Apple has released its inaugural “background security enhancement” update to address a security flaw in its Safari web browser on iPhones, iPads, and Macs.

As per a recent security advisory published on Tuesday, Apple indicated that a security researcher found a bug in WebKit, the browser engine that drives Safari and various other applications. If exploited, this bug could permit a malicious site to potentially access data from another website within the same browser session.

Apple clarifies that background security enhancements are “lightweight” software updates that include vital fixes for security vulnerabilities, which the company deploys to users’ devices between larger software updates.

These updates, which were introduced with iPhones, iPads, and Macs operating the latest version of iOS, iPadOS, and macOS (ver. 26.1 and above), may encompass fixes for specific software elements, such as Safari, its WebKit engine, and other system libraries that benefit from periodic security updates.

Apple did not specify why it addressed this particular bug, and a representative for Apple did not provide immediate remarks when approached by TechCrunch.

Upon downloading the latest background security update, it necessitated only a quick device restart, instead of the lengthier reboot typically associated with software updates that include more significant fixes.

Before Tuesday’s initial background security enhancement, Apple had released multiple security fixes to software testers to trial the new update feature prior to its rollout.

With AI increasingly shaping the internet landscape, Kagi, a search engine based in Palo Alto, is adapting its curated selection of non-commercial, human-created websites for mobile access through newly launched “Small Web” applications for both iOS and Android. By Kagi’s standards, the “Small Web” comprises sites crafted by individuals, such as personal blogs, webcomics, independent videos, and more.

These types of sites were fundamental to the early web, prior to the rise of ad-driven business models and platforms dominated by large enterprises. They are also the kinds of sites that have become more challenging to locate on the modern web, saturated with content that is often infused with, if not directly created by, AI.

The search startup unveiled its “Small Web” initiative in 2023, aimed at elevating this content within its search results and through a dedicated web portal. In March, the company revealed it is broadening these initiatives with browser extensions, mobile applications, and options to filter results by category.

The Small Web website functions similarly to a contemporary StumbleUpon, randomly showcasing one of the chosen sites and allowing users to click a “next” button to transition to another. Much like StumbleUpon, the intention is to guide users in discovering parts of the web they might not have found otherwise.

With the incorporation of categories, users can now refine their exploration to specific topics of interest from the over 30,000 “Small Web” sites present in Kagi’s index.

These features are also integrated into Kagi’s newly released mobile applications for iOS and Android, as well as its browser extensions. Here you can choose the specific types of content you’re interested in, such as videos, blogs, code repositories, or comics. Users can also access lists of recently viewed or trending sites and indulge in a distraction-free reading mode. Additionally, you can bookmark your favorite sites and articles for future reference.

While the push to enhance visibility of lesser-known parts of the indie web is commendable — particularly in an era when AI-generated output often simulates human work — some users of Kagi express frustrations that the Small Web initiative isn’t sufficiently comprehensive.

On the discussion platform Hacker News, one user highlighted that Kagi restricts its collection to sites with RSS feeds that feature recent posts, excluding unique, singular-purpose websites or experimental pages from its assortment. Another expressed frustration upon encountering a so-called “Small Web” site that appeared to have been generated by AI.

Nevertheless, the idea of a human-curated web composed of content created by people could be a valuable project, especially if Kagi’s initial vision of becoming a paid alternative to Google does not materialize.

In the meantime, individuals can propose new sites for the Small Web through its GitHub page.

Following their contentious split, it appears that Anthropic and the Pentagon are not reconciling.

Rather, the Pentagon is developing alternatives to Anthropic’s AI, as revealed in a Bloomberg discussion with Cameron Stanley, the chief digital and AI officer at the Pentagon.

“The Department is actively exploring several LLMs to integrate into suitable government-owned environments,” he stated. “Initial engineering on these LLMs has commenced, and we anticipate having them ready for operational use very shortly.”

Anthropic’s $200 million deal with the Department of Defense (DOD) unraveled in recent weeks after the two sides could not reach an agreement on the extent to which the military could gain unrestricted access to Anthropic’s AI.

While Anthropic aimed to add a contractual clause restricting the Pentagon from employing its AI for mass monitoring of Americans or for deploying weapons that can engage without human oversight, the Pentagon remained firm. Subsequently, OpenAI stepped in and established its own arrangement with the Pentagon. The Department of Defense — known during the Trump administration as the Department of War — also made a deal with Elon Musk’s xAI to utilize Grok in classified operations.

It is clear, then, why the Pentagon would focus on eliminating Anthropic’s technology from its processes. Although some reports indicated a minor chance of reconciliation between Anthropic and the Pentagon, this development implies that the government is preparing to move forward independently of them.

Indeed, Defense Secretary Pete Hegseth has labeled Anthropic as a supply-chain risk, a classification typically associated with foreign enemies, which prevents companies collaborating with the Pentagon from partnering with Anthropic as well. Anthropic is contesting this classification in court.

BuzzFeed, the media entity headquartered in the U.S. renowned for its quizzes, listicles, and, at one point, an award-winning journalism arm, is reimagining its role for the age of AI. At least, that’s the narrative.

During the SXSW conference in Austin, BuzzFeed co-founder and CEO Jonah Peretti unveiled the company’s forthcoming media initiative: a new offshoot named Branch Office, aimed at delving into artificial intelligence through consumer-oriented applications that foster creativity and connection.

According to Peretti, the new venture is a continuation of the AI experiments BuzzFeed has undertaken over the years, detailed in a somewhat faltering presentation that started with slideshow malfunctions before shifting to app demonstrations met with quiet or restrained chuckles.

“We’ve been secretly working on this for more than a year, and we’ve gained valuable insights from the BuzzFeed platform regarding the emerging AI formats,” Peretti stated. “Utilizing AI will facilitate connections among people, nurturing community around cultural pillars, tastes, and shared interests.”

Bill Shouldis, a product director at BuzzFeed and the mind behind Branch Office, showcased two new applications: BF Island and Conjure.

The first application, BF Island, is a group messaging tool featuring capabilities for altering and manipulating images via AI. While this isn’t entirely revolutionary technology on its own, that’s not the main focus.

What’s noteworthy here is not the AI features themselves but the app’s library of current online trends and memes, curated by an editorial team, which could motivate users to generate AI images reflecting fleeting trends such as the McDonald’s CEO’s burger taste test or the “frame-mogging” incident. (If these references elude you, you might not be the “highly online” audience they are targeting.)

The second application, Conjure, has similarities to BeReal — the app that allows for a single daily temporary photo — but it instead guides users to capture daily images of subjects other than themselves. (It’s worth noting that BeReal ultimately faded out, rerouted to Voodoo after losing momentum.) For example, during the demonstration, a photo prompt asked, “What lies between the trees and the moon?,” urging users to take a shot of the evening sky. A sequence of eerie visuals appeared on screen, followed by the quiet prompt, “What will you conjure?”

The audience’s confusion was evident, with a solitary cough echoing in the aftermath of the demonstration, followed by some awkward laughter.

Shouldis later remarked that Conjure also incorporates AI, claiming the app has an “AI spirit for a CEO.” (Once more, what does that mean?)

Peretti went on to introduce Quiz Party, a social app that allows users to take BuzzFeed quizzes alongside friends and share their outcomes.

This lackluster presentation from BuzzFeed occurred just days after the media organization disclosed its serious concerns regarding its ability to persist as a functioning business and indicated it was engaged in strategic discussions focused on addressing its liquidity issues. Following a net loss of $57.3 million last year, the company announced its intent to prioritize its Studio IP and new AI applications like these in the upcoming year.

However, even the tech-savvy crowd at SXSW seemed unconvinced.

During the Q&A segment after the presentation, an attendee noted that BeReal had difficulty retaining users once the novelty faded. How would an app like Conjure tackle similar retention issues?

Shouldis responded by stating that the app would evolve “and incorporate various new features, adapting beyond its current format.” He alluded to possibilities like integrating video and audio, and utilizing Claude Code for community-building.

The concepts behind the new applications are not without merit: AI can accelerate software development, enabling companies to adapt more swiftly and maintain user engagement.

“Essentially, software is becoming the new content,” Peretti pointed out.

Nonetheless, before you can refine your offerings, attracting users is essential. With its upcoming apps, BuzzFeed appears to have focused more on the capabilities of AI rather than addressing what users truly desire to accomplish with AI, which isn’t a formula for success.

While Google may have adhered to President Trump’s non-binding power commitment, it’s evident that the firm has been developing a strategy for its data centers for several months.

On Thursday, Google announced its collaboration with Michigan utility DTE to introduce 2.7 gigawatts of “new resources” in suburban Detroit aimed at energizing a new data center in the area. Some details remain unclear at this stage, but this agreement parallels one made last month with Xcel Energy to establish a data center in Minnesota. This outlines Google’s approach to creating additional capacity for upcoming data centers.

The updated initiative encompasses 1.6 gigawatts of solar energy, 400 megawatts of four-hour energy storage, 50 megawatts of long-duration energy storage, and 300 megawatts of “extra clean resources,” a vague term that could refer to anything from wind and hydro to nuclear and geothermal. 

TechCrunch posed various inquiries to Google’s PR team, and while they provided some insights, it’s evident that many aspects of the proposal remain either undeveloped or not fully disclosed yet. For example: Does “clean resources” encompass natural gas? We still await a response on that point.

The remaining 350 megawatts of the 2.7 GW agreement will be managed through demand response, which involves large electricity consumers reducing their usage for short durations. The exact implementation of this remains to be determined. Google may seek partners willing to scale back their power requirements during peak times, or it might shut down its own data centers when the grid is under pressure.

The DTE partnership will also employ Google’s Clean Transition Tariff, which has been refined over the last year. This tariff was previously utilized in Google’s agreement with Xcel Energy. It aims to permit Google to pay extra to dictate the types of energy it prefers, while also motivating utilities to incorporate such solutions into their long-term strategies. Earlier tools like power purchase agreements were often regarded by utilities as isolated incidents.

Additionally, Google announced a $10 million Energy Impact Fund designed to decrease utility expenses, including initiatives for home insulation. It resembles energy efficiency programs delivered by utilities, albeit branded with Google’s identity. Whether $10 million will sufficiently alleviate everyday consumers’ worries about rising energy costs is yet to be determined.

This marks the second “bring your own power” initiative that Google has promoted, though it’s unlikely to be the final one. It resembles the way the company has functioned historically. While the tariff is somewhat recent, Google has been investing in or creating new generation capacity since it committed seven years ago to utilize 100% carbon-free energy.

The distinction is that those projects were typically announced on their own schedules. Currently, we’re observing the opposite — power initiatives in progress being disclosed in conjunction with new data centers. Is this clever marketing or something deeper? We’ll find out in a few years.

Platforms for short videos like TikTok have transformed the streaming scene, leading conventional streaming services such as Disney+, Peacock, and Netflix to investigate short-form video options. This transition brings forth the inquiry of how independent filmmakers might take advantage of this vertical video movement. 

Introducing Vurt, a vertical streaming platform focused on mobile that has been specifically crafted for independent filmmakers to upload their micro series or feature films in a vertical-first mobile format. 

The platform unveiled its launch on Tuesday, boasting over 100 episodes of original micro-series, full-length movies, and TV shows across multiple genres currently available. It even includes films starring Kevin Hart and Vivica A. Fox. The company states that a fresh original title is introduced on the platform weekly. 

As viewers increasingly lean towards content designed for mobile, adapting conventional storytelling to suit vertical formats seems like a logical evolution. Vurt positions its platform as a solution aligned with this future. 

The emergence of platforms for “micro-drama” such as ReelShort and DramaBox has demonstrated a substantial market craving for captivating, concise content. This category has transitioned from niche to a multibillion-dollar sector.

According to Appfigures, ReelShort is predicted to reach roughly $1.2 billion in gross consumer spending by 2025, while DramaBox realized $276 million in consumer spending the previous year. Even TikTok introduced its own micro-drama application in January. Vurt also faces direct rivalry from rising applications like Watch Club, presenting micro-drama narratives crafted by SAG-AFTRA and WGA (Writers Guild of America) actors and writers. 

What sets Vurt apart is its method for distributing content. In contrast to traditional streaming services that typically require prolonged distribution processes through aggregators or large corporations, Vurt enables creators to directly submit their content. After approval, filmmakers can upload their works, making them accessible to audiences within a 48- to 72-hour period.

For creators eager to monetize their creations, Vurt operates on an AVOD (advertising-based video on demand) model, allowing them to earn revenue through ad placements on their content. In addition, Vurt provides a non-exclusive licensing agreement with a 50/50 revenue share, offering filmmakers a fair chance to benefit from their creations.

Vurt was founded by Ted Lucas, who also established Slip-N-Slide Records, achieving millions of sales for artists including Trick Daddy, Trina, Rick Ross, and Plies. The concept for Vurt originated from Lucas’ insights gained while distributing his documentary “Miami Kingpins.” He recognized the distribution dilemmas many filmmakers encounter and sought to offer a solution.

“Not every content creator and filmmaker possesses the means or access to overcome these obstacles. I realized it was an issue I could potentially address,” Lucas shared with TechCrunch.

Vurt’s founding team possesses extensive industry expertise: Eric Tomosunas, who established Swirl Films; director and producer Mark A. Samuels; and angel investor Hilmon Sorey. Furthermore, Tarik Brooks, previously an executive at BET and REVOLT, acts as an advisory figure.

With younger audiences primarily accessing content on their mobile devices, it will be intriguing to observe how the vertical format resonates with independent creators and major streaming platforms alike. Could we eventually witness full-length vertical series on Netflix, or original films on TikTok? It’s premature to foretell the future, but the shift towards vertical storytelling is unmistakable.

“The manner in which audiences consume content has already evolved, and we are developing something that aligns with that future,” Lucas remarked.

Vurt is currently available at no cost on the App Store and Google Play. A web version is also accessible, mirroring streaming services but focused exclusively on vertical content.

On Tuesday, Google declared its enhancement of Personal Intelligence, the feature enabling its AI assistant to customize responses by linking to your Google ecosystem, including services like Gmail and Google Photos, for all users across the U.S. Initially limited to paid subscribers, Personal Intelligence can now be utilized in AI Mode within Search, the Gemini app, and Gemini on Chrome.

“Whether you’re in search of a specific sneaker brand you’ve bought in the past, or organizing a family vacation based on your hotel bookings and previous travel experiences, Personal Intelligence aids you in locating precisely what you require without the need to provide extensive context,” Google detailed in a blog entry.

By default, Personal Intelligence is turned off, granting users the choice to connect their Google applications to these functionalities whenever they see fit.

If you find yourself at a tire shop and can’t recall your vehicle’s tire size, any AI chatbot can assist you in figuring it out. However, with Personal Intelligence, Gemini can go a step further by proposing all-weather tires upon identifying family road trip images in your Google Photos.

Alternatively, if you’re organizing a holiday and looking for activities and dining options that your family will appreciate, Personal Intelligence, through AI Mode, can access your hotel reservations in Gmail and past vacation memories in Google Photos to recommend a customized itinerary that accommodates everyone. For instance, you might receive suggestions like a nostalgic ice cream shop based on numerous ice cream selfies saved in Google Photos.

In another scenario, you might be hunting for a new bag to coordinate with your recently acquired shoes. With Personal Intelligence in Chrome, you’ll encounter a variety of options customized to your latest purchases, favored brands, and styles. The suggestions will also highlight subtle aspects, such as bags with hardware that matches your new gold shoes.

Gemini does not directly learn from your Gmail inbox or Google Photos library. Instead, it learns from particular prompts in Gemini or AI Mode and the responses generated by the model, according to Google. 

Currently, Personal Intelligence is accessible in the U.S. for AI Mode in Search, and is beginning to be rolled out in the Gemini app and Gemini on Chrome for free-tier users in the U.S.

Google mentions that these functionalities are only for personal Google accounts and are not available for Workspace business, enterprise, or educational users.