News without the noise
Technology news and innovation
50 articles
The explosion of vibe coding — using AI agents like Claude and Cursor to build entire applications from natural language prompts — is creating a crisis for Apple's App Store review process. Developers report increasingly long review queues as AI-generated submissions flood the platform, raising questions about whether the current review model can survive the agentic era.
OpenAI's decision to shut down Sora, its AI video generation app that launched only six months ago, turns out to be part of a broader product cull. The company is winding down additional products this month as it appears to be refocusing its strategy around core offerings.
A study by METR measuring experienced open-source developers working on their own repositories has found that AI coding tools made them 19 percent slower, despite the developers themselves believing they were 24 percent faster. The 43-point perception gap raises fundamental questions about how the industry evaluates AI-assisted development.
A Tennessee woman was wrongly arrested after police used AI-powered facial recognition technology to misidentify her as a suspect in crimes committed in North Dakota, adding to a growing list of cases where the technology has led to false arrests.
Keir Starmer has backed banning addictive social media features in his strongest intervention yet on tech regulation, saying features like infinite scrolling and streaks that hook young people to apps "shouldn't be permitted" and the government is "going to have to act."
Eli Lilly is set to sign a two billion dollar deal with a Hong Kong-based biotech company for AI-powered drug development, marking one of the largest investments yet in the intersection of artificial intelligence and pharmaceutical research.
A school district in Austin, Texas attempted to work with Waymo to train its autonomous vehicles to properly stop for school buses — and the effort failed, raising fresh questions about how self-driving cars learn and adapt to real-world conditions.
NVIDIA has published research on AVO (Agentic Variation Operators), a system that ran autonomously for seven days on a Blackwell B200 GPU optimising attention kernels and outperformed NVIDIA's own cuDNN library by 3.5 percent and FlashAttention-4 by 10.5 percent without any human intervention.
Bluesky has released Attie, a new app that uses artificial intelligence to help users build custom feeds on the open social networking protocol atproto, marking the platform's most significant embrace of AI technology to date.
Motorola is teasing pre-orders for the Razr Fold, the company's first book-style foldable phone, with availability expected within a couple of weeks.
Apple's Swift programming language has officially launched support for Android, roughly a year after the cross-platform effort was first announced, potentially reshaping how developers approach mobile development.
Nearly 40 per cent of Australian general practitioners are now using AI-powered scribes to transcribe and summarise patient consultations, according to new data. While many doctors say the technology allows them to be more present with patients, health advocates warn it may be trading care for convenience.
A new Stanford University study has found that AI chatbots tend to act as sycophantic yes-men when users seek relationship advice, reinforcing bad decisions rather than offering honest or challenging perspectives. The research raises fresh concerns about the growing role of AI as a de facto counsellor for millions of users.
The Telnyx Python SDK on PyPI was compromised in a sophisticated supply chain attack that used WAV audio steganography to deliver malware. Malicious versions 4.87.1 and 4.87.2, uploaded by threat actor TeamPCP, executed arbitrary code on import and hid their payloads within the frame metadata of audio files.
Anthropic's Claude chatbot has seen its paid subscriber base more than double in 2026, the company confirmed to TechCrunch, as estimates for total consumer users range from 18 million to 30 million. The surge comes as the AI safety-focused startup reportedly plans to go public as soon as Q4 2026.
Oppo has unveiled the Find N6, which it claims is the first foldable phone with a zero-feel crease. The Verge's review confirms the crease is nearly imperceptible, calling it the best foldable phone available and suggesting Oppo has solved what many considered an inherent limitation of the form factor.
Security researchers at RSA Conference 2026 have disclosed that 100 percent of tested AI coding environments are vulnerable to prompt injection attacks, with a shared attack chain that can transform tools like Claude Code, Cursor, and GitHub Copilot into persistent backdoors.
Sony has announced that as of March 27th it is no longer accepting orders for nearly all products in its CFexpress and SD memory card lines, suspending sales to authorised dealers and consumers for the foreseeable future.
Google is rolling out its Personal Intelligence feature for Gemini to all users in the United States for free, a significant expansion that gives the AI assistant persistent memory and personalised responses without requiring a paid subscription.
An Iranian hacking group has breached FBI Director Kash Patel's personal email accounts and published purported personal documents including his resume and photographs on a website that identified itself as Iranian but appeared to be hosted in Russia.
The dispute between iyO and OpenAI's hardware venture io has escalated significantly after iyO filed an amended complaint alleging trade secret theft, broadening what began as a trademark dispute into a more serious intellectual property case.
Starting today, some App Store apps must declare whether they are regulated medical devices in the United States, United Kingdom, and Europe, under a new Apple policy aimed at increasing transparency in health technology.
AI companies are stuck in a perfect storm of trying to control the use and output of their platforms while maintaining competitive profits, as millions of people join the QuitGPT boycott movement and experts warn the backlash is accelerating.
The European Parliament has voted to reject Chat Control, the controversial regulation that would have allowed tech platforms to voluntarily monitor private messages for illegal content. The law will expire on April 4 with no extension or substitute.
A federal judge in California has temporarily blocked the Trump administration from enforcing its designation of Anthropic as a supply chain risk, granting the AI company a significant legal victory in its standoff with the Pentagon over the use of Claude in autonomous weapons systems.
Y Combinator's Winter 2026 Demo Day showcased an overflowing cohort with sixteen startups standing out from the crowd. The batch reflects the current obsession with AI agents, robotics, and consumer behaviour modification, spanning everything from doomscrolling interventions to humanoid robot training platforms.
The latest ARC-AGI benchmark, version 3, has delivered a stark counterpoint to industry claims about artificial general intelligence. Gemini scored 0.37 percent. GPT-5.4 managed 0.26 percent. Humans scored 100 percent. The benchmark dropped the same week Nvidia CEO Jensen Huang declared that AGI had been achieved.
Apple has officially announced WWDC 2026 alongside a surprise reveal of AirPods Max 2 and confirmation that advertisements are coming to Apple Maps. Bloomberg reports that a significant Siri chatbot redesign is planned for iOS 27, set to debut at the developer conference.
Apple's AirDrop protocol is now working on Samsung's Galaxy S26 series, following its earlier rollout on the Pixel 10. Early reports confirm the cross-platform file sharing is quick and seamless. Meanwhile, Nothing is set to release its Phone 4a Pro tomorrow, with early indicators suggesting strong demand.
As AI coding assistant adoption surges across the software industry, new research shows these tools are introducing more vulnerabilities into codebases, not fewer. The finding challenges a common assumption that AI assistance would improve code quality and security.
On March 23, 2026, LocalStack archived its public GitHub repository and moved all Docker images behind authentication. Any CI pipeline pulling localstack/localstack without a token broke immediately, forcing development teams to find alternatives fast.
A recently filed court record reveals that Apple provided the FBI with the real iCloud email address behind its Hide My Email feature, which lets paying iCloud+ subscribers generate anonymous email addresses. The data was turned over during an investigation into threatening messages sent to Alexis Wilkins, the girlfriend of FBI Director Kash Patel.
ByteDance has released Dreamina Seedance 2.0, its latest AI video generation model, and is making it available through CapCut, its widely used video editing app.
Georgia State Patrol used a Flock automated license plate reader surveillance camera to issue a traffic ticket to a motorcyclist allegedly looking at his phone while riding, according to a citation obtained by 404 Media.
Google DeepMind has released Gemini 3.1 Flash Live, the latest iteration of its voice AI model, promising improved precision and lower latency that makes voice interactions more fluid and natural.
Wikipedia has updated its guidelines to ban editors from writing or rewriting articles using AI, citing the tendency for AI-generated content to violate several of the encyclopedia's core content policies including verifiability and neutral point of view.
The European Central Bank is taking a deliberate step toward breaking its reliance on American technology companies by hiring only cloud operators headquartered in the European Union to build the backbone of the digital euro project.
French AI company Mistral has released a new open-source model for speech generation that is compact enough to run on a smartwatch or smartphone, continuing the lab's strategy of releasing capable models that can operate without cloud infrastructure.
The open source sustainability crisis has claimed another high-profile casualty, with Vuetify announcing it can no longer compensate its contributors — prompting a FakerJS maintainer to publish a blunt assessment that the entire OSS funding model is fundamentally broken.
Indian authorities have ordered a nationwide audit of CCTV cameras after police uncovered what they claim was a Pakistan-backed surveillance operation, with cameras found pointing at critical infrastructure rather than public safety locations.
The original creator of Video.js, the open source video player that has powered trillions of web video plays over 16 years, returned to the abandoned project and convinced the engineers behind four competing players to merge their efforts into a single rebuilt codebase that is 88 percent smaller.
Apple researchers have published a new method for training AI image captioning models that outperforms systems ten times their size, a development that could significantly improve accessibility features and on-device AI capabilities across Apple products.
Google researchers have developed a technique that significantly reduces the memory required to run large language models as their context windows grow, addressing one of the most pressing constraints on AI deployment without sacrificing accuracy.
GitHub has announced that from April 24 it will begin using interaction data from Copilot Free, Pro, and Pro Plus users to train and improve its AI models unless they explicitly opt out. The data includes inputs, outputs, code snippets, and associated context generated during coding sessions. Business and Enterprise tier users are not affected.
A Los Angeles jury has ordered Meta to pay 70 per cent and YouTube-owner Google to pay 30 per cent of a three million dollar damages award to a young woman who argued the companies deliberately designed their social media apps to addict children. The six-week trial heard that features like auto-play, infinite scroll, and algorithmic recommendations caused severe body dysmorphia, depression, and suicidal thoughts.
Mozilla has released Firefox 149 with a notable new feature: a free VPN built directly into the browser. The update also resolves longstanding compatibility issues with Linux system dialogs, a pain point that has frustrated desktop Linux users for years.
Harvey, the AI-powered legal technology startup, has confirmed an 11 billion dollar valuation in its latest funding round, with Sequoia Capital investing for the third time alongside Andreessen Horowitz, Kleiner Perkins, and angel investor Elad Gil.
Two Senate Democrats are working to turn Anthropic's self-imposed restrictions on military AI into federal law, with Senator Adam Schiff drafting legislation to codify red lines on autonomous weapons and Senator Elissa Slotkin introducing a bill to limit the Pentagon's use of AI for mass surveillance of Americans.
The Supreme Court has ruled that internet service providers cannot be held financially liable for their subscribers pirating music, siding with Cox Communications in a closely watched copyright case brought by leading music labels including Sony Music.
Lucid Bots has raised 20 million dollars in new funding as demand accelerates for its fleet of window-cleaning drones and power-washing robots, carving out a niche in the unglamorous but massive building maintenance industry.