Browse any day's feed

US consumers are still spending, but JPMorgan says the cushion against higher prices is thinning

JPMorgan's CEO of consumer and community banking, Marianne Lake, indicated that while US consumers are currently resilient, the support systems for their spending are weakening. Metrics such as credit card spending and deposit balances remain strong, but Lake highlighted a concerning trend: an increasing number of consumers are experiencing wage inflation that does not keep pace with rising prices. As energy prices surged in April, many households saw their real wages decline, although some offset these costs with savings from lower tax bills. Lake emphasized that if inflation persists, the ability of wages to keep up could be jeopardized.

Uber Turns Driver Downtime Into an AI Data Factory

Uber is transforming its driver downtime into a resource for AI training, allowing drivers to engage in digital tasks such as data labeling and audio recording. This initiative aligns with CEO Dara Khosrowshahi's prediction that AI could replace 70-80% of human jobs in the next decade. Recently, Uber announced a 23% reduction in its 'People and Places' division, primarily affecting senior-level positions, as part of its strategy to streamline operations and leverage AI. The company aims to expand its AI Solutions platform, which has already grown to over 30 countries since its pilot launch in November 2024.

Microsoft AI chief walks back comments about AI taking over white-collar work

Microsoft AI chief Mustafa Suleyman has retracted his earlier assertion that AI would fully automate white-collar jobs such as those of lawyers, accountants, and project managers within 12 to 18 months. During a recent episode of the Decoder podcast, Suleyman clarified that AI is intended to assist these professionals in completing specific tasks—like sending emails or creating presentations—rather than replacing their jobs entirely. He emphasized the importance of distinguishing between tasks, which can be automated, and the broader roles that will remain essential in the workplace.

Apple’s AI promises are finally, almost, sort of here

At WWDC 2026, Apple revealed its new Siri AI, marking a significant pivot towards AI technology after years of delays. CEO Tim Cook emphasized that the new Siri will serve as a comprehensive virtual assistant, integrating seamlessly across Apple devices with features like multimodal interaction and on-device processing for enhanced privacy. While the rollout is set for later this year, the AI's capabilities, such as scheduling and contextual awareness, echo features already introduced by competitors, raising questions about Apple's ability to compete in the fast-evolving AI landscape.

Apple shares slide after big Siri AI reveal

Apple's shares dropped 3% following the announcement of new artificial intelligence software at the Worldwide Developers Conference (WWDC). The company introduced Apple Intelligence, enabling app developers to utilize AI on iPhones and Macs, and a revamped Siri based on large language models. Despite positive analyst reactions to the improvements, concerns arose over the absence of surprises and the lack of a specific release date for Siri AI, which is set for beta testing later this year. Investors reacted by selling shares, marking a potential worst day for Apple since February.

Introducing Gemma 4 12B: a unified, encoder-free multimodal model

Google DeepMind has unveiled Gemma 4 12B, a new multimodal AI model that combines high-performance capabilities with a compact design suitable for laptops. This model bridges the gap between the edge-friendly E4B and the more advanced 26B Mixture of Experts (MoE), achieving performance close to the larger model while using less than half the memory. Notably, Gemma 4 12B is the first mid-sized model to support native audio inputs, allowing it to process visual and audio data directly without traditional encoders, thus enhancing efficiency and reducing latency.

JPMorgan Chase plans to deploy more powerful AI agents this year

JPMorgan Chase is set to deploy advanced AI agents later this year that can autonomously operate for up to two hours, marking a significant evolution in corporate AI adoption. According to Derek Waldron, the bank's chief analytics officer, these agents will transition from completing single tasks to managing complex workflows across various software systems. The implementation of AI has already resulted in a 20% increase in gross sales within private banking, with the potential for individual bankers to expand their client coverage by as much as 50%. This shift reflects a broader recognition of AI's capacity to enhance revenue rather than merely reduce costs.

Sandstone raises $30M to bring AI to in-house legal teams

Sandstone has announced a $30 million Series A funding round, led by Lightspeed Venture Partners, to develop AI tools specifically for in-house legal teams. This funding comes just six months after a $10 million seed round led by Sequoia. Unlike competitors such as Harvey and Legora, which focus on private practice, Sandstone targets small and mid-sized businesses, helping them manage tasks from multiple intake channels like Slack and Jira. Co-founder Jarryd Strydom emphasizes that their platform enhances workflow automation and relationship management tailored to the unique needs of in-house legal work.

Nintendo Direct June 2026: All the news and trailers

During the June 2026 Nintendo Direct, the company announced several key titles for the Switch 2, including a highly anticipated remake of Ocarina of Time, set to release sometime this year. Additionally, Kingdom Hearts IV was revealed, launching on multiple platforms including the Switch 2, although no specific release date was provided. Other highlights included the upcoming Fire Emblem: Fortune’s Weave, which is scheduled for September 17, 2026, and a new Final Fantasy title, Final Fantasy Resonance, launching on October 22, 2026. Fans can also look forward to a demo of Kingdom Hearts III, available now ahead of its package release on October 8.

Upstart chipmakers keep challenging Nvidia. This time it's Microsoft-backed D-Matrix

D-Matrix has commenced production of its Corsair chip, which it claims can run inference workloads 10 times faster and with five times less energy than Nvidia's standalone GPUs, particularly for small workloads. Founded in 2019 and backed by Microsoft through its M12 venture arm, D-Matrix has raised approximately $500 million, achieving a valuation of around $2 billion. The company is set to begin shipping its chips to various customers, including high-profile hyperscalers and AI labs, with a majority located in the U.S. D-Matrix's approach integrates memory and compute on a single chip, which differentiates it from competitors reliant on DRAM.

Beijing's $295 billion AI buildout would require 80 percent domestic chips, locking out US suppliers

China is set to invest 2 trillion yuan ($295 billion) over the next five years to enhance its AI infrastructure, focusing on a nationwide data center network. The National Development and Reform Commission is drafting a plan that mandates at least 80 percent of the technology, including AI chips, be sourced from domestic companies like Huawei, effectively excluding US suppliers such as Nvidia and AMD. This initiative is part of the broader 'Six Networks' program, which aims to integrate critical infrastructure, with total investments potentially reaching 5 trillion yuan when including power infrastructure.

OpenAI's IPO filing, Apple updates Siri, new screwworm cases and more in Morning Squawk

OpenAI has confidentially filed for an IPO with the Securities and Exchange Commission, positioning itself for a potential public debut in the fourth quarter of 2026. This move follows a similar filing by rival Anthropic and comes amid a broader wave of tech IPOs, including SpaceX's upcoming IPO. Meanwhile, at Apple's Worldwide Developers Conference, the company announced significant updates to Siri, which will be rebranded as Siri AI, featuring new voices and enhanced conversational capabilities. This event also marked Tim Cook's final keynote as CEO, as he plans to step down in September 2026.

How an e-scooter founder raised $5 million to build space data centers

Euwyn Poon's startup Orbital has raised $5 million in seed funding to build space data centers aimed at providing AI compute capabilities. Emerging from a16z's Speedrun accelerator, Orbital plans to leverage the anticipated regular flights of SpaceX's Starship rocket to make space-based inference economically viable. The company, which employs around a dozen experts with backgrounds from Amazon LEO, SpaceX, and Northrop Grumman, aims to deploy 10,000 satellites by 2028, each contributing 100 kW of power for AI processing.

5 takeaways from airline CEOs' biggest annual gathering

At the International Air Transport Association's annual assembly in Brazil, airline CEOs confronted escalating fuel costs and their impact on profitability. IATA reported a staggering $100 billion increase in fuel expenses for 2026, leading to a projected drop in net profits from $45 billion in 2025 to $23 billion this year. Despite these challenges, demand remains resilient, with airlines like Etihad Airways seeing ticket sales stabilize at pre-conflict levels. However, concerns linger about the sustainability of this demand as operational costs rise.

How an Agent Built a 3D Paris Gallery by Chaining Two Hugging Face Spaces

An AI agent has successfully created a 3D gallery showcasing Paris monuments by chaining two Hugging Face Spaces, ideogram-ai/ideogram4 for image generation and VAST-AI/TripoSplat for 3D reconstruction. The agent autonomously produced six images, which were then converted into 3D Gaussian splats, demonstrating the ease of integration in multimedia software development. This approach highlights a shift towards a building-block economy in software, where agents can combine well-documented components to create complex outputs with minimal human input.

ECC wins $545m contract for Ghaf Woods’ Capria projects in Dubai

Engineering Contracting Company (ECC) has been appointed as the main contractor for the Capria East and Capria West projects at Ghaf Woods in Dubai, under a contract valued at Dh2 billion ($545 million). This initiative, led by Majid Al Futtaim, aims to establish Dubai’s first forest living community, featuring a range of residential options including one, two, and three-bedroom apartments, as well as duplex residences. The design emphasizes energy-efficient architecture and wellness-oriented amenities, set within a landscape inspired by the UAE’s Ghaf tree.

OpenAI now says "entirely automating everything is not the future we want"

OpenAI has shifted its strategy away from the goal of fully autonomous AI by 2028, now advocating for a collaborative approach between humans and AI systems. CEO Sam Altman and chief researcher Jakub Pachocki emphasized that while AI can assist in research, human involvement is crucial for setting direction and making judgments. They expressed concerns about the societal implications of total automation, stating, 'Entirely automating everything is not the future we want.' Additionally, OpenAI is calling for the establishment of an international organization to oversee AI development and manage potential risks, reflecting a more cautious stance on the pace of AI advancements.

OnePlus 15 now rolling out AirDrop support over Android’s Quick Share

The OnePlus 15 has begun rolling out AirDrop support over Android's Quick Share, making it one of the latest devices to join this feature alongside Google Pixel and Samsung Galaxy models. Released late last year, the OnePlus 15 is notable for being actively sold in the U.S., where AirDrop is heavily utilized. Users first reported the rollout on OnePlus forums, and confirmation has come from Android Authority, indicating that the support is being implemented as a server-side change. While the specific updates required for activation remain unclear, this addition marks a significant enhancement for OnePlus users.

Five things you need to know about AI

During a recent talk at SXSW London, key themes in AI were discussed, highlighting the rapid integration of generative AI tools into everyday tasks, which are now utilized by millions. This shift raises significant concerns about the future of jobs and the economy, as the actual impact of AI on employment remains uncertain. Additionally, the talk addressed the alarming rise of deepfakes, which have been linked to violence and misinformation, and the growing public backlash against AI technologies, evidenced by organized protests and calls for regulation.

NeuroBait: I fine-tuned a model to spark dopamine for ADHD brain

NeuroBait is an AI tool developed to assist individuals with ADHD by addressing task initiation paralysis. Unlike traditional ADHD tools that focus on diagnosis and structured to-do lists, NeuroBait generates supportive prompts that spark dopamine release, making it easier for users to start tasks. The model, based on google/gemma-3-12b-it, was trained on a small, hand-curated dataset and utilizes a fine-tuned approach to deliver empathetic, flowing responses. The creator aims to expand NeuroBait's reach to a broader audience, including those who feel overwhelmed, and plans to incorporate community feedback for further development.

Evotrex raises $30M to build the RV that doesn’t need a charging station

Evotrex, a Los Angeles-based startup, has secured $30 million in Series A funding to develop its first hybrid RV travel trailers, targeting an annual production of around 1,000 units starting next year. The funding, which brings the company's total to $46 million, was primarily sourced from a consortium of Chinese and Hong Kong investors, including GSR United Capital and Anker. Unlike competitors focusing solely on electric models, Evotrex's hybrid RV features a battery pack that can be recharged by an onboard gas engine, aiming to enable extended off-grid living. With 90% of its existing orders for the Premium trim priced at $160,000, Evotrex is positioning itself to capitalize on a market gap left by slower legacy manufacturers.

Apple says its AI is still private, even when it's running on Google's servers

Apple has confirmed that its upcoming Siri upgrade, branded as 'Siri AI,' will utilize Google's Gemini language models and run on Nvidia hardware in Google data centers. Despite this shift to third-party servers, Apple maintains its commitment to user privacy, employing a new iteration of its Private Cloud Compute system that ensures minimal data is sent off-device. This system will still leverage on-device models for simpler queries while more complex interactions will utilize cloud-based models, including the AFM 3 Cloud Pro, which operates on Google infrastructure. Apple promises to gradually enhance privacy protections during the summer preview period leading up to the fall launch of iOS 27 and related operating systems.

OpenAI says going public is "a complicated set of tradeoffs" and is unsure about the timing

OpenAI has taken a significant step toward going public by confidentially filing an S-1 registration, its first formal move in the IPO process. The company announced this on X, anticipating that the news would leak. OpenAI expressed uncertainty about the timing of the IPO, citing a complex set of tradeoffs, including the benefits of remaining private versus the potential advantages of an earlier public listing. This hesitance is compounded by the recent IPO filing from its main competitor, Anthropic, which is seen as better positioned due to a leaner cost structure and faster revenue growth.

Apple’s AI announcements: One big question, and 11 key features

At WWDC 2026, Apple introduced a suite of AI features for Siri and its devices, marking a significant advancement nearly two years after the initial announcement. Users with devices running iOS 26, including the iPhone 11, can upgrade to iOS 27, but many new AI functionalities will only be available on the iPhone 15 Pro and later. The new Siri can now execute tasks by accessing personal data and interacting with apps, while Apple emphasizes user privacy through on-device processing and Private Cloud Compute, ensuring personal data remains secure and unshared.