Webtagr - Technology News Summarizer

Popular：

Virtualization DNS security formal verification reachability analysis compiler errors macro conflict web extension development framework Bitmap Graphics API inconsistencies All Tags

DeepMind Workers Seek Unionization Over AI Ethics Concerns

2025-04-27

DeepMind Workers Seek Unionization Over AI Ethics Concerns

Around 300 London-based Google DeepMind employees are reportedly seeking to unionize with the Communication Workers Union, citing concerns over Google's removal of a pledge against using AI for weapons or surveillance, and its work with the Israeli military, including a $1.2 billion cloud contract. Employees feel “duped” by these actions, with at least five having resigned. This unionization effort highlights growing ethical concerns among tech workers.

(techcrunch.com)

Tech

Slate Auto: The Anti-Tesla EV Startup Targeting Affordability

2025-04-26

Slate Auto: The Anti-Tesla EV Startup Targeting Affordability

Slate Auto, a new American electric vehicle startup, has launched a radically different approach to EVs. Their truck boasts affordability, deep customization, and a decidedly analog feel—manual windows, no central infotainment screen, and even unpainted bodywork. Transforming from a two-seater pickup to a five-seater SUV, it's priced under $20,000 (with tax credits) and slated for late 2026 delivery. Backed by Jeff Bezos and targeting a market underserved by high-priced tech-focused rivals, Slate Auto plans to make its profit through extensive customization options, offering over 100 accessories and DIY-friendly upgrades.

(techcrunch.com)

Tech Customization

Perplexity's Bold Move: Copying Google's Playbook?

2025-04-25

Perplexity's Bold Move: Copying Google's Playbook?

Perplexity, an AI search engine, is building its own browser, Comet, to collect user data outside its app for targeted advertising, as revealed by CEO Aravind Srinivas. This raises privacy concerns and draws parallels to Google's antitrust lawsuit. Perplexity's partnerships with Motorola and potential deals with Samsung, mirroring Google's strategy with Chrome and Android, aim to build a comprehensive user profile. While Srinivas argues for more relevant ads, this move may fuel distrust in big tech's data tracking practices. OpenAI and Perplexity have expressed interest in acquiring Chrome if Google is forced to divest.

(techcrunch.com)

AI

Bluesky's Censorship Loophole: Turkey's Crackdown and the Client Response

2025-04-25

Bluesky's Censorship Loophole: Turkey's Crackdown and the Client Response

Following a request from the Turkish government to block 72 accounts, questions have arisen about Bluesky's decentralization. While the official Bluesky app complied with the censorship demands, third-party clients based on the AT Protocol, such as Skeets and Ouranos, can bypass censorship because they don't enforce geolocation tagging. However, this workaround is not permanent; as third-party clients gain popularity, they too may face government pressure. Developer Aviva Ruben is building Deer.social, allowing users to disable Bluesky's official moderation service and manually configure their location to counter future censorship attempts. This highlights the complex interplay of technology and politics in the pursuit of decentralized social media.

(techcrunch.com)

Tech

Controversial AI Startup Aims for Total Job Automation

2025-04-20

Controversial AI Startup Aims for Total Job Automation

Silicon Valley startup Mechanize, founded by renowned AI researcher Tamay Besiroglu, has sparked controversy with its ambitious goal: the complete automation of all work. This mission, alongside Besiroglu's connection to the respected AI research institute Epoch, has drawn criticism. Mechanize aims to automate all jobs by providing the necessary data, evaluations, and digital environments, resulting in a massive potential market but raising significant concerns about widespread job displacement. While Besiroglu argues that automation will lead to explosive economic growth and higher living standards, he fails to adequately address how people would maintain income without jobs. Despite the extreme ambition, the underlying technical challenge is real, and many large tech companies are pursuing similar research.

(techcrunch.com)

AI startup controversy

OpenAI's New Models Hallucinate More: Bigger Isn't Always Better

2025-04-18

OpenAI's New Models Hallucinate More: Bigger Isn't Always Better

OpenAI's recently released o3 and o4-mini models, while state-of-the-art in many ways, exhibit a troubling increase in hallucinations compared to their predecessors. Internal tests reveal significantly higher hallucination rates than previous reasoning models (o1, o1-mini, o3-mini) and even traditional non-reasoning models like GPT-4o. OpenAI is unsure of the cause, posing a challenge for industries demanding accuracy. Third-party testing confirms this issue, with o3 fabricating steps in its reasoning process. While excelling in coding and math, the higher hallucination rate limits applicability. Addressing model hallucinations is a key area of AI research, with granting models web search capabilities emerging as a promising approach.

(techcrunch.com)

AI hallucinations

ChatGPT's New Image-Reasoning Models Raise Privacy Concerns

2025-04-17

ChatGPT's New Image-Reasoning Models Raise Privacy Concerns

OpenAI's latest AI models, o3 and o4-mini, can accurately pinpoint the location of photos based on visual clues, even identifying specific restaurants and bars. This capability, showcased on social media, has sparked concerns about privacy risks, as malicious actors could potentially use it for doxxing. Tests reveal that even older models like GPT-4o exhibit similar location-guessing abilities, although o3 sometimes surpasses it. While not flawless, with instances of incorrect guesses and system failures, o3's performance highlights the emerging risks of powerful AI reasoning models. OpenAI hasn't yet addressed these concerns in their safety report.

(techcrunch.com)

AI

Microsoft's Giant 1-Bit AI Model: Impressive Performance, Limited Compatibility

2025-04-17

Microsoft researchers unveiled BitNet b1.58 2B4T, a groundbreaking 2-billion parameter 1-bit AI model. Trained on a massive dataset, it outperforms comparable models from Meta, Google, and Alibaba on benchmarks like GSM8K and PIQA, boasting double the speed and significantly lower memory usage. Surprisingly, it runs on CPUs, including Apple's M2. However, its reliance on Microsoft's custom bitnet.cpp framework, currently incompatible with GPUs, limits its broad adoption. While promising for resource-constrained devices, compatibility remains a major hurdle.

(techcrunch.com)

AI 1-bit AI

Google Suspends 39.2 Million Ad Accounts in 2024, AI-Powered Fraud Fight Intensifies

2025-04-16

Google Suspends 39.2 Million Ad Accounts in 2024, AI-Powered Fraud Fight Intensifies

Google suspended a staggering 39.2 million advertiser accounts in 2024, more than triple the number from the previous year, showcasing its aggressive stance against ad fraud. Leveraging large language models (LLMs) and advanced detection methods, Google proactively identified and suspended the vast majority of fraudulent accounts before they could serve ads. The company also implemented technical countermeasures against deepfakes and updated policies, resulting in a significant drop in reported deepfake ads. While acknowledging concerns about fairness, Google highlighted its appeal process and commitment to improving transparency and communication with advertisers.

(techcrunch.com)

Tech ad fraud AI anti-fraud Google Ads

OpenAI's Codex CLI: A Local AI Coding Agent

2025-04-16

OpenAI's Codex CLI: A Local AI Coding Agent

OpenAI launched Codex CLI, a local coding agent running from your terminal. Connecting OpenAI's models (including the new o3 and o4-mini) to local code and tasks, Codex CLI allows AI to write, edit code, and perform actions like moving files. This represents a step towards OpenAI's vision of an 'agentic software engineer'. Open source and lightweight, Codex CLI is supported by a $1 million API grant program for software development projects. While AI coding tools have inherent risks, Codex CLI offers a new approach to AI-assisted programming.

(techcrunch.com)

Development

Figma Sends Cease and Desist to AI Startup Lovable Over 'Dev Mode'

2025-04-15

Figma Sends Cease and Desist to AI Startup Lovable Over 'Dev Mode'

Design giant Figma has sent a cease-and-desist letter to Lovable, a rising AI no-code startup, over the use of the term "Dev Mode." While "dev mode" is a common term in software development, Figma holds a trademark for the shortcut. This isn't just a trademark dispute; it highlights the potential for larger tech companies to stifle competition. Lovable, using "vibe coding" to let users build with text prompts, directly challenges Figma's market. The outcome will be interesting, especially considering Lovable's seemingly unconcerned reaction.

(techcrunch.com)

Tech AI No-Code

Hertz Data Breach: Thousands of Customers Affected

2025-04-15

Hertz Data Breach: Thousands of Customers Affected

Hertz, a global car rental giant, has confirmed a data breach affecting thousands of customers. Personal information and driver's licenses were stolen due to a cyberattack on its vendor, Cleo, between October and December 2024. Stolen data includes names, birthdates, contact information, driver's licenses, payment card details, and workers' compensation claims. Some customers also had their Social Security numbers and other government-issued IDs compromised. Hertz notified affected customers in Australia, Canada, the EU, New Zealand, and the UK, and also reported the breach to several US states. While Hertz denies its own network was compromised, it confirms data was stolen by a third party exploiting zero-day vulnerabilities in Cleo's platform. This highlights the importance of data security and underscores the significant risk of supply chain vulnerabilities.

(techcrunch.com)

Tech

Google Joins OpenAI in Adopting Anthropic's Model Context Protocol

2025-04-14

Google Joins OpenAI in Adopting Anthropic's Model Context Protocol

Following OpenAI's lead, Google announced that its Gemini models will support Anthropic's Model Context Protocol (MCP). MCP allows AI models to directly access various data sources, including business tools, software, content repositories, and application development environments, enabling more complex task completion. This move signifies industry acceptance of MCP as an open standard and is expected to accelerate the development and adoption of AI applications. Google DeepMind CEO Demis Hassabis expressed excitement about collaborating with Anthropic and others to further develop MCP.

(techcrunch.com)

AI

Ex-OpenAI Employees Oppose For-Profit Conversion: A Battle Over Mission and Profit

2025-04-12

Ex-OpenAI Employees Oppose For-Profit Conversion: A Battle Over Mission and Profit

A group of former OpenAI employees filed an amicus brief supporting Elon Musk's lawsuit against OpenAI, opposing its planned conversion from a non-profit to a for-profit corporation. They argue this violates OpenAI's original mission to ensure AI benefits all of humanity. Several ex-staffers previously criticized OpenAI's lack of transparency and accountability, warning of a reckless pursuit of AI dominance. OpenAI responded that its non-profit arm remains, but it's transitioning to a Public Benefit Corporation (PBC). The lawsuit centers on OpenAI's structure and its impact on AI development, highlighting the complex interplay between commercialization and social responsibility in the AI field.

(techcrunch.com)

AI AI commercialization

Fintech Unicorn Solid Files for Bankruptcy: The Price of Hypergrowth

2025-04-11

Fintech Unicorn Solid Files for Bankruptcy: The Price of Hypergrowth

Solid (formerly Wise), a fintech startup once valued at $330 million, has filed for Chapter 11 bankruptcy protection. Despite claims of 10x revenue growth, profitability, and 100 customers, the company crumbled under the weight of failed fundraising and a costly legal battle with investor FTV Capital. FTV accused Solid of misrepresenting revenue and customer numbers, while Solid countersued, alleging strong-arm tactics by FTV. The lawsuit settled, but Solid, now down to three employees, is pursuing bankruptcy restructuring. Solid's downfall serves as a cautionary tale for rapidly expanding fintechs and highlights the challenges of the current funding environment.

(techcrunch.com)

Startup

AI Debugging Falls Short: Microsoft Study Reveals Limits of Code Generation Models

2025-04-11

Microsoft research reveals that even models from top AI labs like OpenAI and Anthropic struggle to debug software bugs as effectively as experienced developers. A study testing nine models showed that even with debugging tools, these models failed to successfully complete more than half of the debugging tasks in the SWE-bench Lite benchmark. The study points to data scarcity as a major factor; the models lack sufficient training data representing human debugging processes. While AI-assisted programming tools show promise, this research highlights the limitations of AI in coding, underscoring that humans remain essential.

(techcrunch.com)

Development Code Debugging

AI Shopping App Founder Charged with Defrauding Investors: The Nate Saga

2025-04-11

AI Shopping App Founder Charged with Defrauding Investors: The Nate Saga

Albert Saniger, founder of the AI shopping app Nate, has been charged with defrauding investors of over $50 million. Nate claimed its app used AI for one-click purchases across e-commerce sites, but in reality, relied heavily on hundreds of human contractors in the Philippines to manually complete transactions. Investigations revealed Nate's automation rate was effectively 0%, exposing the company's exaggeration of its AI capabilities during fundraising. Nate subsequently went bankrupt, leaving investors with near-total losses. This case highlights the issue of startups overhyping their AI capabilities.

(techcrunch.com)

Startup AI Fraud Startup Failure Investment Scam

Anthropic Launches Premium Claude Max AI Chatbot Subscription

2025-04-09

Anthropic Launches Premium Claude Max AI Chatbot Subscription

Anthropic launched a new, high-priced subscription plan for its AI chatbot, Claude Max, to compete with OpenAI's ChatGPT Pro. Max offers higher usage limits and priority access to new AI models and features compared to Anthropic's $20-per-month Claude Pro. It comes in two tiers: $100/month (5x rate limit increase) and $200/month (20x rate limit increase). This move aims to boost revenue for the costly development of frontier AI models. Anthropic is also exploring other revenue streams, such as Claude for Education, targeting universities. While subscription numbers remain undisclosed, the company's new Claude 3.7 Sonnet model has generated significant demand.

(techcrunch.com)

AI AI subscription

Waymo to Use Robotaxi Data for Generative AI, Raising Privacy Concerns

2025-04-07

Waymo to Use Robotaxi Data for Generative AI, Raising Privacy Concerns

Waymo plans to use data from its robotaxis, including interior camera video linked to rider identities, to train generative AI models, according to a leaked draft privacy policy. This data may also be used for personalized ads, raising concerns about the repurposing of rider behavior inside autonomous vehicles. While users can opt out of data sharing, the practice still sparks debate. Waymo, currently the only autonomous vehicle company generating revenue from robotaxi rides in the US (over 200,000 weekly rides), is still likely losing money, prompting exploration of alternative revenue streams like in-vehicle advertising and AI data sales. This may be a strategy to address its financial challenges.

(techcrunch.com)

Tech

SF Mayor Lurie's Tech-Driven Plan to Revitalize the City

2025-04-06

SF Mayor Lurie's Tech-Driven Plan to Revitalize the City

San Francisco Mayor Daniel Lurie is on a mission to restore the city's former glory, leveraging the tech industry's power to achieve it. He's actively engaging with tech CEOs and entrepreneurs, seeking collaborations to address pressing issues. His approach involves tackling the drug and homelessness crisis, including rolling back a program distributing drug paraphernalia. Simultaneously, he's streamlining building permits with 'Permit SF' and proposing rezoning for taller buildings to increase housing. Lurie is aggressively courting AI companies, securing commitments like Databricks' AI conference through 2030, offering tax incentives while emphasizing community investment. A new 'Partnership for San Francisco' fosters communication between businesses and city hall. Despite labor union challenges, he's pushing for autonomous vehicles. Lurie envisions San Francisco regaining its technological innovation leadership, attracting talent and businesses globally.

(techcrunch.com)

Startup Urban Revitalization

High School Founders' Calorie-Counting App Hits $2M in Monthly Revenue

2025-04-04

High School Founders' Calorie-Counting App Hits $2M in Monthly Revenue

Eighteen-year-old high schoolers Zach Yadegari and Henry Langmack launched Cal AI, a calorie-counting app, which boasts over 5 million downloads in eight months and generated over $2 million in revenue last month. Cal AI uses image recognition to calculate food calories and macros, leveraging large image models from Anthropic and OpenAI for 90% accuracy. The founders' entrepreneurial journey began in high school; Yadegari sold his first company at 16. Cal AI's success showcases the dynamism of young entrepreneurs and the potential of AI in practical applications.

(techcrunch.com)

Startup teen entrepreneurs

Anthropic Launches Claude for Education, Taking on ChatGPT

2025-04-03

Anthropic Launches Claude for Education, Taking on ChatGPT

Anthropic launched Claude for Education, a new AI chatbot service aimed at higher education, directly competing with OpenAI's ChatGPT Edu. This tier offers students and faculty access to Claude, featuring a new 'Learning Mode' to foster critical thinking. It includes enterprise-grade security and already boasts agreements with universities like Northeastern and the London School of Economics. Anthropic aims to boost revenue and increase user adoption among students through this offering.

(techcrunch.com)

AI AI education

Zelle Shuts Down Standalone App, Focusing on Bank Integrations

2025-04-02

Zelle Shuts Down Standalone App, Focusing on Bank Integrations

Payment app Zelle announced it will shut down its standalone app on Tuesday. Despite over 150 million users, only 2% utilize the app for person-to-person payments. This led to the decision to discontinue the app; users will need to re-enroll through their bank or other financial institution. Launched in 2017, Zelle's bank integrations enable instant, fee-free transfers. In 2024, it processed $1 trillion in payments, surpassing all other payment apps. This success stems from its use in larger transactions like rent, contrasting with Venmo's focus on social payments.

(techcrunch.com)

Tech payment app

OpenAI Accused of Training GPT-4o on Unauthorized Paywalled Books

2025-04-02

OpenAI Accused of Training GPT-4o on Unauthorized Paywalled Books

A new paper from the AI Disclosures Project accuses OpenAI of using unlicensed, paywalled books, primarily from O'Reilly Media, to train its GPT-4o model. The paper uses the DE-COP method to demonstrate that GPT-4o exhibits significantly stronger recognition of O'Reilly's paywalled content than GPT-3.5 Turbo, suggesting substantial unauthorized data in its training. While OpenAI holds some data licenses and offers opt-out mechanisms, this adds to existing legal challenges concerning its copyright practices. The authors acknowledge limitations in their methodology, but the findings raise serious concerns about OpenAI's data acquisition methods.

(techcrunch.com)

AI

NSO Group's Pegasus Spyware Fails to Stay Hidden: Journalists Expose Flaws

2025-03-28

A new report details attempted hacks against Serbian journalists using NSO Group's Pegasus spyware. Amnesty International researchers traced phishing links directly to NSO Group's infrastructure, exposing serious flaws in the company's and its clients' attempts at stealth. Pegasus has been used to target at least 130 individuals globally, including journalists and activists, over the years. Apple has also helped expose attacks by notifying victims. NSO Group's sale of its software to countries that misuse it is contributing to its exposure. The incident highlights NSO Group's operational security failures and the threat its spyware poses to human rights.

(techcrunch.com)

Tech

ChatGPT's AI Image Generator Sparks Copyright Debate

2025-03-27

ChatGPT's AI Image Generator Sparks Copyright Debate

ChatGPT's new AI image generator has gone viral, with users creating Studio Ghibli-style images and sparking a copyright debate. The tool can mimic the styles of specific studios, like Studio Ghibli, even transforming uploaded images into the chosen style. This functionality, similar to Google Gemini's AI image feature, raises concerns about copyright infringement, as it easily recreates the styles of copyrighted works. While legal experts argue that style itself isn't copyrighted, the datasets used to train the model may be problematic, leaving the issue in a legal gray area. OpenAI stated it allows mimicking broad styles, not individual artists', but this doesn't fully resolve the controversy.

(techcrunch.com)

AI copyright debate

TechCrunch Acquired by Private Equity Firm Regent

2025-03-25

TechCrunch Acquired by Private Equity Firm Regent

TechCrunch, after years under Yahoo's ownership, has been acquired by Regent, a private equity firm. The acquisition ensures minimal disruption to TechCrunch's operations, with the same team continuing its focus on original reporting and analysis. Regent's support will allow TechCrunch to maintain its position as a leading voice in tech news, covering Silicon Valley and beyond, while upholding its commitment to unbiased, reader-first reporting.

(techcrunch.com)

Tech

Citizen Lab Exposes Israeli Spyware Maker Paragon's Global Reach

2025-03-22

Citizen Lab Exposes Israeli Spyware Maker Paragon's Global Reach

A new Citizen Lab report reveals that Israeli spyware maker Paragon Solutions, despite claiming to sell only to democracies, has likely sold its Graphite spyware to the governments of Australia, Canada, Cyprus, Denmark, Israel, and Singapore. The report, based on analysis of server infrastructure and digital certificates, links Paragon to these governments. Paragon's spyware uniquely targets specific apps, making forensic detection harder. Meta confirmed an indicator linked to Paragon mentioned in the report. The findings raise serious concerns about the misuse of commercial spyware and the need for greater government oversight.

(techcrunch.com)

Tech spyware

Meta's AI-Powered Comment Generator: A Step Too Far?

2025-03-21

Meta's AI-Powered Comment Generator: A Step Too Far?

Meta is testing a new feature that uses AI to suggest comments on Instagram posts. The AI analyzes photos and generates three comment options. While Meta claims this enhances user experience, many users express concern, fearing inauthenticity and a decline in genuine interaction. The feature is currently in testing, with no confirmed release date. This follows previous, less successful AI experiments by Meta, highlighting the ongoing challenges of integrating AI into social media.

(techcrunch.com)

Tech

ChatGPT's Hallucinations Spark Another GDPR Complaint Against OpenAI

2025-03-20

ChatGPT's Hallucinations Spark Another GDPR Complaint Against OpenAI

OpenAI faces another European privacy complaint over ChatGPT's tendency to hallucinate false information. Noyb is supporting a Norwegian user falsely accused by ChatGPT of murdering two children and attempting to kill a third. This highlights the risks of LLMs' 'hallucinations' and GDPR's accuracy requirements. While OpenAI offers remedies like blocking prompts, this is insufficient under GDPR's right to rectification. The case could result in fines up to 4% of annual turnover and force OpenAI to modify its AI products, impacting the entire industry.

(techcrunch.com)

AI

← Previous 1 2 3 4 5 Next →