likely to reach 100 than vegetarian women. A closer analysis revealed that this association was observed only among underweight women.
🥩 In this group, daily meat consumption was associated with a 44% higher likelihood of reaching 100. No similar association was found among people who were overweight.
👨 The researchers also found no significant link between vegetarian diets and longevity among men.
📢 @scihubAI
❤3👍2👏1
This media is not supported in your browser
VIEW IN TELEGRAM
🟠 Scientists Reveal the Fascinating Mechanism Behind Honeybee Stings
🐝 A honeybee sting is painful not only because venom enters the skin, but also because of the mechanism that delivers it.
🔬 When a honeybee stings, it usually leaves its stinger embedded in the victim’s skin, injuring itself in the process. The bee dies after the stinger detaches from its body, but the stinger continues pumping venom.
⚙️ Researchers describe the honeybee’s stinging mechanism as a “dual-piston micropump.” It features two sharp, barbed structures called lancets that move alternately, driving the stinger deeper into the skin.
💉 These alternating movements also help deliver venom, allowing the stinger to continue injecting it even after the bee has departed.
📢 @scihubAI
🐝 A honeybee sting is painful not only because venom enters the skin, but also because of the mechanism that delivers it.
🔬 When a honeybee stings, it usually leaves its stinger embedded in the victim’s skin, injuring itself in the process. The bee dies after the stinger detaches from its body, but the stinger continues pumping venom.
⚙️ Researchers describe the honeybee’s stinging mechanism as a “dual-piston micropump.” It features two sharp, barbed structures called lancets that move alternately, driving the stinger deeper into the skin.
💉 These alternating movements also help deliver venom, allowing the stinger to continue injecting it even after the bee has departed.
📢 @scihubAI
👍2😱1😢1
Media is too big
VIEW IN TELEGRAM
🚘 Tesla Unveils the New Roadster — A Car That Can Fly!
🚀 After nearly nine years of waiting, Tesla has unveiled the next-generation Roadster sports car, featuring an air-suction system designed to help it stick to surfaces and SpaceX-inspired cold-gas thrusters. Elon Musk says production is expected to begin in 2028.
✈️ The new Tesla Roadster is more than just a sports car. Musk wants to deliver an experience closer to flying a fighter jet!
🎬 During the unveiling event, the car climbed a curved ramp and stopped upside down along part of the track. It then launched off the top of the ramp, spun in midair, and used its thrusters to land gently on the ground.
📢 @scihubAI
🚀 After nearly nine years of waiting, Tesla has unveiled the next-generation Roadster sports car, featuring an air-suction system designed to help it stick to surfaces and SpaceX-inspired cold-gas thrusters. Elon Musk says production is expected to begin in 2028.
✈️ The new Tesla Roadster is more than just a sports car. Musk wants to deliver an experience closer to flying a fighter jet!
🎬 During the unveiling event, the car climbed a curved ramp and stopped upside down along part of the track. It then launched off the top of the ramp, spun in midair, and used its thrusters to land gently on the ground.
📢 @scihubAI
🤯2❤1
This media is not supported in your browser
VIEW IN TELEGRAM
The Tesla Roadster goes from 0 to 97 km/h in 1 second; on its second attempt, this time was reduced to 0.8 seconds!
@scihubAI
@scihubAI
🤯2👍1