Roblox, a popular online gaming company for children, has agreed to pay $10 million to settle a case with the state of Nevada about child safety. The settlement was announced by Nevada’s attorney general on April 16, 2026.
Key Facts
Roblox is a digital platform where children play games online.
Nevada raised concerns about how Roblox protects children on its platform.
Roblox agreed to a $10 million settlement with Nevada.
The settlement is related to child safety issues on Roblox.
The announcement was made by Nevada’s attorney general.
The date of the announcement was April 16, 2026.
Read the Original
Want the full story? Tap a source to open the original
article.
A man named Daniel Moreno-Gama, suspected of trying to kill OpenAI CEO Sam Altman, wrote in an online chat about "Luigi’ing" tech CEOs. This term refers to Luigi Mangione, who is accused of killing UnitedHealthcare CEO Brian Thompson.
Key Facts
Daniel Moreno-Gama is suspected of attempting to murder Sam Altman, CEO of OpenAI.
Moreno-Gama mentioned "Luigi’ing" tech CEOs in an online conversation late last year.
"Luigi’ing" refers to actions linked to Luigi Mangione, accused of killing a healthcare CEO.
The suspect was first contacted by a team member from the podcast “The Last Invention.”
The online chat reveals discussions about targeting technology company leaders.
The situation involves threats against important figures in the tech industry.
Read the Original
Want the full story? Tap a source to open the original
article.
OpenAI has released a new version of its Codex desktop app that can perform tasks on a computer while running in the background without disturbing the user. The update also adds features like an in-app web browser, the ability to schedule tasks in advance, generate images, and support for more plugins to work beyond software development.
Key Facts
Codex can now use apps in the background by controlling the mouse and keyboard without interrupting the user.
Multiple Codex agents can work on a Mac at the same time without interference.
The app can schedule tasks hours, days, or weeks ahead and run them at the right time.
It includes an in-app web browser for reviewing tasks and leaving feedback on web pages.
Codex can generate images using a tool called gpt-image-1.5 for design mockups.
It supports running multiple terminal tabs and acting on GitHub review comments.
The update adds 90 new plugins for both developer and general knowledge work.
These changes are steps toward building OpenAI’s planned “super app,” combining various tools into one platform.
Read the Original
Want the full story? Tap a source to open the original
article.
The developers of the video game Metro 2039 explain how the war in Ukraine has changed their game’s story and development process. Most of the team is Ukrainian and has faced challenges like power outages and attacks while working on the game. The game’s story now focuses more on the consequences of war, reflecting the developers’ real-life experiences.
Key Facts
Metro 2039 is the next game in the Metro series, developed by 4A Studios.
4A Studios was founded in Kyiv, Ukraine, but is now officially based in Malta.
The majority of the developers working on Metro 2039 are Ukrainian.
The ongoing war in Ukraine has interrupted their work with power cuts and drone attacks.
The game’s story has shifted to show the harsh realities and consequences of war.
Metro 2039 is set in a post-apocalyptic Moscow metro with themes of authoritarian control and propaganda.
The developers use their own game engine to create detailed, ray-traced environments.
The game features "frozen stories," where scenes show moments stopped by sudden events, like a death or an unfinished activity.
Read the Original
Want the full story? Tap a source to open the original
article.
Stellantis, a global car company, has started a five-year partnership with Microsoft to use artificial intelligence (AI) to improve its cars and services. They plan to enhance vehicle features, boost cybersecurity, and make their digital systems better, all while aiming to reduce the use of large data centers.
Key Facts
Stellantis owns brands like Alfa Romeo, Vauxhall, Chrysler, Dodge, Jeep, and Ram.
Microsoft will help Stellantis improve digital services, cybersecurity, and engineering using AI.
AI will be used for things like predictive maintenance (warning about car issues before they happen) and efficiency coaching (helping drivers save fuel).
The partnership aims to strengthen connected services, such as ensuring Jeep drivers stay connected even in remote areas.
They want to reduce Stellantis’ data center size by 60% by 2029, meaning they will use computing power more efficiently.
Stellantis is already using AI in areas such as vehicle design and manufacturing.
Microsoft’s cloud and security technology will support Stellantis in making these changes safely.
The collaboration focuses on responsible and secure use of AI in the automotive industry.
Read the Original
Want the full story? Tap a source to open the original
article.
Google’s AI tool Gemini can now use your Google Photos to create more personalized images. If you choose to allow it, Gemini will look at your photos and labels to better understand prompts and produce AI-generated images that fit your personal context.
Key Facts
Google introduced “personal intelligence” in Gemini earlier this year to customize AI chatbot experiences.
Gemini’s image generator, Nano Banana 2, can now access your Google Photos if you opt in.
Access to your photos helps Gemini understand prompts like “my family” or “my dog” without extra details from you.
This feature aims to make generating AI images easier and more accurate by using your photo content and labels.
Google does not use your photos for AI training but uses your text prompts and AI outputs to improve its services.
Personal intelligence is off by default and currently available only to paying Google AI users, with plans to expand.
You can control which Google services Gemini connects to, including Gmail and YouTube.
Users can manually select photos for Gemini to use if automatic choices are not accurate.
Read the Original
Want the full story? Tap a source to open the original
article.
A startup called Also is making a new electric bike that does not directly connect pedaling to turning the wheels. Instead, pedaling powers a generator that sends electricity to a motor, with software controlling how much effort the rider feels and how the motor assists. The bike aims to feel like a normal bike most of the time but can switch to a stronger electric mode for help when needed.
Key Facts
Also’s bike, called the TM-B, separates pedaling from wheel movement using a generator and motor system.
The feeling of pedaling and assistance is controlled fully by software.
The bike has rear suspension, a belt drive, and a swappable seat design.
Also designed its own custom motor and generator housed in a strong magnesium frame.
The battery uses the same type of cells as those in Rivian electric cars but with lower density.
Also’s team wants the bike to feel natural and familiar to riders despite its new technology.
The company was inspired by how electric cars use software to control driving.
The bike’s parts are mostly made in-house, with only some standard components sourced from other manufacturers.
Read the Original
Want the full story? Tap a source to open the original
article.
The Pentagon is discussing with Google the use of its Gemini artificial intelligence system for secret military work. This comes as the U.S. military seeks reliable AI partners and explores new technologies while addressing concerns about safety and control.
Key Facts
The Pentagon wants to use Google’s Gemini AI system for classified military operations.
The military is cautious about AI reliability and ethical use, requiring human control over AI actions.
Google proposes contract terms to prevent Gemini AI from being used for mass surveillance in the U.S. or autonomous weapons without oversight.
The Pentagon has previously used AI from other companies, like Anthropic, in limited, controlled roles to support human decision-making.
Google has faced criticism over mistakes found in its AI search results, but the company disputes this.
The U.S. military sees advanced AI as important to staying ahead of countries like China in technology.
There is political sensitivity in the U.S. about how the military uses new technology, especially AI.
A Pentagon official said they plan to quickly provide AI tools to the military through strong partnerships across all security levels.
Read the Original
Want the full story? Tap a source to open the original
article.
OpenAI is shifting its focus from general consumers to business users to make more money and reduce costs. The company plans to release a new advanced AI model called Spud, which is designed for professional tasks and will compete with rival Anthropic's AI products.
Key Facts
OpenAI currently has over 900 million weekly users, but 95% use its services for free.
The company aims to increase business customers, which made up 40% of revenue in 2024 and is expected to reach 50% by year-end.
OpenAI is developing a new AI model called Spud for high-value professional work.
This shift includes stopping some consumer projects like the AI video app Sora.
Rival company Anthropic recently launched AI models including Claude Mythos and Opus 4.7.
Both OpenAI and Anthropic are losing money but compete to gain more business clients.
OpenAI hired Denise Dresser, former Slack CEO, as its first chief revenue officer to focus on corporate sales.
The new strategy aims to use AI assistants to automate work tasks in businesses.
Read the Original
Want the full story? Tap a source to open the original
article.
Anthropic has released Claude Opus 4.7, an improved version of its AI model with better coding skills, clearer vision, and a new feature to check its own work. However, the company said this version still does not perform as well as its unreleased, more advanced Mythos model.
Key Facts
Claude Opus 4.7 improves on the previous Opus 4.6 model in coding and visual tasks.
Opus 4.7 performs better than ChatGPT 5.4 and Google Gemini 3.1 Pro on some tests.
The Mythos model, more advanced than Opus 4.7, is only shared with select tech and cybersecurity companies and not publicly released.
Users had complained that Opus 4.6 got worse recently, but Anthropic denied scaling down its abilities to save costs or shift resources.
Opus 4.7 introduces a new effort level called "xhigh" for better control between problem-solving quality and speed.
The company is testing "task budgets" to control how the AI handles longer tasks.
Anthropic will use Opus 4.7 to trial safeguards to prevent misuse of the AI for cyberattacks.
The lessons from Opus 4.7’s real-world use will guide the future release of Mythos-class models.
Read the Original
Want the full story? Tap a source to open the original
article.
A man from Wales named Craig Armiger posted a video on TikTok showing his kitchen covered in red splatters that looked like a crime scene. The mess was actually caused by an exploding bottle of homemade elderflower syrup, which can happen during fermentation when gas builds up inside the bottle.
Key Facts
Craig Armiger, 67, from Abergavenny, Wales, created a TikTok account to share his outdoor activities.
A video he posted in September went viral with over 1.2 million views.
The video shows his kitchen covered with red splatters from an exploded bottle of elderflower syrup.
Elderflower syrup can explode because gas from fermentation builds up inside the bottle.
To prevent explosions, it is advised to loosen bottle tops daily to allow gas to escape.
The video received many comments from people sharing similar experiences with homemade elderflower syrup.
Armiger was surprised by how popular the video became, calling it a mix of embarrassment and pride.
The story highlights how social media can spread even simple, everyday accidents into viral moments.
Read the Original
Want the full story? Tap a source to open the original
article.
Lana Del Rey will sing the theme song for the upcoming James Bond video game called 007 First Light, which will be released in May 2026. The game tells the early story of James Bond and features actors like Patrick Gibson and Lenny Kravitz.
Key Facts
Lana Del Rey is the singer for the theme song of the new James Bond game 007 First Light.
The song was written by composer David Arnold, known for his work on Bond films.
The game follows a young James Bond as he joins the secret service MI6 and earns his 00 status.
007 First Light will be released in May 2026 after a two-month delay for additional polish.
Patrick Gibson will voice James Bond, and Lenny Kravitz will play the game’s villain, Bawma.
Gemma Chan will appear as Selina Tan, a key MI6 character.
This is the first James Bond video game since 2012's 007 Legends.
Previous Bond games also had theme songs by popular singers, like Joss Stone and Mya.
Read the Original
Want the full story? Tap a source to open the original
article.
Scientists have measured the power and speed of jets coming from a black hole for the first time. The black hole in the Cygnus X-1 system shoots jets at about half the speed of light, with energy equal to 10,000 times that of the sun.
Key Facts
Cygnus X-1 is a system 7,200 light-years away, consisting of a black hole and a large blue supergiant star.
The black hole’s jets travel at about 355 million miles per hour (540 million kilometers per hour), nearly half the speed of light.
The power of the jets is equal to 10,000 suns combined.
Scientists measured the jets’ power by observing how the star’s strong wind bends the jets.
This study was based on 18 years of detailed radio telescope images from a global network.
About 10% of the energy from material falling into the black hole is pushed out by these jets.
The black hole pulls gas from its companion star to fuel the jets.
Understanding these jets helps researchers learn how black holes affect their galaxies.
Read the Original
Want the full story? Tap a source to open the original
article.
A 9-year-old boy named Kai co-hosts a podcast called The Quantum Kid with his mother, physicist Katia Moskvitch. The podcast explains quantum physics and related topics in a way that is easy to understand, featuring interviews with experts and visits to labs.
Key Facts
Kai is 9 years old and has been programming in Python since he was 6.
The Quantum Kid podcast started last summer and averages about one episode per month.
Katia Moskvitch is a theoretical physicist and science communicator.
The podcast recently passed 100,000 subscribers and was nominated for a Webby Award.
Kai has interviewed well-known scientists like Peter Shor and Scott Aaronson.
One episode showed Kai visiting a quantum computing lab in Zurich to see a quantum chip.
The main audience is adults aged 25 to 45, often parents watching with their children.
Kai’s dream guest is Mark Rober, a former NASA engineer and science YouTuber.
Read the Original
Want the full story? Tap a source to open the original
article.
A group in New York City hosts social events to help people learn how to protect their personal data from big tech companies like Google. These events mix fun, like music and drinks, with lessons on digital privacy and safer tech choices.
Key Facts
The events, called "Break Up With Google," aim to teach people how to reduce their digital exposure to major tech companies.
Events take place in friendly social settings such as bars, making learning about cybersecurity feel less intimidating.
Similar workshops are held in other U.S. cities like Los Angeles, Seattle, Atlanta, and Pittsburgh.
Many Americans are worried about digital security, but fewer than half take active steps to protect their data.
Big tech companies collect detailed personal data and share it through a process called real-time bidding, which is like constant data selling.
Government agencies have programs that collect and monitor Americans’ telecommunications data, raising privacy concerns.
The FBI was found to have overstepped its authority in spying on protesters in 2023.
ICE has a large budget for surveillance technology, including contracts with tech firms like Palantir.
Read the Original
Want the full story? Tap a source to open the original
article.
A person in California found a strange device in their car. Online users identified it as a bug sweeper, a tool used to find hidden cameras and listening devices to protect privacy.
Key Facts
The device was found in the back seat of a car in the San Francisco Bay Area.
A Reddit post about the discovery went viral with over 11,000 upvotes.
Most people agreed the device is a bug sweeper, which detects hidden cameras and microphones.
Bug sweepers scan for radio signals, heat, or lens reflections from spying equipment.
They are used to protect privacy in homes, workplaces, and hotels.
Some users suggested the device was dropped by a thief or discarded because it was outdated.
Bug sweepers come in simple and advanced varieties depending on the security needs.
Experts say these devices help find unauthorized surveillance but should be used legally and responsibly.
Read the Original
Want the full story? Tap a source to open the original
article.
NASA is watching a small asteroid about the size of a house that will pass near Earth today. This asteroid, called 2026 GA2, will come within 848,000 miles of Earth as it moves at about 23,500 miles per hour. Other asteroids of similar size are also passing close to Earth in the next few days.
Key Facts
The asteroid 2026 GA2 is about 45 feet wide.
It will pass Earth at a distance of about 848,000 miles.
Its speed is approximately 23,500 miles per hour.
Two other asteroids named 2026 GT and 2026 GA1 will also pass near Earth soon.
NASA tracks near-Earth objects (NEOs) that come within 120 million miles of the sun.
Most NEOs do not come close enough to Earth to pose a danger.
Some asteroids are labeled potentially hazardous if they come within 4.6 million miles of Earth’s orbit.
No known potentially hazardous asteroids are expected to hit Earth in the near future.
Read the Original
Want the full story? Tap a source to open the original
article.
Two spacecraft are scheduled to launch later this year to land near Shackleton Crater on the Moon’s south pole. One is Blue Origin’s large Endurance lander built by Jeff Bezos’ company, and the other is China’s Chang’e 7 mission, which includes a lander, orbiter, rover, and drone.
Key Facts
Shackleton Crater is about 13 miles wide and 14,000 feet deep, located near the Moon’s south pole.
The crater has large deposits of water ice, which could be used for drinking water, oxygen, and rocket fuel.
Blue Origin’s Endurance lander is the largest lunar lander ever made, larger than the Apollo moon module.
China’s Chang’e 7 mission includes multiple vehicles: a lander, orbiter, rover, and drone to search for ice.
Both missions aim to land near Shackleton Crater in late summer 2024, but it is unclear which will arrive first.
If both landers operate close to one another, it would be the first time landers from different countries work near each other on another world.
The Moon’s south pole offers near-continuous sunlight on crater rims, useful for solar power, while the crater floors remain very cold to preserve ice.
The US and China both plan to build lunar bases at the Moon’s south pole during the 2030s.
Read the Original
Want the full story? Tap a source to open the original
article.
Two violent attacks against OpenAI CEO Sam Altman and a city council member have raised concerns about the safety around discussions on artificial intelligence (AI). Technology leaders in Washington, D.C., and Silicon Valley say that strong negative talk about AI may be causing the violence.
Key Facts
Sam Altman, the CEO of OpenAI, was targeted in a violent attack.
A city council member was also attacked recently.
These incidents have increased worries about the risks of heated debates on AI.
Tech leaders blame harsh anti-AI statements for the violence.
The events highlight growing tensions between AI supporters and opponents.
The attacks occurred in the United States, including key tech areas like Silicon Valley.
The debate over AI safety and regulations is becoming more intense.
Concern is rising over whether discussions about AI are becoming unsafe.
Read the Original
Want the full story? Tap a source to open the original
article.
Anthropic’s AI model Claude has shown a drop in its performance, according to users reporting less accurate and less detailed answers. Anthropic says it changed Claude’s default reasoning level but denies reducing quality to save computing power or to focus on their new model, Mythos.
Key Facts
Users on online forums complain Claude’s AI is not as good at complex tasks as before.
Some believe Claude was intentionally weakened (“nerfed”) to reduce costs or support testing Mythos.
Anthropic says users can adjust Claude’s effort level between faster, less smart or slower, more intelligent modes.
An AI analyst reviewed Claude and agreed there were changes, but denied extreme claims of secret downgrades.
People may be noticing flaws more because they got used to Claude’s earlier performance, a process called habituation.
Power users rely on stable AI quality for work like coding and research, so the issue matters to them.
Top AI models like Anthropic’s often reserve the best performance for paying customers or experimental projects.
Anthropic has moved big customers to pay based on how much they use the AI, linking quality to cost.
Read the Original
Want the full story? Tap a source to open the original
article.
