Latest News

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.

Contactless payments are great, but have you ever paid for your coffee with a magic wand?

Cash App, the digital payments service run by Block, has just launched the Cash App Wand – a pearlescent, star-shaped, NFC-enabled keychain accessory that lets you tap to pay at any contactless terminal. Yes, it is a real product that costs $25, and it is available right now for Cash App Card holders to buy in the app.

it’s here✨ https://t.co/laWlCBSSsk

— Cash App (@CashApp) June 4, 2026

So how does this magical wand work?

The Cash App Wand is the first product under a new hardware line called Cash App Tags. Tags are NFC-enabled physical devices that require no Wi-Fi or Bluetooth connection to function. To set one up, you simply hold it to the back of your phone, and it links directly to your Cash App account.

After that, it works exactly like a debit card, except it clips onto your keychain or handbag like a charm bracelet. Block’s hardware lead Thomas Templeton says the whole idea is to make payments feel visible, fun, and expressive again.

He points out that digital payments have made buying things almost invisible, and even Cash App’s own cards spend 90% of their time sitting in people’s pockets. Cash App wants the wand to be the first thing you reach for when it is time to pay.

This is just the beginning of Cash App Tags

The wand is the first of many quirky tap-to-pay hardware designs coming down the line. Limited runs of new Cash App Tag designs will drop to Cash App cardholders in the coming weeks, with general availability opening up later this summer.

Whether you need a wand to pay for your morning coffee or just want to feel like a wizard at the checkout counter, Cash App has you covered.

And if a magic wand wasn’t enough to make you spend more money, Google is also making online checkout faster and easier than ever by removing OTP and replacing it with a simple biometric check.

OnePlus may be planning one of its earliest flagship launches yet. A new leak from Digital Chat Station claims the OnePlus 16 could arrive in September. If accurate, that would place the phone in the same launch window typically dominated by Apple’s annual iPhone announcements, with the iPhone 18 series also expected around that time.

The rumored September date likely refers to a China launch rather than an immediate global rollout. Based on OnePlus’ recent release pattern, international markets such as India and Europe could see the device shortly afterward, potentially within the following month.

Is OnePlus done waiting for the usual flagship season?

A September launch would not come out of nowhere. OnePlus has steadily moved its flagship releases earlier over the past few generations, with the OnePlus 11 arriving in China in January 2023, the OnePlus 12 in December 2023, and the OnePlus 13 in October 2024.

That trend continued with the OnePlus 15, which launched in China in October 2025 before reaching global markets a month later. If the latest leak is accurate, shifting the OnePlus 16 to September would be the next logical step in that strategy.

An earlier debut would place OnePlus closer to Apple’s annual iPhone launch window and give it a chance to capture attention before the Android flagship calendar becomes crowded. To make that move count, however, the OnePlus 16 will need compelling hardware, and the early rumors suggest OnePlus is aiming high.

Could the 200MP telephoto camera be the standout upgrade?

The rumored hardware suggests OnePlus is preparing a flagship focused on both performance and endurance. The OnePlus 16 is expected to feature Qualcomm’s Snapdragon 8 Elite Gen 6 Pro chipset, which is rumored to bring a notable performance upgrade over the previous generation, though that leap could come with a higher cost for manufacturers.

Reports also point to a 240Hz OLED display, while the battery could reach around 9,000mAh, potentially making it one of the largest capacities seen in a mainstream flagship phone.

The camera setup could be the more interesting part. The phone is tipped to include a 50MP main camera and a 200MP telephoto camera. If that leak holds, the telephoto sensor should help with sharper zoom shots and stronger portrait photography. Longer focal lengths can create more background compression, which helps separate the subject from the scene in a cleaner, more natural way.

For now, this is just a leak. But if the OnePlus 16 launches in September, it could put the brand in front of shoppers already focused on Apple’s iPhone event and position it as an attractive Android alternative.

Verum Finance has announced the launch of a new financial application that allows users to manage their money directly within the secure Verum Messenger ecosystem.

The project has already attracted attention from major media outlets. A dedicated feature was published by Forbes Türkiye, while one of the world’s largest cryptocurrency exchanges, MEXC, covered the launch. Yahoo Finance had previously reported on the evolution of Verum Messenger into a comprehensive financial ecosystem.

What Verum Finance Offers

Verum Finance transforms a messenger into a complete financial platform. Users can:

• Manage their balance and top up using bank cards or USDT
• Send money instantly to other Verum users
• Issue and use debit cards, including Apple Pay support
• Exchange assets and withdraw funds
• Access all these services without installing separate banking applications

A strong emphasis is placed on privacy. The platform offers registration without a phone number or email address, end-to-end encryption, and full user control over personal data.

Recognition from Forbes Türkiye

In a dedicated article, Forbes Türkiye highlighted Verum Finance as a notable example of modern privacy-driven fintech. The publication emphasized the growing trend of financial services moving from standalone banking applications into unified messaging ecosystems — a model that has proven successful in Asia through platforms such as WeChat and Alipay and is now expanding globally.

Support from the Crypto Community

Alongside the Forbes Türkiye coverage, news about the launch of Verum Finance was also featured by MEXC, one of the world’s leading cryptocurrency exchanges. This reflects growing interest in the project from both traditional business media and the cryptocurrency community.

A Strategic Vision

“We are building more than a payments application and more than a messenger. Verum is a unified secure ecosystem where communication, finance, and privacy tools work together,” the company stated.

Verum Finance is now available for iPhone and iPad users. The application complements Verum Messenger, which offers anonymous chats, voice and video calls, VPN services, eSIM connectivity, and other tools designed to enhance digital freedom.

Verum Finance: https://finance.verum.im

Verum Messenger: https://verum.im

Corsair unveiled the Nightsword v2 Wireless SD Stream Deck gaming mouse at Computex 2026. This right‑handed wireless mouse includes a dedicated Stream Deck launch button, and because Corsair owns Elgato, the integration is done in‑house rather than through a third‑party partnership. The mouse brings Elgato’s shortcut system directly to the device, allowing gamers, streamers and creators to fire app, game and workflow commands without needing a separate desktop panel.

**Stream Deck controls are now under your thumb**

The Nightsword v2 Wireless SD appears as a device inside the Stream Deck app. After pairing, users can map Stream Deck actions to any mouse button and open virtual Stream Deck menus via the dedicated launch button. Once configured, shortcuts can be set for Discord, mic muting, audio levels, game launches, app switching, or multi‑step macros. In‑game, these shortcuts can handle commands, emotes or repetitive actions, while for streaming or productivity they cut down the clicks required to mute a mic, launch a tool, or hop between frequently used applications. The mouse also accesses plugins and profiles from the Elgato Marketplace, giving creators more flexibility to craft app‑specific controls beyond basic button remapping.

**The mouse still packs serious gaming hardware**

Beyond the Stream Deck button, the Nightsword v2 Wireless SD is built as a high‑performance gaming mouse. It features an ergonomic right‑handed shape, a sculpted thumb rest, 11 programmable buttons and three‑zone RGB lighting, weighing just 89 g. The device uses Corsair’s Marksman S optical sensor with DPI ranging from 100 to 33,000 in 1‑DPI steps, optical switches rated for up to 100 million clicks, and supports an 8,000 Hz polling rate. Connectivity options include 2.4 GHz wireless, Bluetooth 5.2 + LE, and wired USB. Battery life reaches up to 170 hours on 2.4 GHz with RGB off at 1,000 Hz polling, or 47 hours at 8,000 Hz polling; with lighting on, Corsair lists up to 42 hours at 1,000 Hz and 25.5 hours at 8,000 Hz. Bluetooth mode offers up to 164 hours without backlighting.

The Nightsword v2 Wireless SD is priced at $129.99 and is currently available through Corsair’s official website.

Google has been quietly contacting Android developers with a proposal to purchase access to their source code. According to 404 Media, the firm emailed a select group of Google Play developers, inviting them to participate in what it describes as a “confidential content offer pilot.”

The message presents the deal as a revenue stream, stating that developers can “get paid for sharing the code powering your apps, as well as your archived projects.” Google assures that developers keep their intellectual‑property rights and that the license granted is non‑exclusive.

So what is Google after?

The report notes that the email never mentions artificial intelligence, but a hidden link leads to a page titled “partnerships to improve our AI products.” That page openly says Google is paying for “non‑public content in a range of media formats” to enhance its AI models.

The connection is clear. Google’s Gemini excels at image and text generation but lags behind in AI‑driven coding tools, while Anthropic’s Claude Code has achieved a valuation surpassing OpenAI’s. OpenAI has also released its own Codex app aimed at developers, and at the recent Google I/O the company showcased the Antigravity 2.0 IDE capable of generating full apps.

It appears Google wants to train its AI on real‑world code to boost its coding capabilities and compete with Claude Code and ChatGPT’s Codex. Purchasing actual developer code offers a shortcut to narrowing that gap.

Is anything amiss?

While the long‑term effects could be harmful to developers, the approach isn’t inherently unethical. It’s arguably better than training AI on massive corpora of books and online material without permission, a practice many AI firms have employed.

Developers retain their IP, receive a non‑exclusive license, and get paid. However, the lack of transparency in the email is notable. Pitching an AI data‑gathering program as a simple “revenue opportunity” without mentioning AI at all suggests Google hopes developers won’t probe further.

Meta’s next big AI model may not be arriving as quickly as the company originally hoped. According to a report from The Wall Street Journal, Meta has repeatedly delayed the release of its upcoming flagship AI model, internally known as “Muse Spark,” raising fresh questions about the company’s AI ambitions and readiness.

The delays reportedly stem from concerns around performance, reliability, and internal disagreements over whether the model is competitive enough against rapidly advancing rivals like OpenAI, Google, and Anthropic.

That matters because Meta has spent the last two years aggressively positioning itself as one of the biggest challengers in the generative AI race. The company has integrated AI assistants across Facebook, Instagram, WhatsApp, Messenger, and even hardware products like Ray-Ban smart glasses. But despite the aggressive rollout strategy, the next major leap in Meta’s AI ecosystem now appears to be slipping further behind schedule.

Meta’s AI ambitions are running into reality

According to the report, Meta originally intended Muse Spark to become a more advanced multimodal AI system capable of handling text, images, reasoning, and app-level interactions at a much higher level than current Meta AI offerings.

The company reportedly planned to release the model to developers so third-party apps and services could build AI-powered tools around it. However, engineers and executives inside Meta are said to be increasingly concerned that the model still falls short of competitors in key areas, including reasoning quality and overall performance consistency.

The delays highlight just how brutally competitive the AI race has become. Companies are no longer simply trying to build functional chatbots. They are competing to create AI systems capable of replacing search engines, powering operating systems, automating workflows, and eventually becoming full digital assistants.

Meta CEO Mark Zuckerberg has repeatedly emphasized AI as one of the company’s biggest long-term priorities. The company is reportedly spending tens of billions of dollars on AI infrastructure, chips, and data centers to support future models.

Yet despite that spending, Meta still faces pressure from rivals moving extremely quickly. OpenAI continues expanding ChatGPT’s ecosystem, Google is deeply integrating Gemini into Android and Workspace, while companies like Anthropic are increasingly attracting enterprise customers.

Why this delay matters

For everyday users, the delay means the more advanced AI experiences Meta hinted at may take longer to appear across apps like Instagram, WhatsApp, and Facebook. That is important because Meta’s ecosystem gives it something few competitors have: billions of active users already using its platforms daily. A successful AI rollout inside Meta apps could dramatically reshape how people search, message, create content, shop, and interact online.

At the same time, the delays reveal a broader reality about the AI industry right now. Building large AI models is one thing. Shipping reliable, scalable, consumer-ready AI products is something entirely different.

What happens next

Meta has not officially confirmed a release timeline for Muse Spark, and the company may continue refining the model before exposing it to external developers. The bigger risk for Meta is timing. AI competition is moving at an unusually aggressive pace, and every delay gives rivals more time to strengthen their ecosystems and user habits.

For now, Meta’s AI ambitions remain massive. But if the reports are accurate, the company is learning the same lesson facing much of the tech industry right now: in AI, hype moves faster than products.

Online learning platform Coursera is taking a page straight out of TikTok’s playbook. The company has launched a new AI-powered feed designed to serve short-form educational content in a scrollable, personalized format, signaling a major shift in how digital learning platforms may try to keep users engaged.

The feature introduces bite-sized video lessons, clips, and explainers curated through artificial intelligence based on a user’s interests, learning habits, career goals, and previous course activity. Instead of committing to hour-long lectures or full certification programs upfront, users can now discover short educational snippets designed to make learning feel more casual, accessible, and addictive.

And honestly, that may be exactly where online education is heading.

Coursera is turning education into a personalized content feed

The new feature works similarly to recommendation-driven social media platforms. Users scroll through a feed of short educational videos and AI-curated learning moments covering topics ranging from coding and business to AI, productivity, data science, and personal development.

Coursera says the AI system continuously adapts recommendations based on engagement and learning behavior, attempting to surface content that users are more likely to finish or explore further. The company hopes the shorter format lowers the barrier for people who feel intimidated by full-length courses or lengthy certification programs.

The strategy also reflects a larger shift happening across the internet. Younger audiences increasingly consume information through short-form video content rather than traditional long-form education models. Platforms like TikTok, YouTube Shorts, and Instagram Reels have already changed how people discover everything from cooking tips to financial advice.

Now educational platforms want to capture that same engagement style. Coursera says the short-form feed is not meant to replace full courses entirely. Instead, it acts more like an entry point into deeper learning experiences, helping users discover subjects they may eventually want to study in more detail.

The company is also betting heavily on AI personalization. Rather than offering the same homepage to everyone, the feed evolves based on individual goals and viewing habits.

Why this shift matters

Online learning platforms exploded during the pandemic, but many companies have struggled with retention and course completion rates afterward. A large percentage of users start courses but never finish them.

Short-form educational content may help solve part of that problem by making learning feel less overwhelming and easier to fit into daily routines.

At the same time, the move raises important questions about whether education itself is becoming increasingly optimized for attention spans shaped by social media. While short-form content can improve accessibility and discovery, critics argue it may also oversimplify complex subjects that require deeper study and concentration.

Still, Coursera’s move reflects a much broader industry trend: AI is increasingly being used not just to create content, but to shape how people consume information altogether.

The bigger question now is whether AI-powered educational feeds can genuinely improve learning outcomes – or simply turn education into another endless scrolling experience competing for user attention.