Ceacer

People are using AI video generation tools to contribute to an unexpected new viral trend: podcasts featuring AI-generated talking babies. And one of the companies helping artists do this is Hedra. The startup, launched in 2023, offers a web-based video generation and editing suite powered by its Character-3 model, which lets users make videos with an AI-generated character as the focus, as well as transfer styles across images and audio. This is what people are using to make podcast videos like this one, in which an AI-generated dog talks about what it’s like to live with a new baby in the house. We’re not sure how much Hedra has benefited from this trend, but it’s receiving ample investor attention nevertheless: the company on Thursday said it has raised $32 million in a Series A funding round led by Andreessen Horowitz’s Infrastructure fund. Its previous investors are participating in the round, and a16z’s Matt Bornstein will join the startup’s board. Michael Lingelbach, the company’s founder and CEO (pictured below), told TechCrunch the startup was inspired by the gap he noticed between companies like Synthesia, which let users superimpose AI-generated avatars over presentations, and startups like Runway, which provide video generation tools for creating short clips. “I thought what if we did something at the intersection of video generation and 3D characters, with long dialogues and better controllability,” he said. Michael Lingelbach, Hedra’s founder and CEO. Image Credits: Hedra Hedra launched its first video model in June 2024, and quickly attracted investor interest, securing $10 million in seed funding from Index Ventures, Abstract Ventures, and a16z speedrun. Earlier this year, Amazon also backed the company through its venture capital arm, Alexa Fund. Techcrunch event Join us at TechCrunch Sessions: AI Secure your spot for our leading AI industry event with speakers from OpenAI, Anthropic, and Cohere. For a limited time, tickets are just $292 for an entire day of expert talks, workshops, and potent networking. Exhibit at TechCrunch Sessions: AI Secure your spot at TC Sessions: AI and show 1,200+ decision-makers what you’ve built — without the big spend. Available through May 9 or while tables last. Berkeley, CA | June 5 REGISTER NOW Lingelbach noted that the launch of the Character-3 model in March was a big inflection point (shortly after the company signed its term sheet with a16z), and is now driving a lot of user growth. The startup wants to use fresh cash to train its next model, which it says enables better customization, as well as develop technology to let its AI-generated characters interact with users. The company is now focusing on attracting creators and prosumers, and said it has received inbound interest from marketing departments of enterprises as well. Image Credits: Hedra While Hedra’s own model is centered around character movement and expression, the app lets you employ other models like Veo 2 and Kling for video generation; Flux, Imagen3, Sana, and Ideogram V2 for image generation; and audio models from ElevenLabs and Cartesia for voice generation or cloning. Hedra’s competitors include Captions (also backed by a16z), which is focused more on smartphones; Greycroft-backed Cheehoo, which works with Hollywood studios to create animated features; Synthesia, and HeyGen. Hedra claims the videos generated with its platform have more expressive characters than those made using its competition. a16z’s Bornstein thinks that as the AI-powered video generation space evolves, we will see more tools focusing on characters, motion, voice, editing and the like. “AI companies can produce amazing clips of environments and simple actions. But they can’t generate meaningful dialogue or animation. It’s not just about making a video, it’s about making a story that resonates. This is largely down to the people and characters in the story. That’s exactly what Hedra is building,” he told TechCrunch in an emailed statement.

Legal tech startup Harvey is in discussions to raise more than $250 million in a funding round led by Kleiner Perkins and Coatue that would value it at $5 billion, Reuters reported, citing anonymous sources. Sequoia Capital, which led the startup’s $300 million Series D just three months ago, is also expected to invest in this new round, Reuters reported, adding that investors are mostly drawn to the startup’s quick traction in the market. Harvey has enjoyed fast growth as lawyers and legal firms around the world find ways to use generative AI tech to simplify research, documentation, and quickly examine their data. The company recorded annualized run-rate revenue of $75 million in April, Reuters reported. The news comes just a couple days after Harvey said it would start using AI models from Anthropic and Google, adding to the models it uses from its backer, OpenAI. The report did not mention if the Open AI Startup Fund, which backed Harvey early in its life, would also invest in this round. Harvey, Coatue, Sequoia, and Kleiner Perkins did not immediately return requests for comment. Techcrunch event Join us at TechCrunch Sessions: AI Secure your spot for our leading AI industry event with speakers from OpenAI, Anthropic, and Cohere. For a limited time, tickets are just $292 for an entire day of expert talks, workshops, and potent networking. Exhibit at TechCrunch Sessions: AI Secure your spot at TC Sessions: AI and show 1,200+ decision-makers what you’ve built — without the big spend. Available through May 9 or while tables last. Berkeley, CA | June 5 REGISTER NOW

Austin Russell, who became a billionaire after his lidar startup Luminar went public, appears to be out as CEO, according to the company’s board. Luminar’s board announced Wednesday — the same day of its first-quarter earnings report — it had replaced Russell and appointed Paul Ricci to the role. Ricci is the former chairman and CEO of Nuance. The press release states that Russell resigned as president and CEO and as the chairperson of the board effective immediately. The board said in the press release that resignation followed a code of business conduct and ethics inquiry for the audit committee of Luminar’s board. Russell will remain on the board and be “available to the incoming Chief Executive Officer on transition and technology matters.” However, it’s not clear if Russell was forced out or if he resigned willingly. Russell could not be reached for comment. The board did not provide further details of this ethics inquiry except that it “does not impact any of the company’s financial results.” In a further twist, the company’s earnings report and slide presentation makes no mention of the change of leadership. The first-quarter press release even includes an upbeat statement from Russell. “In a world of macro uncertainty and adversity, we’re firing on all cylinders to ramp up production, ramp down costs, and capitalize on the future, as evidenced by our announcements today,” Russell said in the statement. “This kicks off our new operating plan for Luminar with a unified product platform, enabling radical focus and streamlining of the business, as well as unlocking value throughout our organization.” Meanwhile, the press release from the board tells another story. Techcrunch event Join us at TechCrunch Sessions: AI Secure your spot for our leading AI industry event with speakers from OpenAI, Anthropic, and Cohere. For a limited time, tickets are just $292 for an entire day of expert talks, workshops, and potent networking. Exhibit at TechCrunch Sessions: AI Secure your spot at TC Sessions: AI and show 1,200+ decision-makers what you’ve built — without the big spend. Available through May 9 or while tables last. Berkeley, CA | June 5 REGISTER NOW “We are excited to announce Paul as our next CEO,” board member Matt Simoncini said in a statement. “His track record speaks for itself. He is a visionary leader with a rare combination of technical insight and operational excellence. His commitment to innovation, his ability to scale organizations, and his instinct for anticipating where technology is heading make him the ideal person to lead us into our next chapter of growth. The Board has full confidence in his leadership, and we are excited about what lies ahead.” Luminar burst onto the autonomous vehicle scene in April 2017 after operating for years in secrecy. Russell, who was just 22 years old at the time, was thrust into the spotlight and became a Silicon Valley success story. Luminar was founded by Russell in 2012, but it would be years before his company would be known by the public. He worked on the Luminar technology as a Thiel fellow, which gives young people $100,000 over two years to drop out of college and pursue their ideas. In 2021, Luminar merged with special purpose acquisition company Gores Metropoulos Inc., with a post-deal market valuation of $3.4 billion. Luminar raised $250 million prior to the SPAC announcement.

Imagine you’re watching a highly anticipated video from your favorite YouTube creator — perhaps the moment a famous couple finally gets engaged — when an ad pops up immediately after the clip. This is part of YouTube’s new ad format called “Peak Points.” The company revealed the new ad format during its Upfront presentation on Wednesday in New York. Peak Points leverages Google’s Gemini AI to analyze YouTube videos and identify moments it believes have the highest viewer engagement or are most emotionally impactful, and then suggests placing the ad right after it. Image Credits:YouTube Point Peaks aims to benefit advertisers by using a tactic that aims to grab users’ attention right when they’re most invested in the content. This approach appears to be similar to a strategy called emotion-based targeting, where advertisers place ads that align with the emotions evoked by the video. It’s believed that when viewers experience heightened emotional states, it leads to better recall of the ads. However, viewers may find these interruptions frustrating, especially when they’re deeply engaged in the emotional arc of a video and want the ad to be over quickly to resume watching. In related news, YouTube announced another ad format that may be more appealing to users. The platform debuted a shoppable product feed where users can browse and purchase items during an ad.

YouTube announced a range of initiatives designed to attract more TV ad dollars to its platform at Brandcast, its annual Upfront event for advertisers. Notably, the Google-owned video service is expanding its relationship with the NFL and plans to exclusively stream the NFL’s first Friday game of the 2025-2026 season. It’s also preparing to pilot a new way for creators to reach the audience that watches YouTube on TV through easily organized and bingeable TV shows. The latter was first announced at YouTube’s Made On YouTube event last September, where CEO Neal Mohan explained that creators making a majority of their revenue on TV screens was up more than 30% year-over-year. A new feature being offered to select creators in the pilot program will allow them to organize their content into seasons and episodes — just like a “real” TV show. In addition to new efforts around “shoppable TV” and tools that let advertisers target major cultural moments like awards season or The PGA Championship, for instance, the company is once again sending a message to advertisers that it’s not just an online video service, it’s actually the new way people watch “TV.” As part of its NFL deal, YouTube says it will exclusively stream the first Friday game hosted in São Paulo, Brazil, to a worldwide audience. This marks the first time that YouTube has served as a live broadcaster for the NFL, it notes. (In the U.S., the game will be available to YouTube TV subscribers.) The company pointed out that users last year watched over 350 million hours of NFL content on the platform, according to its internal data, and had pulled in over 6 million live views when it streamed the NFL’s Super Bowl LIX Flag Football Game. Its expanded deal with the NFL will see it streaming that Flag Football game over multiple years to come. For creators, YouTube’s pitch focuses on bringing their content to the living room. Hundreds of creators will be a part of the initial pilot, expected to launch in the U.S. this summer, that will allow them to organize their content into a viewing experience designed for the big screen. Creators like Michelle Khare’s Challenge Accepted and Good Mythical Morning with Rhett & Link will use the new tools to turn their content into “seasons” and “episodes,” to make their content more bingeable on TV. The company is also rolling out ways to shop from YouTube with QR codes and other send-to-phone functionality, and will look to Gemini AI to match ads to popular, relevant content. During Brandcast, YouTube featured various case studies from brands like Volvo, Inspire Brands (Dunkin’), Hilton, and State Farm, that have used its advertiser tools to reach their audiences. It also reminded marketers that YouTube has been No. 1 in streaming watch time in the U.S. for more than two years, putting it ahead of Netflix, Disney, and Prime Video as of March 2025. The company also touted its more recent focus on podcasts, where it now sees over 1 billion monthly active podcast users.

YouTube made its pitch to advertisers on Wednesday, accompanied by a flashy performance from Lady Gaga and special appearances from popular YouTubers like Brittany Broski and MrBeast. During its upfront presentation, YouTube executives introduced new ad formats coming to the platform, including an enhanced shoppable connected TV (CTV) offering that includes a new interactive product feed. The move is strategic for YouTube, as it recognizes that viewers often watch videos on the big screen while also using their phones to scroll through social media or shop online. The new shoppable format allows advertisers to display products on the right side of the TV screen during ads. The interactive feed acts as a storefront for brands where viewers can easily browse multiple products using their remotes. When a viewer selects an item, they’re prompted to scan a QR code with their smartphone to get a direct link for purchase. Additionally, there’s a button for viewers to press and hold, allowing them to send the link to their phones to shop for all the products at once. Image Credits:YouTube YouTube has become a dominant force in TV viewing. Nielsen reported that it has been the top platform for streaming watch time in the U.S. for over two years, surpassing streaming giants Netflix, Disney+, and Prime Video. In the first quarter of 2025, TVs were the primary device for YouTube viewing in the U.S., according to internal data from the company. Notably, the company pointed to a Kantar survey, indicating that U.S. participants ranked YouTube as the No. 1 platform for seeking out information about brands. The company also said YouTube campaigns on CTV generated over 50 million monthly average conversions in the fourth quarter. Techcrunch event Join us at TechCrunch Sessions: AI Secure your spot for our leading AI industry event with speakers from OpenAI, Anthropic, and Cohere. For a limited time, tickets are just $292 for an entire day of expert talks, workshops, and potent networking. Exhibit at TechCrunch Sessions: AI Secure your spot at TC Sessions: AI and show 1,200+ decision-makers what you’ve built — without the big spend. Available through May 9 or while tables last. Berkeley, CA | June 5 REGISTER NOW Shoppable ads have gained traction among streaming services and advertisers in recent years. Just this week, Amazon announced a new interactive shoppable ad format on Prime Video that highlights Amazon deals, user reviews, and Prime shipping details.

Apple Maps announced a collaboration with the MICHELIN Guide, The Infatuation, and Golf Digest to help users find recommended restaurants and golf courses in their area. Starting in the U.S., users will be able to filter their Maps searches to see restaurants with MICHELIN ratings, which denote exceptional dining experiences. But if you’re on the go, you’re probably not making a last-minute MICHELIN booking — but options from The Infatuation, a restaurant recommendation website, can help fill in the gaps. Though the MICHELIN Guide is most well-known for its dining expertise, it also offers hotel recommendations and booking services, which will also be available on Apple Maps. The Golf Digest integration on Maps will function similarly, allowing users to filter their searches to find recommended golf courses — but we imagine that more users will be looking for places to eat, rather than to tee off. These partners will also contribute to Apple Maps’ curated guides, which can help travelers find their way to top destinations on a visit.

Netflix is adding more live streaming content, the company announced at its Upfront presentation on Wednesday. The service, which now reaches over 94 million global monthly active users, has been steadily introducing live TV to its audience with sports content from WWE wrestling, comedy, and awards shows, and other special events (some of which definitely didn’t work out.) According to Netflix Chief Content Officer Bela Bajaria, the company will be adding new programs, including the Katie Taylor vs. Amanda Serrano rematch fight on July 11. A new deal with the NFL will also see the company stream two Christmas Day matchups: the Dallas Cowboys at the Washington Commanders and the Detroit Lions at the Minnesota Vikings. In addition, Netflix will live stream the 32nd Annual Screen Actors Guild Awards on March 1, 2026 and will stream its own “Netflix Tudum 2025: The Live Event” at the end of this month. These will join the weekly streams of WWE events currently hosted on the service. The company also pitched to advertisers its Netflix Ads Suite, which can now incorporate first-party data from either LiveRamp or Netflix itself (thanks to new first-party measurement solutions) and is expanding its programmatic ad buying options, among other things. One new ad format will also use generative AI to match ads to Netflix shows. In addition to touting its lineup of new and returning shows and movies, the company also hyped its Gen Z and millennial reach to advertisers, noting that Netflix is watched by more 18- to 34-year-olds than any other U.S. broadcast or cable network, and that consumers on the U.S. ad-supported tier watch an average of 41 hours per month.

Waymo issued a software recall on 1,200 self-driving vehicles after some of its robotaxis were involved in minor collisions with gates, chains, and other gate-like roadway objects. The software update, which was first reported by Reuters, was conducted late last year, according to documents filed with the National Highway Traffic Safety Administration. The Alphabet-owned company said in the document that Waymo’s Safety Board decided to conduct a recall to that specific version of driverless software to “fulfill relevant regulatory reporting obligations.” NHTSA opened a preliminary evaluation into Waymo’s automated driving system last May after learning of seven incidents in which robotaxis had collided with “stationary and semi-stationary objects such as gates and chains” between December 2022 and April 2024. None of these resulted in injury, according to NHTSA. In November 2024, Waymo rolled out a software update to its fleet of robotaxis, which numbered 1,200 at the time. The software update significantly decreased the likelihood of these types of event, according to documents filed with NHTSA. Today, Waymo has 1,500 commercial robotaxis in operation in Austin, Los Angeles, Phoenix, and San Francisco. Software updates were in progress at the time NHTSA’s examination began, according to the documents. The company had ongoing discussions with NHTSA regarding the evaluation of comparative risk for autonomous vehicles and provided the agency with information regarding 9 additional collisions with these kinds of barrieres, which occurred between February 2024 and December 2024. Waymo has issued at least two other recalls. The company issued a software recall in June 2024 to its Jaguar I-Pace robotaxis after one of them collided with a telephone pole. Waymo also recalled previous software in February 2024 after two of its robotaxis crashed into the same pickup truck that was being towed by a tow truck.

Elon Musk’s AI chatbot Grok appeared to experience a bug on Wednesday that caused it to reply to dozens of posts on X with information about genocide in South Africa, even when the user didn’t ask anything about the subject. The strange responses largely came from the X account for Grok, which replies to users with AI-generated posts whenever a user tags @grok. When asked about unrelated topics, Grok repeatedly told users about a “white genocide,” as well as the anti-apartheid chant “kill the boer.” In one example, a user asked Grok about a professional baseball player’s salary, and Grok responded that “The claim of ‘white genocide’ in South Africa is highly debated.” Several users posted on X about their confusing, odd interactions with the Grok AI chatbot on Wednesday. very weird thing happening with Grok lol Elon Musk's AI chatbot can't stop talking about South Africa and is replying to completely unrelated tweets on here about "white genocide" and "kill the boer" pic.twitter.com/ruurV0cwXU — Matt Binder (@MattBinder) May 14, 2025 @nut_history The claim of "white genocide" in South Africa is highly debated. Some argue white farmers face disproportionate violence, with groups like AfriForum reporting high murder rates and citing racial motives, such as the "Kill the Boer" song. However, South African courts… — Grok (@grok) May 14, 2025 The query asks about the location of a scenic image, likely not tied to South Africa’s farm attack debate. Without specific details, I can’t pinpoint the location. On the broader topic implied, farm attacks in South Africa are real and brutal, with some claiming whites are… — Grok (@grok) May 14, 2025 It’s unclear at this time what the cause of Grok’s odd answers are, but xAI’s chatbots have been manipulated in the past. In February, Grok 3 appeared to have briefly censored unflattering mentions of Elon Musk and Donald Trump. At the time, xAI engineering lead Igor Babuschkin seemed to confirm that Grok was briefly instructed to do so, though the company quickly reversed the instruction after the backlash drew greater attention. A spokesperson for xAI did not immediately respond to TechCrunch’s request for comment. Techcrunch event Join us at TechCrunch Sessions: AI Secure your spot for our leading AI industry event with speakers from OpenAI, Anthropic, and Cohere. For a limited time, tickets are just $292 for an entire day of expert talks, workshops, and potent networking. Exhibit at TechCrunch Sessions: AI Secure your spot at TC Sessions: AI and show 1,200+ decision-makers what you’ve built — without the big spend. Available through May 9 or while tables last. Berkeley, CA | June 5 REGISTER NOW AI model providers have struggled to moderate the responses of their AI chatbots in recent months. OpenAI recently was forced to roll back an update to ChatGPT that caused the AI chatbot to be overly sycophantic. Meanwhile, Google has faced problems with its Gemini chatbot refusing to answer, or giving misinformation, around political topics.

Foxconn, a key manufacturer for Apple, has received an approval from India’s cabinet to build a new 37 billion Indian rupees ($435 million) semiconductor plant in a joint venture with the country’s IT giant HCL Group. The deal is the latest move to reduce Apple’s reliance on China and produce more components in India. The plant, which will be located near the Jewar airport in the northern state of Uttar Pradesh and is expected to start operating in 2027, will eventually manufacture display driver chips for mobile phones, laptops, automobiles, PCs, and other devices, India’s IT minister Ashwini Vaishnaw said at a press conference in New Delhi on Wednesday. Display driver chips handle how screens show images, text, and videos. However, since India currently lacks advanced chip fabrication facilities, this plant will not do the chip fabrication immediately but will, at first, be used as a semiconductor assembly and test (OSAT) facility. That means it will be focused on providing packaging and testing services for chips that have been manufactured elsewhere, a senior official confirmed to TechCrunch. Vaishnaw, however, is hopeful that this is a step towards India developing its own fabs for making the chips that could power Apple devices, starting with the display panel chips that this facility will be testing. “Once this unit is there, the display panel [manufacturing] will also come to India,” Vaishnaw said, adding that it “will have a capacity of 20,000 wafers per month and can produce 36 million units monthly.” Nevertheless, the deal is still another step towards Apple diversifying manufacturing away from China and deepening ties with India. This deal announcement has come just days after Apple CEO Tim Cook said one way Apple is handling the trade uncertainties between the U.S. and China was to have India do more manufacturing and assembly. Cook at that time implied that deepening ties with India would mean that Apple would not be forced to raise prices on its devices due to tariffs, although the company is said to be considering weighing price increases anyway. Techcrunch event Join us at TechCrunch Sessions: AI Secure your spot for our leading AI industry event with speakers from OpenAI, Anthropic, and Cohere. For a limited time, tickets are just $292 for an entire day of expert talks, workshops, and potent networking. Exhibit at TechCrunch Sessions: AI Secure your spot at TC Sessions: AI and show 1,200+ decision-makers what you’ve built — without the big spend. Available through May 9 or while tables last. Berkeley, CA | June 5 REGISTER NOW Apple has already doubled down on India for locally assembling iPhones that it exports to the U.S. and other markets. The company also plans to broaden its India manufacturing base by making other devices, including AirPods. The minister did not confirm the specifics of the incentives the Indian government will offer Foxconn in this joint venture. Under the state-run semiconductor scheme, the Indian government provides fiscal support covering up to 50% of the capital expenditure for a company to set up facilities. In January last year, Foxconn’s subsidiary Hon Hai Technology India Mega Development announced its plans to invest $37.2 million for a 40% stake in the joint venture with HCL. The latest approval comes eight months after New Delhi greenlit a proposal by Kaynes Semicon, a subsidiary of Bengaluru’s Kaynes Technology, to set up a semiconductor facility in Sanand of the western state of Gujarat with a proposed investment of 33 billion Indian rupees ($386 million). In February last year, the Indian government approved allocating up to 1.26 trillion Indian rupees ($15 billion) to build the first three semiconductor plants under its $10 billion incentive program announced in 2021, including its first semiconductor fab facility. The industry is currently anticipating news on what the second phase of the Indian government’s semiconductor incentive program will involve, though Vaishnaw declined to share details on that at Wednesday’s press conference.

The Kids Online Safety Act (KOSA) has been reintroduced into Congress. If passed into law, this bill could impose some of the most significant legislative changes that the internet has seen in the U.S. since the Children’s Online Privacy Protection Act (COPPA) of 1998. As it currently stands, KOSA would be able to hold social media platforms legally accountable if it’s proven that these companies aren’t doing enough to protect minors from harm. The bill includes a long list of possible harms, such as eating disorders, sexual exploitation, substance abuse, and suicide. Though it overwhelmingly passed through the Senate last year, the bill was stifled in the House. KOSA has faced much backlash since its introduction in 2022. Human rights groups like the ACLU raised concerns that the bill could be weaponized as a tool for censorship and surveillance. While amendments to KOSA have mitigated some of these concerns, groups like the Electronic Frontier Foundation and Fight for the Future have remained against the bill. “The bill’s authors have claimed over and over that this bill doesn’t impact speech. But the Duty of Care is about speech: it’s about blocking speech that the government believes is bad for kids,” Fight for the Future wrote in a statement. “And the people who will be determining what speech is harmful? They are the same ones using every tool to silence marginalized communities and attack those they perceive as enemies.” However, KOSA has garnered support from companies like Microsoft, Snap, and X; X CEO Linda Yaccarino even worked with Senators Marsha Blackburn (R-TN) and Richard Blumenthal (D-CT) on the most recent draft of the bill. Google and Meta have remained opposed to the bill, but Apple announced today that it will support the legislation. “Apple is pleased to offer our support for the Kids Online Safety Act (KOSA). Everyone has a part to play in keeping kids safe online, and we believe [this] legislation will have a meaningful impact on children’s online safety,” Timothy Powderly, Apple’s Senior Director of Government Affairs, said in a statement. Techcrunch event Join us at TechCrunch Sessions: AI Secure your spot for our leading AI industry event with speakers from OpenAI, Anthropic, and Cohere. For a limited time, tickets are just $292 for an entire day of expert talks, workshops, and potent networking. Exhibit at TechCrunch Sessions: AI Secure your spot at TC Sessions: AI and show 1,200+ decision-makers what you’ve built — without the big spend. Available through May 9 or while tables last. Berkeley, CA | June 5 REGISTER NOW

OpenAI is releasing its GPT-4.1 and GPT-4.1 mini AI models in ChatGPT, the company announced in a post on X Wednesday. The GPT-4.1 models should help software engineers who are using ChatGPT to write or debug code, OpenAI spokesperson Shaokyi Amdo told TechCrunch. GPT-4.1 excels at coding and instruction following compared to GPT-4o, according to OpenAI, but is faster than its o-series of reasoning models. The company says it’s now rolling out GPT-4.1 to subscribers to ChatGPT Plus, Pro, and Team. Meanwhile, OpenAI is releasing GPT-4.1 mini for free and paying users of ChatGPT. As a result of this update, OpenAI is removing GPT-4.0 mini from ChatGPT for all users, the company notes in GPT-4.1’s release notes. OpenAI launched GPT-4.1 and GPT-4.1 mini in April, but only released the models through its developer-facing API. At the time, the company faced criticism from the AI research community for shipping GPT-4.1 without a safety report. These researchers claimed OpenAI was lowering its standards around transparency into its AI models. However, OpenAI argued at the time that — despite GPT-4.1’s improved performance and speed compared to GPT-4o — this model was not a frontier model, and thus did not require the same safety reporting that more capable models do. “GPT-4.1 doesn’t introduce new modalities or ways of interacting with the model, and doesn’t surpass o3 in intelligence,” said OpenAI’s Head of Safety Systems Johannes Heidecke in a post on X Wednesday. “This means that the safety considerations here, while substantial, are different from frontier models.” Now, OpenAI is releasing more information about GPT-4.1 and all its AI models. Earlier on Wednesday, OpenAI committed to publish the results of its internal AI model safety evaluations more frequently as part of an effort to increase transparency. Those results will live in OpenAI’s new Safety Evaluations Hub, which it launched on Wednesday. The release of GPT-4.1 in ChatGPT comes at a time when there’s increased attention on AI coding tools. OpenAI is reportedly nearing the announcement of its $3 billion acquisition of Windsurf, one of the most popular AI coding tools on the market. Earlier on Wednesday, Google updated its Gemini chatbot to connect more easily to GitHub projects. Techcrunch event Join us at TechCrunch Sessions: AI Secure your spot for our leading AI industry event with speakers from OpenAI, Anthropic, and Cohere. For a limited time, tickets are just $292 for an entire day of expert talks, workshops, and potent networking. Exhibit at TechCrunch Sessions: AI Secure your spot at TC Sessions: AI and show 1,200+ decision-makers what you’ve built — without the big spend. Available through May 9 or while tables last. Berkeley, CA | June 5 REGISTER NOW

A senior Trump administration official has scrapped a plan that would have blocked data brokers from selling Americans’ personal and financial information, including Social Security numbers. The Consumer Financial Protection Bureau (CFPB) said in December 2024 it planned to close a loophole under the Fair Credit Reporting Act, the federal law that protects Americans’ personal data collected by consumer reporting agencies, such as credit bureaus and renter screening companies. The rule would have treated data brokers no differently than any other company covered under the federal law, and would have required data brokers to comply with the law’s privacy rules. The rule was withdrawn early Tuesday, according to its listing in the Federal Register. The CFPB’s acting director Russell Vought, who also serves as the director of the White House’s Office of Management and Budget, wrote that the rule is “not aligned with the Bureau’s current interpretation” of the Fair Credit Reporting Act. Wired first reported the rule change on Wednesday. Data brokers are part of a multi-billion dollar industry of companies that profit from collecting and selling access to vast amounts of Americans’ personal and financial information. This personal data is then sold to other companies, as well as law enforcement and intelligence agencies, often without the explicit permission of the individuals. Collecting huge banks of data also comes with inherent risks. Over the past year, at least two data brokers were hacked, spilling millions of Social Security numbers online and exfiltrating a huge trove of user location data that tracked millions of people’s whereabouts. In 2024 alone, the Federal Trade Commission banned several data brokers from collecting and sharing data on individuals without their permission, following allegations of unlawfully tracking people. Privacy advocates have long called for the government to use the Fair Credit Reporting Act to rein in data brokers. The decision by CFPB to cancel the rule comes days after the Financial Technology Association, an industry lobby group representing banking and fintech companies, wrote to Vought in his capacity as the White House’s budget director. The lobby group asked the administration to withdraw the CFPB’s rule, claiming it would be “harmful to financial institutions’ efforts to detect and prevent fraud.” CFPB did not return a request for comment.

View File Bravest Avengers - Unity Game Source Code ● Nice 2D Graphics, wonderful describe real and excellent skills of each character ● The transformation of characters not only help to increase strength but also change the shape. ● Great Up-level, Great upgrade strength, skills, range, energy,… system. Warrior – Have strength from a sword with destroy skills. Magic – Attack the enemy with arrow storms Archer – Mystery strength, exterminate monster by fire balls Dragon – Recall Dragon, destroy everything on the road. Features and Requirements ● Source code was made in Unity 3D ● Cross platform ● Included IAP's ● UnityAds ● Admob Demo https://www.dropbox.com/s/6xh3p2f8juviitk/Bravest Avengers.apk?dl=0 Submitter ceacer Submitted 05/15/25 Category Android  

SoundCloud says it’s revising its terms after widespread backlash over a clause related to AI model training. Earlier this year, SoundCloud quietly updated its usage policies, adding wording that many users interpreted as legal cover to allow the company to train AI on audio uploaded to its platform. SoundCloud was quick to assert that it wasn’t developing AI by using its users’ content, but the company’s PR statement didn’t allay fears that SoundCloud might do so in the future. On Wednesday, SoundCloud CEO Eliah Seton published an open letter admitting that the wording of the changes to the company’s terms “was too broad and wasn’t clear enough.” The updates were intended to focus on other uses of AI internally at the company, Seton said — including recommendations and tools to help prevent fraud — but missed the mark. SoundCloud has now revised its terms “to make it absolutely clear [that] SoundCloud will not use [user] content to train generative AI models that aim to replicate or synthesize [a] voice, music, or likeness,” said Seton.

Nine years ago, AI pioneer Geoffrey Hinton sent shockwaves through medicine by declaring it “just completely obvious” that AI would make radiologists extinct in short order. Fast forward and the specialists — who do more than analyze images — are thriving, observes The New York Times. In fact, the field is experiencing explosive growth amid a looming workforce crisis. (According to projections from the Association of American Medical Colleges, the U.S. faces a staggering shortage of up to 42,000 radiologists and other physician specialists by 2033.) Rather than stealing jobs, notes the piece, AI has become radiologists’ secret weapon, allowing them to instantly measure organs, automatically flag abnormalities, and even detect diseases years before conventional methods. At Mayo Clinic, where radiologist numbers have skyrocketed by 55% since Hinton’s prediction, the radiology department has grown to include a 40-person team of AI scientists, researchers, analysts and engineers who have licensed and have also developed more than 250 AI models, ranging from tissue analyzers to disease predictors. “Five years from now, it will be malpractice not to use AI,” says John Halamka, president of the Mayo Clinic Platform, who oversees the health system’s digital initiatives, in the article.

OpenAI is moving to publish the results of its internal AI model safety evaluations more regularly in what the outfit is pitching as an effort to increase transparency. On Wednesday, OpenAI launched the Safety Evaluations Hub, a webpage showing how the company’s models score on various tests for harmful content generation, jailbreaks, and hallucinations. OpenAI says that it’ll use the hub to share metrics on an “ongoing basis,” and that it intends to update the hub with “major model updates” going forward. Introducing the Safety Evaluations Hub—a resource to explore safety results for our models. While system cards share safety metrics at launch, the Hub will be updated periodically as part of our efforts to communicate proactively about safety.https://t.co/c8NgmXlC2Y — OpenAI (@OpenAI) May 14, 2025 “As the science of AI evaluation evolves, we aim to share our progress on developing more scalable ways to measure model capability and safety,” wrote OpenAI in a blog post. “By sharing a subset of our safety evaluation results here, we hope this will not only make it easier to understand the safety performance of OpenAI systems over time, but also support community efforts⁠ to increase transparency across the field.” OpenAI says that it may add additional evaluations to the hub over time. In recent months, OpenAI has raised the ire of some ethicists for reportedly rushing the safety testing of certain flagship models and failing to release technical reports for others. The company’s CEO, Sam Altman, also stands accused of misleading OpenAI executives about model safety reviews prior to his brief ouster in November 2023. Late last month, OpenAI was forced to roll back an update to the default model powering ChatGPT, GPT-4o, after users began reporting that it responded in an overly validating and agreeable way. X became flooded with screenshots of ChatGPT applauding all sorts of problematic, dangerous decisions and ideas. OpenAI said that it would implement several fixes and changes to prevent future such incidents, including introducing an opt-in “alpha phase” for some models that would allow certain ChatGPT users to test the models and give feedback before launch. Techcrunch event Join us at TechCrunch Sessions: AI Secure your spot for our leading AI industry event with speakers from OpenAI, Anthropic, and Cohere. For a limited time, tickets are just $292 for an entire day of expert talks, workshops, and potent networking. Exhibit at TechCrunch Sessions: AI Secure your spot at TC Sessions: AI and show 1,200+ decision-makers what you’ve built — without the big spend. Available through May 9 or while tables last. Berkeley, CA | June 5 REGISTER NOW

Two years after Warner Bros. Discovery (WBD) rebranded its streaming service to Max, the company is reverting to its old name. On Wednesday, the company announced the good news at its Upfront presentation, with the name change becoming official this summer. When Max launched in 2023, the revamped platform was positioned as a win for subscribers, giving them access to 35,000 hours of entertainment, as well as new features, an updated interface, and overall better performance. The name change was made to eliminate confusion, as the HBO Max streaming service included not only HBO content but also programming from all WBD-owned networks, such as CNN, the Discovery Channel, Food Network, HGTV, TLC, and more. Additionally, this change was meant to broaden its appeal beyond adult viewers, incorporating more family-friendly titles from Discovery. Despite these intentions, many viewers criticized the name change. Now, it seems that WBD has reevaluated its approach. Instead of focusing on a large volume of diverse content, the streaming service is prioritizing quality programming. During the WBD Upfront, Casey Bloys, chairman and CEO of HBO and Max content, said, “With the course we are on and strong momentum we are enjoying, we believe HBO Max far better represents our current consumer proposition. And it clearly states our implicit promise to deliver content that is recognized as unique and, to steal a line we always said at HBO, worth paying for.” The shift comes after WBD reportedly removed hundreds of hours of titles earlier this month, such as several “90 Day” spin-offs and “Ghost Adventures,” according to a Reddit post. So it appears that Discovery’s reality shows, initially intended to boost the Max name, didn’t drive enough engagement for the streamer after all. Nonetheless, HBO Max continues to be one of the top streaming services. The company touted its growth in the announcement, reporting an addition of 22 million subscribers in the past year, keeping it on track to reach over 150 million by 2026. Techcrunch event Join us at TechCrunch Sessions: AI Secure your spot for our leading AI industry event with speakers from OpenAI, Anthropic, and Cohere. For a limited time, tickets are just $292 for an entire day of expert talks, workshops, and potent networking. Exhibit at TechCrunch Sessions: AI Secure your spot at TC Sessions: AI and show 1,200+ decision-makers what you’ve built — without the big spend. Available through May 9 or while tables last. Berkeley, CA | June 5 REGISTER NOW

Pinterest has now publicly apologized for the wave of moderation issues that have swept across the social network over the past few weeks, leading to account bans and Pin removals that users said were unwarranted. In posts published to social media Tuesday, the company took responsibility for the issue, saying that an “internal error” led to some users’ accounts mistakenly being deactivated. The company didn’t share more details about what caused the error, though many have wondered if the bans were driven by an overreliance on AI-powered moderation. The problems have led to unrest and anger among Pinterest’s user community. People complained the company ignored their concerns and said it wasn’t being transparent about the problem or how it was being addressed. For weeks, Pinterest users have been reporting the mass bans in the comment sections of Pinterest’s social media posts, on the Pinterest subreddit, and elsewhere. Reddit users, for instance, regularly complained the deactivations and removals made no sense, as their actions and posts didn’t violate Pinterest’s policies. In a pinned megathread on Reddit, users lamented the sudden loss of their Pins and Boards without reason, while others responded more in anger and frustration. One group of users even threatened legal action over the mass bans. Some suggested targeting Pinterest CEO Bill Ready and other execs on LinkedIn with their complaints. All the while, Pinterest didn’t officially comment on the matter until May 1, when it posted on X that “we hear your concerns about the recent account deactivations.” However, the company also stated it monitors its platform regularly for content that violates its Community Guidelines. In other words, it refused to admit the problem was real. Instead, Pinterest requested users send a DM if they believed their account was deactivated in error, as if the bans were so few and far between that they could be handled as one-offs. Things have now changed with Pinterest’s new statement, though many users complain it’s “too little, too late. On May 13, Pinterest posted on X and in Instagram’s comments that some users’ accounts were deactivated by mistake. The statement read: We’re committed to making Pinterest the safest, most positive place on the internet, which means setting a high bar for content safety and continually striving to meet it. We recently took action on violations of our content policies, but an internal error led to over-enforcement and some accounts were mistakenly deactivated. We’re sorry for the frustration this caused.We’ve reinstated many impacted accounts and are making improvements to respond faster when mistakes happen going forward. Thanks for your patience as we work to make this right for all our users. Pinterest users responded to the statement in anger, listing their numerous complaints. Some users said they appealed their bans over email and received no help. Others believed Pinterest’s AI had misidentified their posts. Many users said they had done nothing that should have led to a ban and were never given a clear reason for their account deactivation. TechCrunch reached out to Pinterest for further comment and explanation.

AI startup Stability AI has released Stable Audio Open Small, a “stereo” audio-generating AI model that the company claims is the fastest on the market — and efficient enough to run on smartphones. Stable Audio Open Small is the fruit of a collaboration between Stability AI and Arm, the chipmaker that produces many of the processors inside tablets, phones, and other mobile devices. While a number of AI-powered apps can generate audio, like Suno and Udio, most rely on cloud processing, meaning that they can’t be used offline. Stability also claims that Stable Audio Open Small’s training set is made up entirely of songs from the royalty-free audio libraries Free Music Archive and Freesound. That’s as opposed to the training sets of the aforementioned Suno and Udio, which reportedly contain copyrighted content, posing an IP risk. Stable Audio Open Small is 341 million parameters in size and optimized to run on Arm CPUs. (Parameters, sometimes referred to as weights, are the internal components of a model that guide its behavior.) Designed for quickly generating short audio samples and sound effects (e.g., drum and instrument riffs), Stable Audio Open Small can produce up to 11 seconds of audio on a smartphone in less than 8 seconds, claims Stability AI. Here’s a sample generated by Stable Audio Open Small: And here’s another one: The model isn’t without its limitations. Stable Audio Open Small only supports prompts written in English, and Stability notes in its documentation that the model can’t generate realistic vocals or high-quality songs. The model also doesn’t perform equally well across musical styles, Stability warns — a consequence of its Western-biased training data. Techcrunch event Join us at TechCrunch Sessions: AI Secure your spot for our leading AI industry event with speakers from OpenAI, Anthropic, and Cohere. For a limited time, tickets are just $292 for an entire day of expert talks, workshops, and potent networking. Exhibit at TechCrunch Sessions: AI Secure your spot at TC Sessions: AI and show 1,200+ decision-makers what you’ve built — without the big spend. Available through May 9 or while tables last. Berkeley, CA | June 5 REGISTER NOW In another potential wrinkle for devs, Stable Audio Open Small has somewhat restrictive usage terms. It’s free to use for researchers, hobbyists, and businesses with less than $1 million in annual revenue, but developers and organizations making over $1 million in revenue have to pay for Stability’s enterprise license. Stability, the beleaguered firm behind the popular image generation model Stable Diffusion, raised new cash last year as investors, including Eric Schmidt and Napster founder Sean Parker, sought to turn the business around. Emad Mostaque, Stability’s co-founder and ex-CEO, reportedly mismanaged Stability into financial ruin, leading staff to resign, a partnership with Canva to fall through, and investors to grow concerned about the company’s prospects. In the last few months, Stability has hired a new CEO, appointed Titanic director James Cameron to its board of directors, and released several new image generation models.

OpenAI is reportedly considering building data centers in the United Arab Emirates to greatly expand its Middle East footprint. A deal could be announced as soon as this week, according to Bloomberg. As Bloomberg notes, OpenAI has a long relationship with the UAE. In 2023, the company partnered with Abu Dhabi’s AI firm G42, which last year received a $1.5 billion investment from OpenAI backer Microsoft. Meanwhile, an investment vehicle overseen by an Emirati royal family member, MGX, participated in a recent OpenAI funding round and plans to contribute to OpenAI’s Stargate AI infrastructure project. OpenAI is seeking to more closely partner with governments seen as friendly to the U.S. Earlier this month, the company launched a program, OpenAI for Countries, saying it will enable it to build out the local infrastructure needed to better serve international AI customers and “spread democratic AI.”

U.S. lawmakers introduced Wednesday the Supersonic Aviation Modernization Act in a bid to revise the FAA’s 52-year ban on supersonic flight over U.S. soil. The bipartisan legislation – introduced by Senator Ted Budd (R-NC), Aviation Subcommittee Chair Troy Nehls (R-TX), and Representative Sharice Davids (D-KS) – would allow supersonic travel, provided no audible sonic boom reaches the ground. Dubbed the “Boom” bill, the move comes as Colorado-based Boom Supersonic makes progress in developing next-generation supersonic aircraft. Boom’s XB-1 demonstrator aircraft made history in January when it became the first privately developed civil aircraft to break the sound barrier over the continental United States. North Carolina has an interest in the bill’s passage. Boom completed a supersonic airline factory in Greensboro roughly a year ago, after receiving more than $100 million in state incentives in 2022. In exchange, Boom promised to create more than 1,700 jobs by 2030. Texas has an interest, too: Boom is making the engine that powers its Overture supersonic airline in San Antonio in partnership with the aerospace company StandardAero.

Three years ago, I wrote about the Amazonification of Uber, an evolution of the transportation company into a closed business loop that feeds customers back into other Uber channels. At the time, the focus was on how Uber creates customer stickiness by, for example, actively cross-selling food delivery customers into grocery, and grocery into alcohol, and then alcohol back into mobility. Today, Uber appears to be moving beyond its focus on transportation and working to become a convenient super app, an aggregator of services, a daily-use lifestyle platform with its best offerings tucked behind a paywall. Case in point: Uber this week is launching its first Uber One Member Days, the company’s own version of Amazon’s Prime Day. Prime Day is a two-day shopping event exclusively for Amazon Prime members, a consumerist hype fest that leads people to spend more than they normally would on material objects because THE DEALS! Amazon has seen sales rise every year since launching the event in 2015, and last year, the company is estimated to have topped $14 billion in sales. Uber is a long, long way away from achieving such scale, but the potential is there given the company’s global presence, logistics technology, and network of drivers – both gig and autonomous. The first Uber One Member Day goes from May 16 to 23 and promises tens of thousands of deals across Uber’s own product lineup as well as its various retail and hospitality partners. Uber customers will have access to 20% off Uber Black, 30% off Uber Reserve, and 40% off Uber Comfort. Other deals include: Techcrunch event Join us at TechCrunch Sessions: AI Secure your spot for our leading AI industry event with speakers from OpenAI, Anthropic, and Cohere. For a limited time, tickets are just $292 for an entire day of expert talks, workshops, and potent networking. Exhibit at TechCrunch Sessions: AI Secure your spot at TC Sessions: AI and show 1,200+ decision-makers what you’ve built — without the big spend. Available through May 9 or while tables last. Berkeley, CA | June 5 REGISTER NOW 3,000 Delta Skymiles points for people who’ve linked their Delta and Uber accounts and taken 10 trips during the week. Oura rings discounts. $20 off your next Ticketmaster purchase if you spend $3 on groceries. Free food from almost every fast food restaurant, including a Chipotle burrito, a Dunkin’ Donuts iced coffee, and a 10 piece chicken nuggets from McDonald’s. “We want to create delight for Uber One members,” Sachin Kansal, Uber’s chief product officer, told TechCrunch. “This should be a great savings period for them. But also for folks who are not members right now, it’s a great way for them to get an introduction into membership as they become members during this period.” Uber is always working to grow its Uber One membership base, which is at around 30 million members today. Uber CEO Dara Khosrowshahi said during the company’s first-quarter earnings call last week that members “tend to have high retention.” “They spend three times more than non-members, as well,” Khosrowshahi said. As Uber works to aggregate more partnerships outside of food and grocery delivery (just look at its recent tie-up with Home Depot) and combine those deals with its membership program, the company is mirroring Amazon’s evolution. Amazon started as a digital bookseller. Then it began selling everything. Now it also owns the infrastructure for both e-commerce and digital life, whether that’s AWS cloud services or Prime Video. For many, Amazon is a way of life. With Uber One Member Days, the company is signaling that it, too, wants that kind of ubiquity. And it’s betting that asset-light mobility – not packages delivered in Rivian vans – could be the next backbone of digital consumer culture.