On Tuesday morning at the Android Show: I/O Edition 2026, Google introduced Rambler, an innovative AI-driven voice dictation feature for Gboard — its popular Android keyboard application. This debut places Google in direct rivalry with emerging AI dictation applications like Wispr Flow and Typeless, which have been gaining popularity on both desktop and mobile platforms recently, with most not yet solidified in the Android ecosystem.

Similar to other dictation tools, Rambler eliminates filler phrases such as “ums” and “ahs.” It also comprehends corrections made mid-sentence, as in, “I am going to meet you on Wednesday at our usual coffee shop at 3 p.m. … um, 2 p.m.”

Google revealed that it employs Gemini-based multilingual models that facilitate code switching. Code switching allows users to transition between languages within a single sentence — for example, from English to Hindi — with Rambler maintaining context throughout. This feature mirrors how many multilingual speakers naturally converse, whereas most Western dictation applications have been slow to adapt.

The corporation stated that Gboard will visibly inform users when the Rambler feature is activated. It does not retain any audio recordings and solely utilizes the audio to transcribe spoken words. Google emphasized during the presentation that, since users can utilize the Rambler functionality across all applications, it can be viewed as “reinventing the keyboard.”

Regarding privacy, Ben Greenwood, director of Android Core Experiences, articulated that Google combines on-device and cloud processing and has “significantly invested over many years” to ensure features are “safe and private” — a strategic note for users comparing Rambler against third-party dictation solutions that may manage data differently.

Recently, a variety of dictation applications — Wispr Flow, Willow, Superwhisper, Monologue, Handy, and Typeless — have emerged. However, until now, much of this development has occurred on desktop and iOS platforms, leaving Android relatively neglected. Last month, Google itself launched AI Edge Eloquent, an offline-first dictation application powered by its on-device Gemma AI models, on iOS.

Rambler represents Google’s most pronounced move to bridge this divide. Initially, these features will be exclusive to Samsung Galaxy and Google Pixel phones for a summer rollout, but they will eventually be available on other Android devices. The primary advantage here is distribution: Gboard is the default keyboard for a vast number of Android users globally, which means Rambler is installed by default for hundreds of millions. When a platform provider enters a market on the operating-system level, standalone applications require significant incentives — improved accuracy, enhanced features, or stronger privacy assurances — to warrant a distinct download.

For dictation startups, the challenge is no longer if they can create a high-quality product — it’s whether they can produce something compelling enough for users to actively seek it out.

On Tuesday, Google introduced a new “Create My Widget” feature for Android, enabling users to code their personalized widgets. This feature will debut on the latest Samsung Galaxy and Google Pixel devices this summer.

Users can create a widget by articulating their desires in natural language. For instance, one might request the feature to “recommend three high-protein meal prep recipes every week” to generate a custom dashboard that can be added and resized on the home screen.

Alternatively, if you’re a cyclist focused solely on wind speed and rain, you can create a weather widget that displays only those specific metrics on your home screen.

Gemini also has the ability to retrieve information from the internet and integrate with Google applications such as Gmail and Calendar to assemble a unified, personalized dashboard. For example, if you’re organizing a family reunion in Berlin, it can compile your flight and hotel information, highlight restaurant bookings, and even include a countdown.

This feature represents Google’s latest effort to integrate generative AI more deeply into the Android ecosystem, as technology firms compete to enhance customization tools for everyday users.

“It’s akin to asking your personal assistant a question and having them repeatedly deliver the answer,” stated Ben Greenwood, director, PM, Android Core Experiences, during a briefing with reporters. “Consider it as querying Gemini about the world, its understanding of current events, and aspects related to your personal data. These two categories unlock a wealth of use cases that we’re incredibly enthusiastic about.”

The company announced this new functionality alongside the introduction of Gemini Intelligence, which will offer more utilities like advanced autofill, a voice dictation feature powered by AI for Gboard, and other capabilities.

An anti-doomscrolling mechanism has now been integrated into Android. (Indeed, the situation has become that dire.)

On Tuesday, Google revealed Pause Point, a feature aimed at preventing users from becoming absorbed in addictive applications on Android, the mobile operating system that powers Google’s Pixel devices, Samsung smartphones, and more.

Pause Point operates by mandating a 10-second delay after opening any app identified as a distraction by the user. TikTok, Instagram, X, and even Google’s YouTube might be categorized as “distracting” by Android users concerned about the influence of time-consuming algorithms on their daily routines.

With this feature, Google isn’t solely focused on users’ well-being.

It is also responding to mounting regulatory scrutiny regarding social media risks and algorithmic threats. Presently, numerous countries and U.S. states have enacted legislation to restrict or prohibit minors from accessing social media, as the effects of these applications on the mental health of young individuals have become more apparent.

Google can now highlight a feature like Pause Point to assert its role in being part of the solution rather than the issue.

“Android is more advanced than ever, but we also want to provide you with the tools to disconnect when necessary,” said Dieter Bohn, former executive editor at The Verge and current director of product operations for Google’s Platforms & Ecosystems team, during a press briefing regarding the Android 17 update.

“I believe we are all guilty of picking up our phone and opening an app, only to get stuck in autopilot, with an hour disappearing,” he mentioned.

Until now, social media application developers, including YouTube, have turned to app timers to remind users to take breaks or stop scrolling. Pause Point reverses that concept, interrupting app launches — and the dopamine rush that ensues — to compel users to pause and reconsider if this is truly what they wish to do, or merely a habit they want to overcome.

Google suggests you could utilize the time allowed by Pause Point to engage in a brief breathing exercise or to contemplate alternative activities instead of scrolling. For example, the feature can recommend more productive applications, such as a preferred fitness app, an audiobook application, the Kindle or Google Play Books application, and others.

Alternatively, you can browse through some favorite photos for inspiration — perhaps ones that remind you of other enjoyable activities, such as walking outdoors, spending time with your pets, or engaging in crafts.

Additionally, Pause Point allows users to set an app timer before they begin, making the time spent in the app feel more deliberate from the outset. This could be more effective than a standard timer, which is fixed for the same duration, regardless of the situations prompting you to take a break from scrolling.

The feature is also more challenging to disable than typical app timers, many of which can be easily ignored. In contrast, Pause Point requires a phone restart to turn it off, according to Google, which also encourages reflection before deactivation.

While Pause Point may not be as entertaining (or charming) as screen-time-focused or self-care applications like Finch or Hank Green’s Focus Friend, it benefits from being integrated into Android itself, which could enhance its adoption.

Looking to leave billionaire-controlled social media but unsure of how to begin? A new application named Indigo is being launched today to enable users to navigate the open social web across various platforms, all within a single application.

Decentralized social media is gaining traction, as a growing number of individuals recognize their desire for control over algorithms, feeds, moderation preferences, and other elements of engaging in online communities. Nevertheless, numerous options exist. Users can select decentralized servers that form the social network Mastodon, utilizing the older ActivityPub protocol (also currently implemented by Meta’s Threads), or they can choose the newer social platform Bluesky, which operates on the AT Protocol.

Indigo does not compel users to make a choice. Rather, the app integrates the distinct social networks into a cohesive experience, featuring a synchronized timeline across devices and a composer that enables simultaneous cross-posting to both platforms.

Created by Soapbox Software — the same group responsible for the cross-posting application Croissant, which simplifies posting to Mastodon, Bluesky, and Threads from a single interface. Indigo co-creator Ben McCarthy, who also developed the Obscura apps and more, mentioned to TechCrunch that the new app leverages the team’s prior experience with Croissant while enhancing it with additional features that make it an excellent stand-alone solution for users on both open social networks.

When utilizing Indigo, it is possible to differentiate between Bluesky and Mastodon posts in your timeline: Bluesky profiles are highlighted in blue, while Mastodon profiles are marked in purple.

The application also features additional tools and functionalities, such as options that allow users to automatically scroll to the top when new posts arrive, conceal NSFW content, mute specific users or keywords, filter visible replies, activate a dark mode, and more.

Additionally, you can manage all your curated feeds and lists from both social networks through the app’s “more” menu, keep track of your notifications across both platforms, and conduct searches across Bluesky and Mastodon utilizing Indigo’s universal search feature.

The ability to cross-post, synchronize timelines across devices, and monitor trends between networks requires a subscription fee of either $4.99 monthly or $34.99 annually. (There’s a $119.99 one-time purchase available for those committed to the open social web.) Presently, Indigo is compatible with iPhone, iPad, and Mac.

Threads is experimenting with a Meta AI integration that functions similarly to X’s Grok. Users with a public profile will have the ability to mention Meta AI in a post or reply for additional context. This feature is currently in beta testing across Malaysia, Saudi Arabia, Mexico, Argentina, and Singapore.

Meta informed TechCrunch via email that the capability aims to assist individuals in gaining real-time insight into trends and breaking news, as well as provide recommendations, all within chats.

Users can now invoke Meta AI to inquire about topics such as, “Why is there buzz about the World Cup this month?,” “Which Met Gala outfits are gaining attention right now?,” or “What’s the Knicks’ status in the playoffs?” Meta AI will then process the request and reply as a public response posted by the @meta.ai account. It will reply in the language used in the original post.

By incorporating Meta AI into its platform, Threads is positioning itself not merely as a location for discussing news and trends but also as a source for information and recommendations without necessitating leaving the app.

The concept mirrors Grok’s function on X, which is populated with user posts posing questions to the AI chatbot, such as, “Is this authentic?” or “Clarify this.” Naturally, giving an AI chatbot this degree of exposure poses risks, as witnessed on X when Grok produced posts that praised Hitler. Nevertheless, Meta AI is equipped with notably stronger safeguards than Grok, though it is yet to be determined if it will encounter similar problems.

Meta emphasizes that if users wish to see fewer replies from Meta AI in their feed, they can mute @meta.ai, select the “Not interested” option on any Meta AI post, or hide a Meta AI response that appears directly on their post.

The company states it intends to gather insights from early feedback and will work on enhancing the experience before broadening it to a wider audience.

As malicious entities harness AI to take advantage of software flaws with unmatched speed, businesses are increasingly acknowledging the necessity to strengthen their cybersecurity measures.

Luckily, these same AI technologies are aiding companies in their defense efforts.

The rising demand for such features has allowed Exaforce, an AI startup focused on identifying and preventing attacks in real time, to secure a $125 million Series B round. This funding round valued the three-year-old startup at $725 million, with investments from HarbourVest, Peak XV, Mayfield, Khosla Ventures, and Seligman Ventures.

This substantial funding round comes just a year after Exaforce obtained a $75 million Series A, totaling its capital to $200 million. This influx highlights both the high expenses involved in establishing and marketing an AI-powered security operations center (SOC) and the significant market potential investors perceive in this sector.

Exaforce states that it employs AI agents, referred to as “Exabots,” utilizing deep data analytics to automate security workflows, thus alleviating the workload on human analysts.

For co-founder and CEO Ankur Singla, the objective is clear: Utilize AI to detect and thwart threats in real time. “It’s a very straightforward mandate, but it’s quite complex to implement,” he noted.

The primary challenge for security teams lies in the fact that most threat alerts turn out to be false alarms. “A security operations professional receives hundreds of alerts. How can one discern which is a genuine, high-priority alert?” Umesh Padval, a managing partner at Seligman Ventures, illustrated the task of security teams as akin to searching for a needle in a haystack.

Exaforce asserts that its AI platform can minimize manual, tedious tasks by up to 90%.

In light of the increasing cyberattacks, the startup recently launched “vibe hunting,” a feature allowing security teams to query its AI platform using natural language to explore potential threats based on simple intuition. “With vibe hunting, you can pose a straightforward question like, ‘Did we receive any new attacks from Iran?’” Singla explained.

Exaforce officially launched its product in the market in the last quarter of the previous year, after two years of testing with design collaborators. Since then, the startup has onboarded 20 customers, including prominent names like Replit and Guardant Health. Due to the surge in cyberattacks, Singla mentioned to TechCrunch that Exaforce anticipates acquiring 40 to 50 customers by the end of this year.

High-profile attacks have “accelerated our capacity to engage with customers, as they no longer question, ‘Why do I need this?’” Singla noted. The more common question he is encountering now is: “How do I put it into operation?”

Exaforce is not the only one employing AI for security operations. The firm contends with competition from startups like 7AI, Dropzone AI, and Prophet Security, as well as established industry players like Palo Alto Networks and CrowdStrike.

Waymo has released a software upgrade for its fleet of nearly 4,000 vehicles to assist them in evading flooded streets as a component of a recall declared by the National Highway Traffic Safety Administration (NHTSA) on Tuesday.

However, the company has not completely addressed the issue regarding how its vehicles respond in these situations. Documents provided by NHTSA indicate that Waymo is still “developing the final solution for this recall.”

The concern seems to be that Waymo’s robotaxis were decelerating but not halting when they encountered flooded roads that were impassable, according to the NHTSA. Robotaxis utilizing both Waymo’s fifth- and sixth-generation autonomous driving systems are involved.

The regulator noted that the recall pertains to 3,791 vehicles — giving us a clearer picture of how many vehicles Waymo operates on the streets across approximately a dozen U.S. cities.

Waymo has now initiated several recalls for its self-driving vehicles. The company’s first recall occurred in February 2024 when it found two robotaxis in Phoenix had independently crashed into the same towed vehicle. Since then, Waymo has conducted recalls to rectify low-speed collisions with parking barriers and telephone poles, as well as to tackle unlawful driving near school buses.

Waymo chose to initiate the recall in late April after its robotaxis had difficulty navigating floods in central Texas; during one incident, an empty robotaxi was carried away in San Antonio. The company has also halted operations in the city.

The initial update sent to its fleet imposes “restrictions at times and at locations where there is an increased likelihood of encountering a flooded, higher-speed roadway,” according to the NHTSA.

“We have pinpointed an area for enhancement regarding impassable flooded lanes on higher-speed roadways, and have made the choice to submit a voluntary software recall with NHTSA concerning this scenario,” Waymo stated in a release. “We are striving to implement additional software protections and have introduced mitigations, including refining our extreme weather protocols during heavy rain, limiting access to regions where flash flooding could occur.”