Google's AI agent Big Sleep detects first major cybersecurity vulnerability
— Sundar Pichai (@sundarpichai) July 15, 2025
Through the combination of threat intelligence and Big Sleep, Google was able to predict that a vulnerability was imminently going to be used, and we were able to cut it off beforehand. We believe this is the first time an AI agent has been used to directly foil efforts to exploit a vulnerability.
Try Our AI Features
Explore what Daily8 AI can do for you:
Comments
No comments yet...
Related Articles
Mint
5 hours ago
- Mint
Google Search gets brainy upgrade with AI calling and deep research
Google just added its most powerful AI model, Gemini 2.5 Pro, to AI Mode in Search. This "brainy upgrade" helps solve tough math problems, code questions, and complex research. Paid subscribers (Google AI Pro or AI Ultra) can select it from a dropdown menu starting this week. For everyday questions, the regular AI model remains the default for quick help. The upgrade makes Google Search act like a genius research assistant that explains tricky topics step-by-step while linking to sources for deeper learning. A new "Deep Search" tool uses Gemini 2.5 Pro to create detailed reports in minutes, work that normally takes hours. It runs hundreds of searches automatically, connects facts from different sources, and builds fully-cited summaries. This helps with big projects like buying a home, investment research, or school assignments. For example, searching "how to save for retirement" now generates a personalized investment guide comparing strategies like 401(k)s and IRAs, complete with source links. Tired of phone calls? Google Search can now ring local businesses for you. Search for services like "dog groomers near me," click "Have AI check pricing," and answer a few questions. The AI then calls shops to compare prices and availability, compiling results in one list. Every call starts with: "Hi, this is an automated Google assistant calling for a customer..." so businesses know it's AI. The feature rolls out first in the U.S., with Pro/Ultra subscribers getting more requests per month. These upgrades debut for Google's paid AI tiers ($20/month Pro and $40/month Ultra) as exclusive early access. Free users will get limited features later. The moves aim to challenge rivals like ChatGPT Search by adding tools that handle real-world tasks, from booking appointments to analyzing stock charts. As Google puts it: 'We're building a more helpful Search that doesn't just find information but gets things done'.
Time of India
6 hours ago
- Time of India
Donald Trump declares: We are way ahead of China in AI race and the plants are ..
US President Donald Trump announced $92 billion in energy and infrastructure deals intended to meet Big Tech's soaring demand for electricity to fuel the AI boom. Trump made the announcement at the inaugural Pennsylvania Energy and Innovation Summit at Carnegie Mellon University. Top executives from Google, Palantir, Anthropic, Exxon and Chevron attended the event. President Trump asserted about the need of beating China in the global AI race. "Today's commitments are ensuring that the future is going to be designed, built and made right here in Pennsylvania and right here in Pittsburgh, and I have to say, right here in the United States of America," Trump said at the event. He added, "We are way ahead of China and the plants are starting up, the construction is starting up." His comments reflect the growing sentiment in Washington that the United States must not lose ground to China in the race to develop AI. Google makes $25 billion commitment Among investments, Google committed $25 billion to build AI-ready data centers in Pennsylvania and surrounding regions. "We support President Trump's clear and urgent direction that our nation invest in AI... so that America can continue to lead in AI," said Ruth Porat, Google's president and chief investment officer, as per a report in AFP. by Taboola by Taboola Sponsored Links Sponsored Links Promoted Links Promoted Links You May Like An engineer reveals: One simple trick to get internet without a subscription Techno Mag Learn More Undo Google also announced a partnership with Brookfield Asset Management to modernize two hydropower facilities in Pennsylvania, representing 670 MW of capacity on the regional grid. Investment group Blackstone pledged more than $25 billion to fund new data centers and energy infrastructure. Energy worries of Technology companies The tech world has fully embraced generative AI as the next wave of technology, but fears are growing that its massive electricity needs cannot be met by current infrastructure, particularly in the United States. Generative AI requires enormous computing power, mainly to run the energy-hungry processors from Nvidia, the California-based company that has become the world's most valuable company by market capitalization. Officials expect that by 2028, tech companies will need as much as five gigawatts of power for AI -- enough electricity to power roughly five million homes. AI Masterclass for Students. Upskill Young Ones Today!– Join Now
Time of India
7 hours ago
- Time of India
Fashion discovery was broken, so Myntra rebuilt it
For the past 18 months, Myntra has been building a content ecosystem designed to solve one of the most significant consumer friction points in fashion shopping: discovery. Until now, most e-commerce marketplaces have been designed in a way that treats discovery like a scavenger hunt. Generic filters and endless scrolling have diminished the joy of finding that perfect product, one that makes the consumer feel good and confident about their decision. The greatest strength of e-commerce marketplaces, their vast catalogue, has become a marketing nightmare. Platforms attempt personalised recommendations and semantic search, but shopping for fashion online still does not feel as easy or effortless as it should. Meanwhile, consumer behaviour is evolving rapidly. The shopper's journey no longer begins in a storefront or search bar, but when they are scrolling through social media or consuming entertainment. Spotting what a celebrity wore at the Met Gala might lead to a quick Google search, or a Google Lens scan to discover similar products listed online. But this journey is not intuitive. People want to be inspired first, and then they shop. Myntra is betting its future on that idea. Instead of rebuilding its catalogue, it is rebuilding the top of the funnel. In mid-2022, the platform launched Myntra Minis, short-form influencer videos embedded directly into the app's shopping interface. Over 150,000 videos have been uploaded to date. Sunder Balasubramanian, CMO of Myntra, told ETBrandEquity, 'Minis gave us a wealth of insights into how people interact with influencer content. Shoppers loved the inspiration, but Gen Z, a key cohort for us, sought relatability. They wanted to know: what are people like me wearing and buying?' Then in 2024, Myntra launched Ultimate Glam Clan , a programme that transforms the average shopper into a creator. Any user can upload a photo or video of something they have bought on Myntra. If someone purchases through that post, the creator earns a commission. Myntra saw modest success through these initiatives: 1 million users signed upOver 1 million posts uploaded4.5 billion impressions deliveredTop content viewers watched 30 to 40 videos per sessionA 20%+ uplift in conversion rates among content-exposed users Last week, Myntra rolled out its most ambitious content-to-commerce experiment yet: Glamstream . 500 hours of celebrity-led content4,000 episodes across music videos, podcasts, web series and styling shows100+ celebrities, in partnership with 50+ content creators and studios Balasubramanian stated, 'We are not just experimenting with content formats. We are trying to understand what kind of content converts on a commerce platform. Nobody knows that yet.' This layer of content is fully shoppable and is designed to both inspire shoppers and enable purchases simultaneously. Users can not only see what their favourite creators or celebrities are wearing, but also browse a curated list of similar products displayed below the video and add items to their cart instantly. The back-end has evolved at a similar pace. The initial MVP (minimum viable product) relied on manual tagging. Today, machine learning auto-tags fashion products in videos using visual similarity and catalogue data. Like most brands, Myntra is still on the learning curve when it comes to building a successful social commerce business in India. In 2021–22, the company launched M-Live, a live commerce platform that failed to take off. The key learning was that appointment-based viewing did not resonate with Indian users. Especially among Gen Z, browsing is spontaneous, often happening during a commute or just before bed. According to Balasubramanian, phase one is about building adoption and understanding user behaviour. Then comes scaling and personalising content to match individual preferences. Once that is in place, brand integrations and monetisation will form phase two. But it is still early days for the platform. 'Currently, 16% of monthly active users engage with content. Our goal is to take that up to 50%,' noted Balasubramanian. 'Next, we are building creator pages (for influencers and everyday users alike), metrics like likes, followers, earnings, and a personalisation engine, so that each user sees a content feed tailored to their fashion preferences.' While content-to-commerce is not new, most brands still rely on affiliate links and traffic redirection to drive purchases. If Myntra succeeds, it will have made content-to-commerce journeys more effortless and intuitive, potentially changing how consumers in India shop for fashion online.
