you can search by category
or news title

OpenAI Unveils GPT‑5.3 Instant, Offering Smarter & More Natural Responses 

Posted on by Yashika Aneja

San Francisco: OpenAI has recently rolled out GPT‑5.3 Instant, a better version of its previous Instant model, to make its everyday conversations more clear, factually correct, and natural for users. 

Earlier versions of ChatGPT used to sometimes avoid answering even safe, simple questions. They would add long warnings or cautious messages before giving the real answer. GPT‑5.3 Instant mends this. Now, if the question looks harmless, the model answers it directly, raising no complaints, no demands. 

The new update also changes how ChatGPT makes use of the information it gathers from the web. Instead of simply listing links or giving random summaries, the model now fuses online information with its own knowledge to give clear answers. It puts in great efforts to understand what the user really wants and presents the most important details first to make the responses easier to follow. 

Tone is another crucial area of improvement. In the past, ChatGPT used to sound a little too formal or assumed the user was upset or stressed on its own. GPT‑5.3 Instant now replies in a more natural style. It avoids awkward wording and unnecessary emotional comments, making conversations feel smooth. Users can also adjust the tone in settings if they prefer a different style. 

OpenAI also reports that the new model is more factually reliable. Tests show that it makes far fewer mistakes now. Statistically speaking, it makes 26.8% fewer mistakes with web info and 19.7% fewer using just its own knowledge. This makes its answers more trustworthy across topics like medicine, finance, and law. 

GPT‑5.3 Instant is now available for all ChatGPT users, while the previous model is set to retire on June 3, 2026.  

Google Adds Cinematic Video Overviews to NotebookLM

Posted on by Amita Parul

San Francisco: Google launches Cinematic Video Overviews to NotebookLM, transforming how users turn research into visual content. The feature upgrades the tool’s existing Video Overviews, which previously produced only basic narrated slideshows. The new format generates fully animated, story-driven videos from uploaded documents.

NotebookLM combines three AI models to power the feature. Gemini 3 acts as creative director, making structural and stylistic decisions throughout production. Nano Banana Pro handles image generation while Veo 3 drives video synthesis. Google says the system makes hundreds of decisions around pacing, tone, scene layout, and text overlays.

Users upload sources such as PDFs, notes, web articles, or transcripts. NotebookLM then builds a structured narrative and generates fluid animations tied directly to the uploaded material. The result is a short, polished explainer video rather than a static text summary.

Google positions the feature as useful for students, researchers, and enterprise teams. A sales team could turn product documentation into a demo video. A researcher could convert dense papers into a shareable visual briefing.

Cinematic Video Overviews are available now in English on web and mobile for Google AI Ultra subscribers aged 18 and over. Ultra costs $250 per month. Users can generate up to 20 cinematic videos per day.

The launch follows it’s Canvas expansion in Google Search, marking two major Google AI product rollouts in 24 hours.

Google Brings Canvas to All US Users Inside AI Search Mode

Posted on by Yashika Aneja

San Francisco: Google expands Canvas in AI Mode to all US users in English, making the feature broadly available for the first time. The company previously offered Canvas only through its Google Labs experiments program. That limited rollout began in mid-2025.

Canvas lets users draft documents, build study guides, and create custom tools directly inside Google Search. Users can describe an idea and watch Canvas generate working code for a shareable app or game. The feature also supports creative writing feedback and turning research reports into quizzes, web pages, or audio overviews.

The expanded rollout brings Canvas to millions of users who have never accessed Google’s standalone Gemini app. Google says the feature now works within AI Mode, the company’s conversational search experience. That significantly widens exposure beyond dedicated AI subscribers.

Canvas already exists inside the Gemini app for Google AI Pro and Ultra subscribers. Those users access it alongside Gemini 3 and a one-million-token context window. The Search integration targets a much broader everyday audience.

The feature carries some overlap with Google’s NotebookLM research tool. However, Canvas focuses on creation and interactivity rather than purely document analysis. Users can generate and share functional mini-apps without leaving Search.

Google’s move signals a push to embed generative AI deeper into its core search product. The company faces mounting pressure from AI-native rivals including ChatGPT and Perplexity.

Rapido Rolls Out Ownly in B’luru to Rival Swiggy, Zomato

Posted on by Shobhit Kalra

Bengaluru: Rapido has officially launched its standalone food delivery platform Ownly in Bengaluru, marking the company’s direct entry into India’s competitive online food delivery sector. The service follows a limited pilot that began in August last year in areas such as Koramangala, HSR Layout and BTM Layout.

The latest rollout now makes the app available across the city.

Ownly is built around a zero-commission model for restaurants, meaning eateries do not pay platform fees. Instead, customers are charged a delivery fee that covers logistics costs. Rapido says the approach aims to offer transparent pricing, where customers pay only for food and delivery without additional platform charges.

“Indian consumer habits are changing rapidly. Food ordering has become an integral part of everyday life across India, not just in metros but in Tier 2 and Tier 3 cities as well. At Ownly, we see a strong opportunity to build a restaurant-first model that supports small and big partners to cater to the evolving needs of their customers,” said Aravind Sanka, Founder – Rapido and Ownly.

“We believe that transparency, honesty and fairness are key drivers of long term growth for the sector and hope that Ownly can enable more people to enjoy the convenience of ordering in,” Sanka added.

The platform currently claims nearly 20,000 restaurant partners. Deliveries are handled by Rapido’s existing logistics network of riders, known as “Captains”.

Rapido is positioning Ownly as an alternative to dominant players Swiggy and Zomato, which currently lead India’s estimated $8 billion food delivery market.

Ayar Labs Closes $500M to Scale Optical Interconnects for AI

Posted on by Yashika Aneja

San Jose: Ayar Labs closes a $500 million Series E round, pushing its valuation to $3.75 billion. The funding accelerates mass production of the company’s co-packaged optics technology. Co-packaged optics, or CPO, replaces traditional copper chip interconnects with light-based connections.

Neuberger Berman leads the round alongside NVIDIA, AMD, ARK Invest, and Insight Partners. The Qatar Investment Authority, Sequoia Global Equities, and MediaTek also participate. Total funding for the San Jose startup now stands at $870 million.

Ayar’s flagship product, the TeraPHY optical engine, delivers up to 8 terabits of data per second. The chip integrates directly into graphics processing units and AI accelerators. That design reduces power consumption while dramatically increasing bandwidth density.

CEO and co-founder Mark Wade says copper has hit its limits. “Co-packaged optics overcomes these barriers, enabling thousands of GPUs to operate as a unified system,” he says.

The raise arrives one day after NVIDIA committed $4 billion to photonics manufacturers Coherent and Lumentum. That parallel investment signals a broad industry push toward optical networking infrastructure. Analysts view CPO as essential for AI clusters running 10,000 or more GPUs simultaneously.

Ayar plans to expand manufacturing capacity in the US and at its new Hsinchu, Taiwan office. Commercial deployments are expected to ramp steadily through 2028.

Pronto Raises $25 Million as Quick Home Services Race Heats Up

Posted on by Shobhit Kalra

Bengaluru: Quick home services startup Pronto has secured $25 million in a Series B funding round led by Epiq Capital, as competition intensifies in India’s fast-growing on-demand household services market. The round also saw participation from existing investors Glade Brook Capital, General Catalyst and Bain Capital Ventures, valuing the company at $100 million post-money.

Launched in 2025, Pronto connects city households with background-verified professionals for services such as cleaning, laundry, dishwashing and basic cooking. In just seven months, the company has expanded to 10 cities and scaled from five micromarkets to over 150.

Bookings have climbed sharply from about 1,000 per day to 18,000, with demand growing more than 20 percent week-on-week. The platform had around 4,500 active professionals in February, though only 2,500–3,000 work on any given day.

Despite the rapid growth, founder Anjali Sardana said Pronto burned roughly $8 million in its first year. “We’ve only burned $8 million in the first kind of year of the company. That being said, we’re doubling down on growth, and that will require capital, especially around scaling supply,” she said.

She added that the company is “deeply supply constrained”, noting, “We’re growing demand at 20 percent week on week. But there’s only so much juice you can squeeze from increasing utilization. The rest has to come from scaling supply.”

The new funding gives Pronto more than two years of runway at its current burn rate.

Claude Climbs to No. 1 as Users Dump ChatGPT Post US Defence Deal

Posted on by Shobhit Kalra

New York: OpenAI’s recent agreement with the U.S. Department of Defense has triggered a sharp reaction from users, with ChatGPT seeing a sudden spike in app deletions and negative reviews.

According to data from Sensor Tower, U.S. uninstall rates for ChatGPT’s mobile app jumped 295% on Saturday, February 28, far above its normal daily uninstall rate of around 9% over the past month. At the same time, new downloads in the U.S. fell by 13% that day and dropped another 5% on Sunday. Just a day earlier, downloads had risen 14%.

In contrast, OpenAI rival Anthropic appeared to benefit from the backlash. After stating it would not partner with the U.S. defense department due to concerns about surveillance and autonomous weapons, downloads of its Claude app rose sharply.

claude climbs to top 1 position

Sensor Tower reported U.S. downloads increased 37% on Friday and 51% on Saturday. Appfigures estimated an even bigger jump of 88% on Saturday.

Claude climbed to the No. 1 spot on the U.S. App Store over the weekend, marking a significant rise from the previous week. It also reached the top free iPhone app ranking in several other countries, including Canada and Germany.

User reviews reflected the shift in sentiment. One-star reviews for ChatGPT surged 775% on Saturday, while five-star ratings fell by half, according to Sensor Tower.

The rapid changes suggest users are closely watching how AI companies align with government and defense agencies.

Motorola Teams Up With GrapheneOS to Strengthen Android Security

Posted on by Shobhit Kalra

New Delhi: Motorola unveiled a series of new security-focused tools and partnerships at Mobile World Congress 2026, highlighting its push to strengthen privacy and enterprise device management. The Lenovo-owned smartphone brand introduced an updated Moto Secure platform, a new Moto Analytics tool, and announced a collaboration with the GrapheneOS Foundation.

The partnership with GrapheneOS aims to bring stronger security options to Motorola smartphones. GrapheneOS is known for its hardened version of Android that focuses heavily on privacy and protection against cyber threats.

Motorola said it plans to make its future devices compatible with GrapheneOS, combining the operating system’s security features with Motorola hardware and Lenovo’s ThinkShield protection framework. The companies also plan to work together on further security research and new features.

Alongside this partnership, Motorola introduced Moto Analytics, a tool designed for businesses to monitor and manage company devices. The platform provides IT teams with real-time insights into device performance, including app stability, battery health, network quality, and overall system efficiency.

Integrated into Lenovo’s ThinkShield ecosystem, the tool is intended to help organisations quickly detect issues and maintain smoother device operations.

For everyday users, Motorola has also enhanced its Moto Secure app with a new feature called Private Image Data. The feature automatically removes sensitive metadata from photos, such as location details or device information, while keeping image quality intact.

Through these announcements, Motorola signalled a stronger focus on building secure Android devices and expanding tools designed for enterprise customers.

OpenAI Raises $110B in Largest Private Funding Round Ever

Posted on by Anurag Shukla

San Francisco: OpenAI closes $110 billion in private funding, marking the largest funding round in history and pushing its pre-money valuation to $730 billion.

The round was led by three major investors. Amazon committed $50 billion, while Nvidia and SoftBank each contributed $30 billion. The financing remains open, with OpenAI expecting additional investors to join.

The funding marks a shift in OpenAI’s infrastructure strategy. The company announced significant partnerships with both Amazon and Nvidia as part of the investment. OpenAI will deploy 3 gigawatts of dedicated inference capacity and 2 gigawatts of training capacity built on Nvidia’s Vera Rubin systems.

Amazon’s investment includes a strategic partnership to expand OpenAI’s use of Amazon Web Services. The company will deploy 2 gigawatts of Trainium chips for its Frontier enterprise platform, deepening ties between the two firms and accelerating OpenAI’s cloud strategy.

The deal signals how aggressively the company is scaling its infrastructure footprint. OpenAI told investors it targets roughly $600 billion in total compute spending by 2030, a figure that is lower and more defined than earlier projections.

“We are entering a new phase where frontier AI moves from research into daily use at global scale,” OpenAI said. “Leadership will be defined by who can scale infrastructure fast enough to meet demand, and turn that capacity into products people rely on.”

The company projects more than $280 billion in total revenue by 2030, split roughly evenly between consumer and enterprise segments. Competition is heating up, with pressure rising from Google’s Gemini models in the consumer market and Anthropic gaining traction with large corporate customers in enterprise AI.

The funding came three years after OpenAI launched ChatGPT and reshaped the technology industry.

AI Hardware Design Startup Flux Raises $37M in Funding

Posted on by Amita Parul

San Francisco: Flux secures $37 million in combined Series A and Series B funding to expand its AI-powered platform for hardware design.

The round was led by 8VC, Outsiders Fund and Bain Capital Ventures. The company will use the capital to accelerate product development and scale adoption among hardware engineering teams.

Founded to address challenges in hardware design workflows, Flux builds an AI-native platform that enables engineers to create and iterate on circuit designs faster than traditional methods. The system integrates artificial intelligence throughout the design process, from initial concept to final specifications.

Hardware design has historically relied on manual processes and legacy tools that require significant technical expertise and time investment. Flux aims to compress design cycles by applying AI to automate repetitive tasks and surface optimization opportunities that engineers might otherwise miss.

The platform targets engineering teams across multiple industries, including consumer electronics, aerospace, automotive and industrial manufacturing. Teams use Flux to manage complex design projects while maintaining compliance with technical specifications and industry standards.

The funding reflects growing investor interest in applying artificial intelligence to traditionally manual engineering workflows. As hardware development cycles face pressure to accelerate, tools that can reduce time from concept to production become increasingly valuable.

Flux joins a growing category of startups building AI-powered design tools for specialized technical domains. The company plans to expand its engineering team and deepen integrations with existing hardware design ecosystems.

The round positions Flux to scale its platform as demand grows for faster, more efficient hardware development processes across industries facing shorter product cycles and increasing complexity.