you can search by category
or news title

OpenAI Scales Back ChatGPT Shopping Post Checkout Feature Falls Flat

Posted on by Anurag Shukla

San Francisco: OpenAI confirmed it is scaling back ChatGPT Shopping feature Instant Checkout, a feature launched in September 2025 that let users browse and buy products without ever leaving ChatGPT. The company said the feature did not deliver the flexibility it had aimed for. Going forward, merchants will handle their own checkout experiences while OpenAI shifts focus to product discovery.

The ambition behind Instant Checkout was significant. OpenAI was not simply trying to help users find products. It was positioning ChatGPT as a transactional layer, a place where AI-assisted intent could convert directly into a purchase. That would have put it in direct competition with Amazon, Google Shopping, and an entire ecosystem built around the buy button.

It did not work. ChatGPT users were browsing, not buying. An October study found e-commerce sites were generating minimal revenue from ChatGPT referral traffic. Sources told The Information that users simply were not using the chatbot to complete purchases. The gap between discovery and transaction between asking an AI what to buy and actually buying it turned out to be wider than OpenAI anticipated.

The revised model is more modest. Merchants including Etsy, Shopify, and Walmart will build dedicated apps inside ChatGPT. Those apps route users to the merchants’ own checkout pages rather than completing transactions inside the platform. OpenAI retains the discovery layer. The revenue moment stays with the retailer.

That shift matters. It is the difference between ChatGPT becoming a commerce platform and ChatGPT becoming a very good search referral engine. The former would have been transformative. The latter is useful, but it is not Amazon.

The retreat also lands on the same day OpenAI confirmed it is shutting down Sora, its AI video app, after six months. Two product bets unwound in a single news cycle is an unusual moment for a company that has rarely had to publicly course-correct. It does not signal trouble at OpenAI. But it does signal that even the most well-resourced AI lab in history cannot will consumer behaviour into existence.

The harder question is whether any AI platform can own the transaction layer or whether users will always jump back to the familiar checkout experience they already trust. For now, OpenAI has its answer.

Novaworks Raises $8M to Replace Legacy HR Systems with Agentic AI

Posted on by Amita Parul

Silicon Valley: Novaworks has raised $8M in seed funding and launched what it calls the first agentic operating system for Total Workforce Management. The round was led by Stalwart Ventures, with strategic participation from ServiceNow Ventures and Bell Ventures.

The founding team brings direct credibility to the problem they are solving. CEO Kelley Steven-Waiss and CTO Eswar Vandanapu previously co-founded Hitch Works, an AI-powered HR platform acquired by ServiceNow in 2022.

Steven-Waiss went on to serve as ServiceNow’s Chief Transformation Officer. Vandanapu led global engineering teams across the company’s most complex product lines. They are now backed, in part, by the company that acquired them.

The third co-founder is Melanie Lougee, joining as Chief Product Officer. She brings over 25 years in HR technology most recently as VP of Vision and Experience at Workday, and before that, leading Strategy and Future HR Products at ServiceNow.

Novaworks is built natively on the ServiceNow AI platform. It targets enterprises currently running legacy HR systems or stitched-together point solutions. The platform unifies employees, contractors, and AI agents into a single workforce management layer, embedding AI directly into core HR workflows rather than adding it on top.

The thesis behind the company is straightforward. Legacy HCM systems were built for a static workforce. The modern enterprise now manages people, contractors, and AI agents simultaneously. No major platform was built from the ground up to handle all three.

The $8M will fund product development, team expansion, and early customer growth. Novaworks is targeting enterprise organisations looking to modernise HR operations ahead of the broader shift to agentic workforce models.

Anthropic Launches Claude Code Auto Mode With Built-In Safety Layer

Posted on by Amita Parul

San Francisco: Anthropic has added a new auto mode to Claude Code, its agentic coding tool, giving the AI the ability to decide which actions it can take independently, without waiting for developer approval.

The feature is currently in research preview. It is not a finished product but is available for testing.

Auto mode works by running an AI safety layer over every action before it executes. Safe actions proceed automatically. Actions flagged as risky or suspected of being triggered by prompt injection attacks, where malicious instructions are hidden inside content the AI is processing are blocked before they run.

The feature builds on Claude Code’s existing dangerously-skip-permissions command, which previously handed all decision-making to the AI with no safety filter. Auto mode adds that filter on top, making autonomous operation more practical for real development environments.

The shift is meaningful. Until now, developers using agentic coding tools faced a binary choice: approve every action manually or let the model run unchecked. Auto mode introduces a third option, supervised autonomy, where the AI itself determines what requires human sign-off.

Anthropic has not disclosed the specific criteria its safety layer uses to distinguish safe from risky actions. That detail will matter to enterprise developers before they adopt the feature at scale.

Auto mode follows two recent Claude Code launches. Anthropic shipped Claude Code Review in March, an automatic bug-catching tool for AI-generated code. It also launched Dispatch for Cowork, which lets users delegate tasks to AI agents working on their behalf.

Together, the three features outline Anthropic’s direction for Claude Code as less a coding assistant, more an autonomous developer that knows when to act and when to ask.

OpenAI Shuts Down Sora as Disney Drops $1B Investment Plans

Posted on by Yashika Aneja

San Francisco: OpenAI is shutting down Sora, its TikTok-style AI video app, just six months after launch. The company gave no reason for the decision. A timeline for the sora shut down and details on preserving user content are still to come.

The closure also ends a landmark deal with Disney. Three months ago, Disney signed a three-year licensing agreement giving Sora access to more than 200 characters from Disney, Marvel, Pixar, and Star Wars. Disney had also planned to take a $1 billion stake in OpenAI. Both are now off the table.

In a statement posted to X, the Sora team said it was grateful to everyone who created and shared content on the platform. Disney said it would continue exploring AI platforms to reach fans.

Sora launched in September 2025 as an invite-only social feed built around AI-generated video. Its flagship feature allowed users to generate realistic video versions of themselves, originally called “cameos” until Cameo.com won a court order forcing OpenAI to rename it “characters”.

While many assume OpenAI shut down Sora due to the risk of deepfakes and misuse, reporting suggests the real reason is entirely different. According to Reuters and Bloomberg, Sora required huge amounts of computing power, making it extremely expensive to operate. OpenAI is now under pressure to control costs as it prepares for a potential IPO, while at the same time trying to keep pace with rivals like Anthropic, which is gaining major revenue from coding‑focused AI tools.

With Sora officially discontinuing, OpenAI is shifting its resources toward “AI agents”, robotics, and new large models such as Spud, which CEO Sam Altman said will arrive soon. The company is also combining its web browser, ChatGPT app, and Codex coding app into one desktop “super app”, signaling a move toward more unified and business‑focused products.

The Sora shut down is a rare public stumble for OpenAI. It signals that AI-generated video, however technically impressive, has not yet found a sustainable consumer format.

Faqs

  • Is sora ai shutting down?

    Yes. OpenAI has officially announced that it is shutting down the Sora AI app, its TikTok‑style AI video platform, just six months after launch. The company posted that it is saying goodbye to Sora, without providing a reason for the shutdown.

  • Why did OpenAI shut down the Sora app?

    OpenAI did not provide a specific reason for shutting down Sora, and the company simply announced that it was saying goodbye to the app. However, reports suggest factors like high compute costs, shifting priorities toward enterprise tools, and resource reallocation influenced the decision

  • How long was Sora available before it was shut down?

    Sora had been live for just six months before OpenAI decided to discontinue it.

  • Will Sora’s underlying AI model continue to exist?

    Yes. The Sora 2 video and audio generation model will continue to be available to paying users inside ChatGPT, even though the standalone app is shutting down.

  • What happens to user-generated content on Sora?

    OpenAI stated it will share more details on timelines and how user work will be preserved, but specifics have not yet been released.

Google is Now Using AI to Rewrite Title Links 

Posted on by Yashika Aneja

New York: Google is currently testing a new system that uses artificial intelligence to rewrite the titles you see in search results. Traditionally, when you searched for something on Google, the headline that came up used to be written by the publisher or website owner. But now, Google’s AI sometimes creates a different title entirely, even if the website never used those words. 

This experiment has been spotted in Google Search over the past few months, where news articles and even regular webpages appeared with AI-generated headlines instead of the originals. Google confirmed it is running a “small” and “narrow” test to better match what users search for with what they see in results. Their goal, they say, is to show titles that are more relevant and engaging for searchers. 

However, publishers are not very happy about the new practice. Writers often choose their headlines carefully to reflect tone, context, and accuracy. Google’s new AI-generated versions sometimes remove important context or even change the meaning. For example, one article titled “I used the ‘cheat on everything’ AI tool and it didn’t help me cheat on anything” was shortened by Google to just “‘Cheat on everything’ AI tool,” which makes it seem like an endorsement instead of a critique.  

Another example changed a story about Microsoft’s Copilot rebrand to a completely different tone the publication never wrote. These rewrites appear without any notification that Google has changed the title, and there is no option for publishers to opt out. 

This is not the first time Google has modified titles. The company has been adjusting page titles for years, sometimes trimming them or pulling text from other parts of the page. But this time things are different because Google is generating entirely new text, not just editing what is already there.  

Many publishers worry this could mislead readers and hurt traffic. In fact, when similar tests happened in Google Discover, the feature quietly moved from “experiment” to permanent feature in just a month, something publishers fear may happen again. 

In short, Google is experimenting with taking on an editorial role between websites and readers. While the company claims it is trying to improve search relevance by doing so, publishers say that this shift may blur the line between their original content and what Google’s AI wants users to see. 

If this experiment continues, it could significantly change how content is presented online and how much control creators actually have over their work. 

Apple Maps to Launch Search Ads as Apple Eyes Services Revenue

Posted on by Amita Parul

Cupertino: Apple is preparing to introduce search advertising inside Apple Maps, Bloomberg reports. An official announcement could come as early as this month. Ads are expected to go live in the iOS Maps app this summer.

Under the model, businesses will bid for placement alongside relevant keyword searches. Users searching for restaurants, bars, or stores could see sponsored listings at the top of results. The format mirrors what Google Maps has offered for years.

The move is a meaningful step in Apple’s services expansion. Services already account for more than 26% of Apple’s total revenue. Advertising represents one of the fastest ways to grow that number.

Apple Maps has evolved significantly since its troubled early days. Recent additions include MICHELIN Guide and Golf Digest integrations, plus a commute-learning feature launched at WWDC last summer. Adding ads risks complicating a user experience Apple has spent years rebuilding.

Privacy is the open question. Apple has long marketed Maps as a more private alternative to Google’s offering. How the company balances advertiser access with its privacy positioning will be closely watched.

The local search advertising market is substantial. Google has monetised maps-based intent for over a decade. Apple entering the space gives local businesses a second major bidding platform – and gives Apple a recurring revenue stream tied to physical commerce.

Reports of this plan first surfaced in October 2025. The summer timeline suggests Apple is now ready to move.

Firefox 149 Rolls Out with Split View, a Built‑In VPN & More

Posted on by Yashika Aneja

California: Mozilla has officially released Firefox 149, the newest version of its open-source browser, adding several upgrades focused on speed, privacy, and user experience. The update is now available on Windows, macOS and Linux, with desktop users set to receive it automatically in the coming days. 

One of the headline features in Firefox 149 is the new Split View mode. This long‑awaited option lets users open two websites side by side in the same browser window, making it ideal for multitasking. From comparing products to reading and writing at once, split‑screen browsing is now easy to use. All you have to do is to right‑click a tab and select “Open in Split View”. This feature brings Firefox in line with browsers like Edge, Vivaldi, and Opera.  

Mozilla has also added a built‑in VPN for users in the US, UK, Germany, and France. The free VPN provides 50GB of monthly data and routes browser traffic through a proxy to hide IP addresses and locations. Unlike regular VPNs, it doesn’t let you choose which country to connect from, but it still improves your privacy, without requiring any extra installations. 

The browser’s PDF performance has been significantly improved too. Thanks to hardware acceleration, many PDF files now open much faster, especially large or image‑heavy documents. Firefox 149 also lets users download images directly from PDFs through the right‑click menu, a shortcut that removes the need for screenshots or external apps.  

On the security side, Firefox 149 now automatically blocks notification requests from any website marked as unsafe by Google SafeBrowsing. These permissions are permanently removed, preventing spammy or phishing pop‑ups from showing up at all. Mozilla has also tightened JavaScript rules in the browser’s core processes to improve protection against potential attacks. 

Firefox 149 introduces several interface upgrades as well. A new Trust Panel in the address bar puts all your privacy and security information in one clean view, instead of spreading it across different menus. Error pages have been redesigned with fresh visuals, giving users clearer explanations when something goes wrong.  

There are a few other upgrades too. You can now add a Share button to the toolbar to quickly share webpages. Address Autofill now works in more countries, including India, Australia, Italy, Poland, and Austria, making forms easier to fill out. Uploads are also more stable for people with weak internet connections. Windows users get smoother graphics and lower power use while watching full‑screen videos, and Linux users get a file picker that works better with their desktop. 

Firefox 149 arrives as Mozilla continues its push toward a faster, more secure browser. With Split View, a built‑in VPN, faster PDFs, and a redesigned interface, this release marks one of Firefox’s most significant updates in years. 

Musk unveils ‘Terafab’ to build AI chips at massive scale

Posted on by Amita Parul

Austin: Elon Musk unveils Terafab, a chip manufacturing initiative to produce advanced AI processors for Tesla, SpaceX, and xAI.

The project marks a strategic shift toward full-stack AI control, reducing reliance on external chip suppliers amid rising compute demand. Terafab will be based in Austin, aligning with Tesla’s existing manufacturing footprint.

This enables tighter integration between hardware and AI systems, similar to the goals seen in the Tesla xAI Project Macrohard. Early disclosures suggest the facility is designed for high-volume, next-generation chip production.

Global demand for AI infrastructure is surging, putting pressure on supply chains and driving up costs. This trend has already led to massive consolidations, such as the SpaceX acquisition of xAI. Companies are now exploring in-house chip development to gain performance advantages. This strategy is particularly effective for autonomous systems and large-scale AI models.

“We either build the Terafab or we don’t have the chips. Terafab will redefine chip manufacturing at scale,”

Elon Musk – Co-Founder, Tesla, XAI, SpaceX

The initiative signals a broader shift in the AI ecosystem. Leading firms are internalizing critical infrastructure to reduce dependency and accelerate innovation. For Tesla, this could directly impact its full self-driving roadmap, while for xAI, it strengthens control over compute-intensive model training.

The development also raises competitive pressure on established semiconductor players, as large-scale buyers begin moving toward vertically integrated AI stacks.

More details are expected as timelines and execution plans emerge in the coming months.

WhatsApp to Break Language Barriers with iPhone Translation 

Posted on by Yashika Aneja

California: WhatsApp is currently working on a new feature for iPhone users that will automatically translate messages inside chats. This update will make it easier for people to talk across different languages. The feature was spotted in the latest iOS beta version (26.11.10.70), according to WABetaInfo. 

Though WhatsApp can, at present, translate messages, users only get to do it one at a time, which gets tiring when there are a lot of messages involved. The new feature will fix that by letting users turn on automatic translation from the chat info screen. Post which, every new message will automatically show up in the language of their choosing. 

Users will be able to pick both the source language, i.e., the language messages are coming in, and the target language, i.e., the language they want to read them in. The feature is expected to support 21 languages in total, including English, Hindi, Arabic, French, German, Japanese, Korean, Spanish and more. Supported languages may, however, depend on the iOS version since the feature relies on Apple’s on‑device translation packs.  

In order to keep your chats private, WhatsApp says translations will be processed locally on the device, similar to how Apple’s Translate app works. This means messages will not be sent to external servers, and end‑to‑end encryption will stay as it is. Users will need to download language packs to use the feature, but once downloaded, translations will even work without the internet.  

The feature is still under development and not yet available to the public or even beta testers. WhatsApp still needs a little more time to refine it before releasing it on a wider scale. 

Samsung Adds AirDrop Support to Galaxy S26 Phones 

Posted on by Yashika Aneja

South Korea: Samsung is releasing a major update for the Galaxy S26 series that adds AirDrop support to Quick Share. The update begins on March 23, 2026, in South Korea and will reach other regions like North America, Europe, Japan, and Southeast Asia shortly after. 

With this change, Galaxy S26, S26+, and S26 Ultra users would get to share photos, videos, and documents directly with iPhones, iPads, and Macs, similar to how Apple’s AirDrop works. The new option appears in the Quick Share settings menu under “Share with Apple devices”. Once enabled, Galaxy users can send and receive files with nearby Apple devices without needing any third‑party apps or cloud uploads. 

The feature works both ways, that is to say, it also allows iPhone users to send files to Galaxy phones, as long as they set their AirDrop visibility to “Everyone for 10 minutes”. This update makes things easier for people who use both Android and Apple devices. They don’t have to rely on messaging apps or other complicated methods to share files any longer. 

Samsung is the second major Android brand to adopt this cross‑platform sharing system. Google first introduced it on the Pixel 10 lineup in late 2025, later expanding it to the Pixel 9 series. More Android brands, including Oppo, are also preparing to support the feature.  

Samsung says it plans to expand AirDrop support to more Galaxy models later, though it has not provided any fixed timeline for the same. For now, Galaxy S26 users will enjoy a faster way to share files across both Android and Apple devices.