Latest AI News
Mozilla Wants the World to Have 7 Billion AGIs
CTO Raffi Krikorian believes Mozilla is best positioned to make open-source AI systems more accessible, like how it did with Firefox.
View
Meta Delays Launch of AI Model ‘Avocado’: Report
Meta has postponed the release of text-based LLM Avocado to May or June as performance lags behind rivals.
View
Alexa+ gets a new ‘adults only’ personality option that curses but won’t do NSFW content
Amazon’s AI assistant Alexa+ is getting another new personality. On Thursday, the company announced it’s expanding its lineup of personality styles for users to choose from to include a “Sassy” option, which is for adults only. Notes Amazon, before opting to use the Sassy personality, users will be required to go through additional security checks in the Alexa app. The personality style will also not be available when Amazon Kids is enabled, Amazon says. The new option joins others like Brief, Chill, and Sweet, launched last month. When you toggle on the option for Sassy in the Alexa mobile app, you’re warned that the Sassy style uses explicit language, which is why it requires a security check. On iOS, this involved a Face ID scan. The AI assistant explained its style to us like this: “The Sassy style is built on one premise: help first, judge always. Every answer comes wrapped in wit and a well-placed roast — it’ll answer your question; it’ll just make you feel something about it first. Expect reality checks delivered with charm, compliments that somehow sting, and warmth you didn’t see coming. Equal-opportunity irreverence, zero apologies. Honest, sharp, and funny — and somehow that’s more helpful than helpful.” Alexa’s app also had warned that the style could contain “mature subject matter.” However, further investigation discovered this is not Amazon’s version of something likeGrok’s adult AI companions. The AI assistant said the new option won’t get into areas like explicit sexual content, hate speech, illegal activities, personal attacks, or anything that could cause harm to oneself or others. The move is the latest example of how Amazon is trying to make Alexa+ more customizable, as it revamps the assistant for the generative AI era. By offering the assistant different personalities — including one positioned as more adult — Amazon is borrowing from a broader trend in AI, where companies have been experimenting with tone, style, and personas to make their assistants more engaging and personalized to the individual users’ choices.
View
Bumble introduces an AI dating assistant, ‘Bee’
Dating app maker Bumble is venturing into generative AI. During the company’s fourth-quarter earnings on Wednesday, Bumble introduced a new AI assistant it’s calling “Bee,” designed to become a personal matchmaker that learns users’ “values, relationship goals, communication style, lifestyle, and dating intentions” through private chats. It then uses those insights to help find the user more relevant matches. Currently, Bee is in the pilot phase and being tested internally, Bumble founder and CEO Whitney Wolfe Herd told investors, but it’s launching into beta soon. With Bee, the company envisions being able to capture much more information about Bumble users, as it learns more about each individual’s story and what they really want. This could differentiate Bumble’s app from others like Tinder, which also just underwent an overhaul as the dating app market has fizzled with Gen Z users. Bumble says users will interact with Bee much like they do with other AI chatbots, through typing and speaking in a more conversational style. Initially, Bee will be used to power a new dating experience called “Dates” that uses AI to recommend matches, but in the future, Bumble says Bee will move into other areas, like offering date suggestions or requesting anonymous feedback from your prior matches. In “Dates,” Bee will first learn about the user through a private, onboarding conversation. It then identifies two people who have shared intentions, values, and relationship goals. Both users are notified in the app with a description of why they make a great match. The addition is part of a broader tech and AI-focused overhaul of the dating app, which to date has marketed itself as more focused on women’s needs. The company pioneered features like “women message first,” body-shaming bans, and tools that blurred unsolicited explicit images, among others. Now it’s looking to use AI to return to user growth amid a dating market that sees younger users, particularly Gen Z, growing tired of the swipe. In fact, Herd said that Bumble would experiment with removing the long-popular swipe mechanism in select markets to see how users react. Instead of prioritizing swipes as a binary “yes” or “no,” Bumble is looking to leverage other features, like new “chapter-based” profiles where members can connect with one another on different parts of a user’s life story. This will give Bumble more data to feed into its AI system and algorithms. “We will be introducing more dynamic ways for somebody to express interest in your story, rather than just your profile, and this is going to drive more dynamic engagement, spark better conversation, and ultimately drive better KPIs across the board — like engagement and chances to get better conversations going,” Wolfe Herd said. “You will also see us take a much more deliberate approach to getting people offline versus just in what people refer to as dead-end chat zones.” The company is also looking into other ways to better cater to Gen Z, a cohort that often prefers group socializing over one-on-one dates to get to know people. The company has been working to add AI to its app for years, rolling out changes likeAI photo selectionandfeedbacktools, for instance, as well as in areas likesafety. Wolfe Herd told investors that Bumble’s back-end infrastructure had been overhauled as the app infused itself with AI. The company reported better-than-expectedearningsin Q4, withrevenueof $224.2 million and average revenue per paying user up 7.9% to $22.20. The stock rallied some 40% on the news.
View
Sales automation startup Rox AI hits $1.2B valuation, sources say
Rox, a startup developing autonomous AI agents to boost sales productivity, has raised a round valuing the company at $1.2 billion, according to multiple sources. The funding included a lead investment from returning backer General Catalyst, two of the people said. Rox AI and General Catalyst did not respond to TechCrunch’s request for comment. At the time of the fundraise, which closed last year, RoxAI was projected to close 2025 with $8 million in annual recurring revenue (ARR), according to two people familiar with the deal. In November 2024, Rox announced it had raised a total of$50 million, including a seed round led by Sequoia and a Series A round led by General Catalyst, with participation from GV. Rox was founded in 2024 by former chief growth officer of New Relic, Ishan Mukherjee. Mukherjee joined New Relic following its acquisition of Pixie, an observability startup he co-founded. The startup positions itself as an intelligent revenue operating system that plugs into a company’s current software setup—from Salesforce to Zendesk, and then deploys hundreds of AI agents. These agents monitor existing accounts, research prospects, and update the CRM. By consolidating these functions, Rox aims to replace and consolidate numerous fragmented software solutions currently used by sales teams. “Rox’s unique system of AI agents levels up the CRM experience. These agents work constantly behind the scenes to monitor customer activity, identify potential risks and opportunities, and even suggest the best course of action,” GV investorDave Munichiello wrote in a 2024 blog postwhen announcing the Series A round. Rox’s competition spans several categories, including established revenue intelligence providers like Gong and Clari, as well as AI sales development platforms such as 11x and Artisan. Plus, there seems to be a steady stream of new AI-native, all-in-one CRM competitors joining the field, such asSam Blond’s new startup, Monaco, launched out of stealth last month. According to Rox’s website, the company’s customers include Ramp, MongoDB, and New Relic.
View
How to watch Jensen Huang’s Nvidia GTC 2026 keynote
Nvidia kicks off its annual GTC developer conference in San Jose, California, next week with CEO Jensen Huang’s keynote scheduled for Monday at 11am PT / 2pm ET. GTC — which stands for GPU Technology Conference — is Nvidia’s flagship annual event, where the chipmaker typically uses the spotlight to announce new products, champion partnerships, and lay out its vision for the future of computing. Huang’s keynote will focus on Nvidia’s role in the future of computing and AI. You can watch the two-hour address in person at theSAP Centerorlivestream the talkon the event’s website. The broader three-day event is focused on what’s coming next for AI across industries including healthcare, robotics, and autonomous vehicles, among others. On the software side, it’s rumored that Nvidia will release an open source platform for enterprise AI agents, dubbedNemoClaw, as originally reported by Wired. The platform would give businesses a structured way to build and deploy AI agents (software that can carry out multi-step tasks autonomously) and would position Nvidia to mirrorsimilar offeringsfrom companies like OpenAI. On the hardware side, the company is also rumored to be releasing anew chip designed to accelerate the AI inference process— the process by which an AI model applies what it has learned to generate responses or make decisions, as distinct from the initial training process, which requires far more computing power. Faster, cheaper inference is widely seen as one of the last bottlenecks to scaling AI applications broadly. The chip, if confirmed, would represent Nvidia’s latest bid to dominate not just the training market, where it already commands an estimated 80% share, but the inference market as well, where competition from custom chips built by Google, Amazon and others is fast intensifying. Kevin Cook, a senior equity strategist at Zacks Investment Research, told TechCrunch that attendees should also expect to learn what the company plans to do with its relationship with Groq, the inference company Nvidia reportedly paid $20 billion late last year to license its technology. There’s a lot of curiosity around this tie-up, given that Jonathan Ross, Groq’s founder, Sunny Madra, Groq’s President, and other members of the Groq team agreed to join Nvidia to help advance and scale that licensed tech. There will, of course, also be a range of partnership announcements and demonstrations showcasing Nvidia’s AI capabilities across industries.
View
A writer is suing Grammarly for turning her and other authors into ‘AI editors’ without consent
Grammarly released acontroversial featurelast week that uses AI to simulate editorial feedback, making it seem like you’re getting a critique from novelist Stephen King, the late scientist Carl Sagan, or tech journalist Kara Swisher. But Grammarly did not get permission from the hundreds of experts it included in this feature, called “Expert Review,” to use their names. One of the affected writers, journalist Julia Angwin, has filed aclass action lawsuitagainst Superhuman, the parent company that owns Grammarly, arguing that the company violated the privacy and publicity rights of her and the other writers it impersonated. A class action lawsuit allows writers to join Angwin in her case. “I have worked for decades honing my skills as a writer and editor, and I am distressed to discover that a tech company is selling an imposter version of my hard-earned expertise,” Angwin said in astatement. The situation is more than a little ironic — Angwin has spent her career leading investigations into tech companies’ impacts on privacy. Other critics of this kind of technology, like renowned AI ethicistTimnit Gebru, were also included in Grammarly’s “Expert Review.” The “Expert Review” feature, available only to subscribers paying $144 a year, predictably fails to deliver on the promise of thoughtful feedback. Casey Newton, the founder and editor of the tech newsletter Platformer and another person impersonated by Grammarly,fed one of his articlesinto the tool and got feedback from Grammarly’s approximation of tech journalist Kara Swisher. Grammarly’s imitation of Swisher produced “feedback” so generic that it raises the question of why the company would go through the rigmarole of using these writers’ likenesses in the first place. Here is what Grammarly’s approximation of Kara Swisher told him: “Could you briefly compare how daily AI users versus AI skeptics articulate risk, creating a through-line readers can follow?” Newton relayed the message from the AI approximation of Kara Swisher to the actual, real human being, Kara Swisher. “You rapacious information and identity thieves better get ready for me to go full McConaughey on you,” Swisher texted Newton (referring to Grammarly). “Also, you suck.” Grammarly has since disabled the “Expert Review” feature, according to aLinkedIn postby Superhuman CEO Shishir Mehrotra. While Mehrotra offered an apology, he continued to defend the idea of the feature. “Imagine your professor sharpening your essay, your sales leader reshaping a customer pitch, a thoughtful critic challenging your arguments, or a leading expert elevating your proposal,” he wrote. “For experts, this is a chance to build that same ubiquitous bond with users, much like Grammarly has.”
View
Bumble to launch an AI dating assistant, ‘Bee’
Dating app maker Bumble is venturing into generative AI. During the company’s fourth-quarter earnings on Wednesday, Bumble introduced a new AI assistant it’s calling “Bee,” designed to become a personal matchmaker that learns users’ “values, relationship goals, communication style, lifestyle, and dating intentions” through private chats. It then uses those insights to help find the user more relevant matches. Currently, Bee is in the pilot phase and being tested internally, Bumble founder and CEO Whitney Wolfe Herd told investors, but it’s launching into beta soon. With Bee, the company envisions being able to capture much more information about Bumble users, as it learns more about each individual’s story and what they really want. This could differentiate Bumble’s app from others like Tinder, which also just underwent an overhaul as the dating app market has fizzled with Gen Z users. Bumble says users will interact with Bee much like they do with other AI chatbots, through typing and speaking in a more conversational style. Initially, Bee will be used to power a new dating experience called “Dates” that uses AI to recommend matches, but in the future, Bumble says Bee will move into other areas, like offering date suggestions or requesting anonymous feedback from your prior matches. In “Dates,” Bee will first learn about the user through a private, onboarding conversation. It then identifies two people who have shared intentions, values, and relationship goals. Both users are notified in the app with a description of why they make a great match. The addition is part of a broader tech and AI-focused overhaul of the dating app, which to date has marketed itself as more focused on women’s needs. The company pioneered features like “women message first,” body-shaming bans, and tools that blurred unsolicited explicit images, among others. Now it’s looking to use AI to return to user growth amid a dating market that sees younger users, particularly Gen Z, growing tired of the swipe. In fact, Herd said that Bumble would experiment with removing the long-popular swipe mechanism in select markets to see how users react. Instead of prioritizing swipes as a binary “yes” or “no,” Bumble is looking to leverage other features, like new “chapter-based” profiles where members can connect with one another on different parts of a user’s life story. This will give Bumble more data to feed into its AI system and algorithms. “We will be introducing more dynamic ways for somebody to express interest in your story, rather than just your profile, and this is going to drive more dynamic engagement, spark better conversation, and ultimately drive better KPIs across the board — like engagement and chances to get better conversations going,” Wolfe Herd said. “You will also see us take a much more deliberate approach to getting people offline versus just in what people refer to as dead-end chat zones.” The company is also looking into other ways to better cater to Gen Z, a cohort that often prefers group socializing over one-on-one dates to get to know people. The company has been working to add AI to its app for years, rolling out changes likeAI photo selectionandfeedbacktools, for instance, as well as in areas likesafety. Wolfe Herd told investors that Bumble’s back-end infrastructure had been overhauled as the app infused itself with AI. The company reported better-than-expectedearningsin Q4, withrevenueof $224.2 million and average revenue per paying user up 7.9% to $22.20. The stock rallied some 40% on the news.
View
Atlassian follows Block’s footsteps and cuts staff in the name of AI
Australian productivity software company Atlassian held layoffs as the company looks to funnel more money into AI. Atlassian announced it’scutting 10% of its workforce, around 1,600 people, on March 11. The company said this decision allows it to spend more funds on AI and enterprise sales and to strengthen its finances. More specifically, Atlassian said that it’s doing well but is choosing to adapt to market conditions. “The bar for what ‘great’ looks like for software companies — on growth, on profitability, on speed, on value creation — has gone up,” Atlassian CEO Mike Cannon-Brookes wrote in a press release related to the layoffs. TechCrunch reached out to Atlassian for more information regarding which types of roles were cut and what happens next. Atlassian declined to comment beyond the release. This news comes just a few weeks after a similar, albeit more drastic, statement was made by Block CEO Jack Dorsey. In February, the payments company announced it wascutting more than 4,000 employees, nearly half of its 10,000 employees at the time. Dorsey said the cuts were being driven by the fact that AI could automate a lot of the work these employees were doing and predicted that many other companies would come to the same conclusion. Several enterprise-focused VCs predicted to TechCrunch that2026 would be the year that AI would start to take a meaningful toll on labor. So far, their prediction has come true.
View
Tinder tries to lure people back to online dating with IRL events, virtual speed dating
Tinder held its inaugural product keynote on Thursday, unveiling an ambitious slate of updates designed to reinvigorate its product, bolster safety, and harness AI. This comes on the heels of Match Group’s $50 million investment in product development,announced last August, as the parent company looks to re-engage its user base and win over younger Gen Z daters. Among the updates are innovative features for discovering in-person events and meeting people in real life, alongside a new virtual speed dating experience being tested in Los Angeles. Additionally, a series of AI enhancements were introduced to improve the matching algorithm and enhance user safety. One of the most notable features is the new Events tab, which will be in beta for users in Los Angeles starting in late May or early June. This feature lets users discover curated local events— such as speakeasies, bowling, raves, and pottery classes — where they can connect with matches in person. This is a nod to Gen Z’s growing appetite for real-world encounters over endless swiping. There has been a shift away from traditional dating apps, as young people seek authentic offline experiences orunconventionalwaysto meet potential partners. Other apps, likeBreeze,222, Timeleft,andThursday, have also tapped into this in-real-life (IRL) trend. “We’re really trying to tap into meeting younger users at the places where they’re already hanging out,” Hillary Paine, senior vice president of product at Tinder, told TechCrunch. “You can go to an event with your friend and have a good time, or you could meet somebody new. Instead of asking users to choose between their dating life and their social life, we’re trying to blend these things together and create a more social community first experience.” Profiles of event attendees will be available on the app after the event for users to like and swipe through, a concept reminiscent of“Missed Connections” ads, allowing users who may have lacked the courage to approach someone or simply missed the opportunity to reconnect. Speed datingis also experiencing a resurgence. To hop on the bandwagon, Tinder revealed it’s now piloting a video speed dating experience in LA, where users can join scheduled three-minute video chats with potential matches. This serves as a “vibe check,” designed to help people break the ice and gauge chemistry before committing to an in-person meeting. The company noted that users will have the option to extend promising conversations beyond three minutes. Your profile photo has to be verified to join the experience. Many daters appear to have become fatigued by video chats, making it intriguing to see if this experiment, introducedlate in the game, will achieve success. During the COVID-19 pandemic, Tinder launched aFace-to-Facefeature that was later discontinued, indicating a decline in interest. AI was also a major theme at the keynote. Tinder is continuing to invest in the technology, starting with its “Chemistry” feature, which utilizes AI to learn about users through questions and, with their permission, their camera rolls. This feature curates daily matches to help reduce swipe fatigue and is now rolling out in the U.S. and Canada afterinitial testingin Australia and New Zealand. In the future, aspects of Chemistry will grow from being just one feature into something that shapes the entire Tinder experience, making it more personalized, the company said. Additionally, Tinder introduced a new “Learning Mode” that presents more relevant matches earlier on. The system is designed to quickly gain insights into what users are seeking in potential matches, adapting recommendations to better suit personal preferences. Previously, Tinder needed multiple swiping sessions to gather enough signals to personalize well. With Learning Mode, Paine notes, it can start to understand a user from the very first session. She said, “We’re hoping that this is something that makes Tinder really feel like it understands you from the very first time you use it, or if you’re returning to Tinder after some time away, it feels like it gets me, and I don’t have to spend a lot of time telling Tinder what I’m looking for again.” Tinder is also enhancing safety features like “Does This Bother You?,” which now uses large language models to better detect harmful messages and auto-blurs disrespectful content, while “Are You Sure?” prompts are being fine-tuned to more accurately identify potentially harmful interactions. Visually, Tinder is getting a sleek redesign: edge-to-edge profile photos, a subtle blur effect, and a Liquid Glass aesthetic for the Like and Nope bar. New modes are also on the horizon: “Music Mode” will allow up to 20 Spotify songs to auto-populate a user’s profile, and “Astrology Mode” will let users add birth details to unlock their Sun, Moon, and Rising signs and check compatibility. This follows the recentlaunchof Double Date Mode and College Mode. Overall, the slew of announcements signals a pivotal shift in Tinder’s approach. While Match reported a positive earnings result inQ4 2025, with $878 million in revenue, the company has faced consecutive quarters of declining paying subscribers. Consequently, it’s under pressure to retain users and restore investor confidence, even as its outlook remains cautious, acknowledging changing user preferences and heightened competition. It remains to be seen whether these changes will help maintain daters’ interest in the app. However, one thing is clear: Tinder is making a significant commitment to the future of dating, shifting away from solely relying on swiping and adapting to what it believes its young users want.
View
Facebook Marketplace now lets Meta AI respond to buyers’ messages
Facebook Marketplace is rolling out new Meta AI features, including one that automatically replies to messages from interested buyers, FacebookannouncedThursday. The social network is also using AI to help sellers list items faster and summarize profiles and is giving sellers the option to offer shipping on their listings. The company says that since sellers receive a lot of buyer inquiries, it’s making it easier to respond to initial messages with auto-replies powered by Meta AI. Now when buyers inquire about an item’s availability, sellers can use Meta AI to automatically draft replies using information from their listing, such as the description, availability, pickup location, and price. Sellers can enable, preview, and edit these auto-replies during listing creation. This new feature is designed to address a common pain point for sellers, who often don’t want to spend excessive amounts of time responding to initial or redundant messages, such as asking if an item is available when it’s already marked as available. One seller had evencreated their own AI toollast year to handle these initial inquiries from interested buyers. Facebook Marketplace also now lets sellers list their items automatically using Meta AI. After uploading an image of an item, Meta AI can create a draft listing, fill in the details, and suggest a price based on similar items in the area. Additionally, buyers will now see a summary of a seller’s Facebook profile at the top of their Marketplace page, including how long they’ve been on Facebook and their number of friends. The overview also includes a summary of their Marketplace activity, such as listing history, the types of items they sell, and their seller ratings. Sellers can now also expand their reach by offering shipping on their listings, Facebook says. And they can generate prepaid shipping labels and easily track every order from a simple dashboard. The new AI features join Facebook Marketplace’sexisting Meta AI integrations, including a tool that helps buyers ask the right questions when making a purchase, as well as AI-powered insights for vehicle listings.
View
Nvidia Unveils Nemotron 3 Super Open-Source AI Model for Agentic AI Systems
Nvidia released a new open-source artificial intelligence (AI) designed to handle complex agentic workflows. Dubbed Nemotron 3 Super, it is a hybrid mixture-of-experts (MoE) model that combines advanced reasoning capabilities and is said to complete tasks with high accuracy for autonomous agents. The new model is already being deployed by several AI firms, including Perplexity, for its new agentic Computer platform. Additionally, it is also being hosted on public repositories to let interested individuals download and run the model locally.
View
