A team of researchers at Georgia Tech has developed a new smartphone‑based system that could dramatically simplify how people interact with robots. Called COBALT, the platform allows users with little to no computing experience to remotely control robot arms from virtually anywhere in the world using just a phone and an internet connection.
The project, developed at Georgia Tech’s People, AI & Robotics (PAIR) Lab, transforms smartphones into motion controllers for robotic arms. Users simply move their phones in different directions, and the robot mirrors those movements in real time. Basic tasks such as grabbing, moving, and releasing objects can be performed through simple on‑screen controls, making the experience feel more like playing a mobile game than operating industrial machinery.
Ayush Agarwal, a Ph.D. student in Georgia Tech’s School of Interactive Computing who leads the COBALT research team, said the system was intentionally designed to make robotics accessible to beginners rather than experts. During testing, participants from countries including India, Indonesia, and Pakistan remotely controlled robot arms located inside Georgia Tech’s lab despite having no prior robotics experience.
Researchers believe crowdsourcing could shape the future of robotics
The broader goal behind COBALT extends beyond convenience. Researchers believe the platform could solve one of robotics’ biggest challenges: collecting enough real‑world training data to improve AI‑powered robotic systems.
Modern robots require enormous amounts of policy training data to learn how to perform physical tasks reliably. According to Assistant Professor Animesh Garg, who directs the PAIR Lab, simulation alone is not enough to train robots for large‑scale deployment. Instead, researchers envision a crowdsourced network where millions of smartphone users passively contribute operational data by remotely interacting with robots.
Garg compared the idea to tapping into the nearly five billion smartphone users worldwide. By lowering the barrier to entry, the team hopes to create a scalable global system capable of accelerating robotic learning and automation.
The technology could also have major educational implications. Georgia Tech researchers recently demonstrated COBALT to students from Midtown High School in Atlanta, allowing them to remotely operate robot arms using smartphones. The simplicity of the interface could make robotics education more accessible in classrooms without expensive equipment or specialized hardware.
A future “gig economy” for robots may not be far away
The researchers also believe COBALT could eventually support entirely new forms of remote work. Garg described the possibility of a robot‑powered gig economy where people remotely operate assistive robots in homes, warehouses, or factories from anywhere in the world.
In practical terms, that could mean a factory robot autonomously handles most tasks but requests human assistance when it encounters a difficult situation. Instead of requiring on‑site workers, remote operators could briefly take control through their phones before handing the operation back to the AI system.
Agarwal said user studies showed smartphones were preferred over VR headsets, keyboards, or traditional controllers because they felt more intuitive while still providing high‑quality control data. The system also minimizes latency by using WebRTC technology, similar to platforms like Zoom and Google Meet, ensuring that robot movements and live video streams remain responsive even across long distances.
The research paper on COBALT is being presented this week at the IEEE International Conference on Robotics and Automation in Vienna, where the team is showcasing not just the technology itself, but the large‑scale remote operation network built around it.
Google may have just accidentally shown everyone where Gemini is headed next. According to TestingCatalog, a new Troubleshooting mode has quietly appeared inside the Gemini model picker menu for some users.
It sits alongside existing options like Gemini 3.5 Flash and 3.1 Pro, which are the standard AI models you already switch between in the app.
GOOGLE 🔥: A new Troubleshooting mode has been spotted on Gemini.
In this mode, Gemini will explain troubleshooting process via text responses and interactive widgets. Even though it is working and available, it still looks like an unintended release and might get reverted… https://t.co/FWQLelYXjupic.twitter.com/Y73PJb7y1e
— 🚨 AI News | TestingCatalog (@testingcatalog) June 4, 2026
What does the Troubleshooting mode in Gemini actually do?
Rather than giving you a wall of text to read, the Troubleshooting mode guides you through a problem step by step using a mix of text responses and interactive widgets.
For example, if you tell Gemini your car will not start, it might identify common causes like a dead battery and then present you with symptom options to tap, such as “clicks or silent,” to help narrow down the issue faster. It is a more structured, guided experience than asking Gemini a question in regular chat mode.
How is this different from just asking Gemini normally?
That is a fair question, and the answer comes down to how the mode is tuned under the hood. Redditors who got early access suggest it runs on a lower temperature setting, which means it sticks closely to the problem at hand and skips the conversational filler.
Its responses are reportedly focused on diagnosis and practical fixes rather than general information. Google has not officially announced the feature, and it remains unclear whether this is a planned rollout or an internal test.
For now, the Troubleshoot feature appears to be an unintended release, meaning Google likely flipped it on by mistake, and could pull it back at any time. More details are expected in the coming weeks.
If you’ve ever written an email draft in ChatGPT only to copy it into Gmail or Outlook to hit send, you can now skip that extra step. OpenAI has introduced a feature that enables you to dispatch emails straight from writing blocks on the web version of ChatGPT, keeping the entire workflow inside a single conversation.
ChatGPT writing blocks now support sending emails
Writing blocks, a formatting tool launched late last year, turn email drafts into a distinct, clean block that resembles a real email editor rather than a plain chat reply. You can highlight any line to request revisions, accept or reject suggestions individually, and edit the text directly within the block without needing to copy it elsewhere. The newest update pushes this further by adding a send button, so you never have to leave the chat to deliver your message.
Draft it. Tweak it. Send it.You can now send emails directly from writing blocks in ChatGPT on the web, without leaving the conversation. pic.twitter.com/GoQtlSFGFG
— ChatGPT (@ChatGPTapp) June 5, 2026
Is it safe to send emails through ChatGPT?
Before you start relying on this capability, note a recent lawsuit filed in California alleging that OpenAI shared user prompts, chat queries, and identifying data with Google and Meta tracking tools without proper consent, potentially violating state privacy law and federal wiretap statutes. You may want to avoid drafting highly sensitive content until the matter is resolved.
What else is ChatGPT working on?
OpenAI continues to roll out new features. The assistant is getting better at remembering details about you by learning from your chat history, making conversations feel more personalized over time. On the productivity front, ChatGPT for Excel and Google Sheets has left beta and is now globally available, allowing users to create workbooks, clean data, and generate formulas using plain language without being spreadsheet experts.
Apple has spent most of the past year telling users that a more capable Siri is on the horizon. A fresh report now indicates the company might be tempering expectations before the assistant finally lands. Bloomberg’s Mark Gurman says internal documents label the overhauled Siri as a “beta” and “preview” product, suggesting Apple may not present it as a polished, finished experience when it rolls out later this year. This mirrors the rollout of the original Siri, which bore the beta tag for about two years after its debut.
**Apple appears to be lowering the stakes**
Choosing this route would break from Apple’s usual practice of unveiling major software features with a high‑gloss launch. While Apple is famed for refined releases, artificial intelligence poses a distinct set of challenges. Branding the new Siri as a preview would give the company leeway to refine the assistant publicly without promising perfection from day one. It also helps explain why Apple has been unusually cautious when discussing the next‑generation capabilities of Siri after earlier delays pushed the project back.
The approach reflects a wider reality confronting the AI sector. Whether it’s chatbots spewing inaccurate data or digital assistants missing context, even the biggest tech firms are still learning how to make AI dependable for everyday use.
**Not everyone may get access immediately**
Gurman’s reporting also hints at another scenario: Apple could roll out a waitlist for users eager to test the upgraded Siri. This wouldn’t be unprecedented—Apple employed a similar tactic when it introduced Apple Intelligence in 2024, gradually expanding access rather than opening it to everyone at once. A waitlist would let the company monitor performance, collect feedback, and manage demand while ironing out bugs behind the scenes.
For users, this means the debut of Apple’s AI‑powered assistant may resemble an early‑access program more than a traditional software launch. While that could disappoint those hoping for an instant upgrade, it may ultimately be the safer path. A smarter Siri that arrives gradually is likely preferable to one that launches quickly and falls short of Apple’s lofty AI ambitions.
For years, smartphone makers have been locked in a race for brighter screens, thinner bezels, and sharper resolutions. Now, it looks like the next battleground could be refresh rates — and things are getting a little absurd.
A new leak suggests OnePlus is exploring a roadmap that could eventually bring 240Hz OLED displays to its flagship phones. That’s a number typically associated with competitive gaming monitors, not devices that spend most of their time scrolling through social media feeds and watching YouTube videos. According to tipster Digital Chat Station, OnePlus is considering a gradual jump through 165Hz and 185Hz panels before ultimately reaching 240Hz in future devices.
The refresh rate race is heating up
Most flagship smartphones today top out at 120Hz, which already feels incredibly smooth for everyday use. Animations are fluid, scrolling feels responsive, and games that support high frame rates look noticeably better. But OnePlus appears interested in pushing beyond what most users would consider necessary.
Recent rumors surrounding the upcoming OnePlus 16 have already hinted at a 165Hz-to-185Hz jump while retaining the company’s preferred 1.5K display resolution. That suggests OnePlus may be prioritizing speed over pixel count, at least for now. It’s not hard to see the appeal. Higher refresh rates can make supported games feel more responsive, especially in fast-paced shooters and racing titles. The challenge is that the benefits become increasingly difficult to notice as the numbers climb.
The real challenge isn’t speed
Getting to 240Hz is one thing; doing it without destroying battery life is another. That’s likely why OnePlus reportedly continues to favor 1.5K panels over sharper 2K displays. Combining ultra-high refresh rates with higher resolutions demands more power, more processing muscle, and more aggressive thermal management.
The company could revisit 2K screens in the future, but only if display technology improves enough to avoid major compromises. For now, the rumored OnePlus 16 is expected to arrive later this year with Qualcomm’s Snapdragon 8 Elite Gen 6 chip and a larger silicon-carbon battery, both of which could help support more demanding display hardware. Whether anyone truly needs a 240Hz smartphone screen is another question entirely. But if the leak is accurate, OnePlus seems determined to find out.
Apple is reportedly gearing up for a potentially disruptive shift in how notifications behave in iOS 27 and iPadOS 27. Bloomberg’s Mark Gurman says internal builds now show incoming alerts sliding in from the left edge of the screen. While that might appear to be a minor visual tweak, it seems to be part of a broader redesign of navigation gestures that could compel long‑time iPhone users to re‑train years of instinctive motions.
The familiar swipe might no longer work as expected
For years, iPhone owners have relied on a simple gesture: swipe down from near the middle of the display to pull down the Notification Center. In iOS 27, that action is slated to open Search or an AI‑driven assistant panel instead. Accessing notifications would require a new motion—swiping down from the left side of the screen. Anyone who has switched to a new smartphone after years on another platform knows how deeply these gestures become ingrained.
Apple’s AI push could be behind the change
The reported redesign hints that Apple wants Search and its AI capabilities to take a much more prominent place in the iPhone experience. Rather than tucking AI tools behind buttons or menus, the company appears to be assigning them one of the most natural gestures on the device, signaling where it sees future user interactions heading.
The animation for notifications also seems crafted to reinforce the new behavior. With alerts now arriving from the{ left side of the screen, the visual cue lines up with the new swipe direction needed to view them. Whether users will welcome the alteration remains to be seen. History shows that even modest tweaks to familiar gestures can provoke strong reactions. If the leak is accurate, iOS 27 may not only look different—it could reshape how millions of people instinctively interact with their iPhones each day.
A New York Times examination of internal documents from lawsuits filed by more than 1,400 school districts against Meta, Snap, TikTok, and YouTube shows that these firms deliberately targeted students, even as their own safety teams warned about the damage being caused.
The evidence is stark. Snapchat sent phone alerts to teens during school hours, urging them to post what was happening in their classrooms. A Snapchat strategy memo even labeled classroom phone use as “under the desk” time.
Meta went further, hiring “teen ambassadors” and paying high‑schoolers $45 gift cards plus branded gear to promote Instagram to their peers. TikTok contributed millions to the National PTA, partially to fund school events focused on online safety.
Did the companies know what they were doing? The answer is yes, and that’s why the revelations are so unsettling. TikTok’s safety team had advocated for years to turn off notifications during school hours, but senior leadership rejected the proposal. A TikTok employee wrote in 2022, “Teachers are going to hate it. Kids already have smartphone addiction in class,” referring to a feature that nudged users to post within three minutes. A manager replied, “If we assume teens are going to do this anyway, we’d rather them be here on TikTok.”
Google was not blameless either. A 2020 internal memo stated that “investing in schools helps onboard kids into Google’s ecosystem,” and YouTube managers were aware that the algorithm was serving off‑topic videos to students during class time.
What’s next? All four companies recently settled with Breathitt County Schools, a small Kentucky district of roughly 1,500 students, for $27 million. However, that is likely only the start. The upcoming case involves Tucson Unified School District, which is seeking more than $1 billion in damages.
Cornell Law professor Alexandra Lahav described the litigation as “massive, massive lawsuits” that could ultimately cost these corporations billions. The companies argue that the pandemic and other factors are to blame for the teen mental‑health crisis, and that parents and schools also share responsibility. Whether a court will concur is a separate question.
Even if a court awards a billion dollars, that sum is a drop in the bucket for these firms, which can easily absorb it while generating 100 times that amount in a single year. Unless criminal charges are pursued for plainly harming children and students, and strict legislation is enacted, such practices are unlikely to cease anytime soon.
Small AI models just received an unexpected boost from a classic board game. MIT researchers set up a Battleship‑style environment to see if AI agents could become better at gathering information before taking a turn. The outcome was a dramatic rise in performance for compact systems, including one model that went from rarely beating humans to winning the majority of games after the researchers altered its board‑search strategy.
This improvement targets a major flaw in today’s AI agents: they are often tasked with problems whose answers depend on details they haven’t yet obtained. MIT’s findings suggest that smarter question planning can make a low‑cost model act far more competently.
How much smarter did it get?
MIT’s experiment used a Battleship variant driven by natural‑language queries. One AI acted as the teammate tasked with locating hidden ships, while another had full board visibility and provided answers.
The most striking gain came from Llama 4 Scout. Initially, the smaller model defeated human opponents in only 8 % of games. After the researchers introduced a more deliberate inference method, its win rate jumped to 82 %, outpacing a larger frontier model while costing roughly 1 % of the expense.
That metric matters for anyone watching AI costs. The model didn’t win by becoming larger; it won by asking sharper questions and extracting more value from each response.
Why does Battleship help AI learn?
Battleship serves as an ideal test because it forces an AI to operate with incomplete information. It can’t see the entire board, so every query must narrow the search space and set up the next move.
This mirrors real‑world AI tools. A support bot, research assistant, or planning agent often needs to ask follow‑up questions before it can help. When that step fails, the model may miss crucial details, repeat itself, or issue premature recommendations.
The MIT approach puts pressure on that weak point by measuring whether an agent can collect the right data before delivering an answer.
Where could this go next?
The tougher question is whether the same technique works outside of games. Battleship is a controlled environment, making scoring easier than evaluating open‑ended agent workflows in search, customer service, or workplace software.
Nevertheless, the trend is worth watching. If smaller models learn to pose better questions before acting, companies could deploy cheaper AI tools that feel more capable in everyday tasks.
The next milestone will be transferring the skill from a game board to real‑world work. Tasks with vague instructions, missing files, and hurried users will pose a far greater challenge.
Valve has announced that Steam Machine will ship this summer, finally giving PC gamers a concrete launch window for its SteamOS living‑room PC. The missing piece is still the price, and that’s the detail many buyers need before they can decide whether it fits their setup.
The update arrived as Valve broadened its Verified program to include Steam Machine and Steam Frame. For Steam Machine, games will be evaluated for default controller support, default graphics settings, and how well they run without manual tweaks. Valve says the hardware is roughly six times as powerful as the Steam Deck, while still running SteamOS, the Steam interface, and Proton.
**How your library will look**
Steam Machine Verified should feel familiar if you’ve used the Steam Deck. The requirements are almost identical, so you’ll get a clearer indication of whether a game is ready for TV play before you spend time adjusting controls or graphics settings.
Valve already has a solid foundation for that work. Tens of thousands of titles have passed Steam Deck verification, and Valve is testing Steam Machine support for games that missed Deck performance targets because of CPU or GPU limits. On stronger hardware, some of those games could meet the new bar without developers changing anything.
**Why the price gap lingers**
The summer timing makes Steam Machine more concrete, but the missing price keeps the comparison unfinished. Buyers still don’t know whether Valve’s living‑room PC will be priced closer to a Steam Deck, a gaming laptop, or a compact Windows gaming PC.
That comparison goes beyond raw performance. Valve must demonstrate that a TV‑connected SteamOS PC can make PC gaming easier in the living room than the options people can already buy. Verified labels should reduce setup uncertainty, but price will decide whether that convenience looks worth paying for.
**When buyers get the rest**
Valve has also added Steam Machine and Steam Frame tabs to the Partner Dashboard, where some games already have Verified results for the new devices. That gives developers more guidance before launch, but it isn’t the full consumer reveal yet.
For now, you shouldn’t allocate budget for Steam Machine until Valve shares the remaining hardware details. Price is the big unknown, but final availability timing and configuration options will also shape whether it’s a smart upgrade or a wait‑and‑see PC gaming box.
Fake reviews pose a serious problem for shoppers on the web. If you’ve ever purchased an item based on glowing feedback only to receive a sub‑par product, you’ve experienced the issue firsthand. A recent study in the International Journal of Information and Communication Technology introduces an AI‑driven system that not only spots counterfeit reviews but also maps their propagation.
Why current solutions fall short
Most existing detection tools concentrate solely on the textual content of a review. This worked for a time, but fraudsters have become more sophisticated, pairing well‑crafted prose with deceptive images to make their posts appear genuine. Text‑only approaches struggle to catch this blend, creating challenges for both consumers and honest merchants.
Multi‑signal approach
The researchers tackled the issue by creating a system that evaluates several cues simultaneously. It processes the review text using two techniques—a convolutional neural network for text and pre‑trained language models—to capture both surface patterns and deeper semantics. It also examines reviewer behavior, noting that fake accounts often use default avatars and auto‑generated usernames, whereas real users tend to personalize their profiles.
Can AI also detect bogus images?
The short answer is yes. Review images are examined separately with a residual network, a deep‑learning architecture commonly applied to visual data. After gathering all the signals, the system fuses them to decide whether a review is legitimate.
When a review is flagged as fake, a Transformer model is activated to trace its origin and follow how far it has spread across the network.
Results
Experiments on a large JD.com dataset showed the model achieved a detection accuracy of 94.2% and a tracing accuracy of 93.5%, surpassing all compared baseline methods. Such performance could eventually lead to fewer deceptive reviews and more trustworthy ratings for shoppers.
