NerdNews

OpenAI's ChatGPT can now connect to users' bank accounts, allowing it to view financial transactions and provide personalized advice. The feature is currently available to select US-based ChatGPT Pro users and will eventually be rolled out to all users. However, this raises concerns about data privacy and security.

OpenAI Symphony is an open-sourced agent orchestrator that uses project-management tools to coordinate multiple coding agents, reducing the 'human attention' bottleneck and allowing for more efficient autonomous coding.

YouTube's AI deepfake detection tool is now available to all creators 18 and older, allowing them to detect and request removal of unauthorized content using their likeness. The tool uses facial recognition to scan uploaded videos and can help creators and individuals protect their identity from being used in malicious or misleading AI videos.

The trial between Elon Musk and OpenAI has revealed several key points, including questions about Sam Altman's honesty, Greg Brockman's diary entries, and Microsoft's involvement in OpenAI. The trial may have significant implications for the future of artificial intelligence and the companies involved.

The Elon Musk-OpenAI trial has raised questions about trust in AI leaders, particularly OpenAI CEO Sam Altman. The trial's closing arguments focused on Altman's trustworthiness, with Musk's lawyer grilling him about past statements. The case highlights the importance of trust in the AI industry, where privately held companies often operate behind a veil of secrecy.

Apple's revamped Siri will include an auto-deleting chat feature, allowing users to save conversations for 30 days, one year, or forever. This move emphasizes Apple's focus on privacy as a differentiator in the AI space.

YouTube is expanding its AI likeness detection program to all users over 18, allowing them to scan for facial matches and request removal of deepfakes. The feature uses a selfie-style scan to monitor YouTube and alerts users if a match is found.

Anthropic's Claude Code is a rapidly evolving tool with no long-term road map, focusing on model improvements and user signals. The company is working to address user frustrations with usage limits and is exploring ways to provide more transparency. Claude Code's product lead, Cat Wu, discusses the tool's development, its 'lean harness' approach, and the challenges of meeting diverse user needs.

OpenAI's ChatGPT now offers personalized financial advice to Pro users in the US who connect their bank accounts. This is made possible through a partnership with Plaid, allowing connections to over 12,000 financial institutions. ChatGPT can read balances, transactions, investments, and liabilities, but cannot see full account numbers or make changes to accounts.

xAI has introduced Grok Build, a coding agent and CLI for professional software engineering, available in early beta to SuperGrok Heavy subscribers. The agent aims to compete with rivals like Anthropic's Claude Code, with xAI taking user feedback to improve the product.

AMD announced that its next-generation FSR 4.1 upscaling tech will be available on older Radeon RX graphics cards, starting with RDNA 3 GPUs in July and RDNA 2 in 2027. The tech uses AI-powered upscaling and has been optimized to run on older cards without native specialized hardware.

A new survey indicates that CEOs are planning to reduce junior roles and focus on hiring mid-level positions, with 43% of CEOs looking to reduce junior roles over the next two years. This shift is driven by the increasing use of AI, which is automating tasks that early-career workers could perform, making them more vulnerable to AI-driven cost-cutting initiatives.

According to data from the Bureau of Labor Statistics, jobs with AI exposure in the US are starting to disappear, with a 0.2% drop from May 2024 to May 2025. This trend is particularly notable in customer service representatives, which saw a 4.8% decrease over the same period.

Apple will relaunch Apple Intelligence and Siri with new Apple Foundation Models, prioritizing user privacy and data security. The company will not compromise on privacy for better AI, and its AI strategy will focus on on-device processing and private cloud compute. Apple's partnership with Google will utilize Gemini for training, but Apple's models will remain the central backbone of its AI efforts.

Eclipse Ventures' investment in Cerebras Systems returned $2.5B, a 17-fold return. The firm's thesis of digitizing the physical world is gaining traction, with portfolio companies raising nearly $15B in 2025. Eclipse's focus on hardware and physical-world tech is paying off, with investments in robotics, energy, and defense.

Intel has begun testing the production of Apple Silicon chips, with a target of mass shipment in 2027. The chips will be produced on Intel's 18A-P process, and the company expects to reach full production by 2028. This move is seen as a strategic partnership between Apple and Intel, and may help Apple diversify its chip supply chain. TSMC is still expected to remain the primary supplier of Apple's processors, but Intel's involvement may help Apple reduce its dependence on a single supplier.

Monzo redesigned its data warehouse to support 100 teams and 12000 dbt models, cutting costs by 40% and improving data delivery speed by 25%. The 'meshy' approach introduces defined modeling layers, explicit interface models, and CI-enforced validation.

Snap, YouTube, and TikTok have settled a lawsuit alleging that social media addiction has harmed public school students. The suit, filed by a Kentucky school district, claims that social media has disrupted learning and created a mental health crisis. The terms of the settlement have not been revealed, and Meta is still facing a trial in the same suit.

Google announces the Googlebook, a new category of premium laptops built for Gemini, with features like a multicolor 'glowbar' and Magic Pointer. The company also reveals Gemini Intelligence, a rebrand of its advanced AI features, and security updates including automatic call spoofing detection and improved app permission management.

The article discusses how tech culture has made contempt for humanity feel enlightened, and how this mindset is driving some people towards transhumanism. Philosopher Shannon Vallor argues that classical humanism is insufficient to meet the current moment and that we need a new humanism that recognizes the dynamic and evolving nature of humanity.