apple, meta and openai Updates

Apple is reportedly shifting its focus from developing a successor to its Vision Pro headset to creating AI-powered smart glasses, aiming to compete with Meta's offerings. The company plans to release a version without a display next year, with a display-equipped model potentially arriving in 2027 or 2028. These glasses will heavily rely on voice commands and AI, utilizing a custom chip. Meanwhile, OpenAI has introduced a benchmark called GDPval to measure AI productivity for business value, finding that models like Claude Opus and GPT-5 perform at near-expert levels for professional tasks. Hybrid human-AI approaches show significant gains in speed and cost reduction. In the legal sector, Kennedys law firm is partnering with Spellbook to offer AI training for junior lawyers, addressing concerns about automation impacting traditional training methods. On the security front, NVIDIA's AI Red Team is advising caution with LLM-generated code due to risks of remote code execution and highlighting vulnerabilities in data access for RAG systems. In healthcare, the Pope emphasizes that AI can assist but not replace the human touch in medicine, stressing the importance of compassion and personal contact. Vanderbilt University Medical Center is also exploring AI's expanding role in healthcare through its annual symposium. Politically, AI has been used to generate images for commentary, with Vice President JD Vance defending President Trump's AI videos of Hakeem Jeffries as jokes, while California Governor Gavin Newsom's office used an AI-generated Marie Antoinette image to mock Trump. The issue of AI-generated 'workslop' is also highlighted, with a study showing it wastes employee time and damages reputations, prompting calls for AI literacy training and organizational guardrails. Finally, BrandPilot AI has expanded its AdAi tool to audit Google Shopping Ads, aiming to reduce wasted ad spend for retailers.

Key Takeaways

• Apple is pausing Vision Pro sequel development to focus on AI smart glasses, with a display-less version planned for next year and a display version for 2027/2028.
• OpenAI's new GDPval benchmark shows AI models like Claude Opus and GPT-5 can perform professional-level tasks, with human-AI collaboration boosting efficiency.
• NVIDIA's AI Red Team warns against directly executing LLM-generated code due to security risks like remote code execution.
• The Pope states that AI can aid healthcare but cannot replace the essential human elements of compassion and personal contact in patient care.
• AI-generated 'workslop' is a growing problem, wasting employee time and harming reputations, necessitating AI literacy training and organizational guidelines.
• Law firm Kennedys is launching an AI training program with Spellbook for junior lawyers to adapt to an AI-driven legal field.
• California Governor Gavin Newsom's office used an AI-generated image of Marie Antoinette to mock President Trump.
• Vice President JD Vance defended President Trump's AI-generated videos of Hakeem Jeffries as jokes, dismissing claims of racism.
• BrandPilot AI's AdAi tool now audits Google Shopping Ads to help retailers reduce wasted ad spend.
• South Marion High School is receiving nearly $260,000 to establish an AI program.

Apple shifts focus from Vision Pro to AI smart glasses

Apple is reportedly pausing development on a lighter version of its Vision Pro headset to concentrate on creating smart glasses. The company plans to release a pair without a display next year, with a version featuring a display possibly arriving in 2027. These smart glasses will include speakers, cameras, and rely heavily on voice commands and AI, similar to Meta's offerings. Apple is also developing a custom chip for these glasses.

Apple prioritizes AI smart glasses over Vision Pro sequel

Apple is reportedly halting work on a successor to its Vision Pro headset to focus on developing new AI-powered smart glasses. This strategic shift aims to compete with other companies in the growing smart glasses market. The company is reallocating resources from the headset project to accelerate the development of its glasses.

Apple halts Vision Pro update for AI smart glasses development

Apple has stopped a planned update for its Vision Pro headset to shift employees to its smart glasses project, according to Bloomberg News. The company was working on a cheaper, lighter Vision Pro but is now prioritizing the glasses, which aim to rival Meta's products. Apple is developing two types of smart glasses: one without a display for next year and another with a display for 2028, focusing on voice interaction and AI.

Apple plans AI glasses to challenge Meta's market lead

Apple is reportedly developing AI-powered smart glasses and may announce a version without a display next year, aiming for a 2027 release. This move shifts focus from a lighter Vision Pro headset. The company is working on two types of glasses, one with a display and another without, both designed to heavily utilize voice commands and AI. This strategy aims to compete directly with Meta's existing smart glasses offerings.

Vance defends Trump's AI videos as jokes, not racist

Vice President JD Vance stated that AI-generated videos posted by President Trump of House Democratic Leader Hakeem Jeffries were intended as a joke. Democrats had criticized the videos, which featured Jeffries with a superimposed mustache and sombrero, as racist. Vance argued that the videos were humorous and reflected a lighthearted approach to political commentary, dismissing claims of racism.

Newsom uses AI to mock Trump with Marie Antoinette image

California Governor Gavin Newsom's press office used an AI-generated image of Marie Antoinette with President Trump's face as their X profile picture. The image was used to mock Trump's plans for a White House ballroom amid a government shutdown and debates over healthcare funding. Newsom's office has frequently used its social media to criticize Trump.

OpenAI benchmark measures AI productivity for business value

OpenAI has introduced GDPval, a benchmark to test if AI models can perform professional-level tasks, addressing CFOs' demands for measurable return on investment. The benchmark evaluated 1,320 tasks across 44 occupations, showing leading AI models like Claude Opus and GPT-5 perform near expert levels in many cases. Hybrid approaches, where humans review AI outputs, significantly speed up task completion and reduce costs, indicating AI's growing economic value.

AI 'workslop' creates productivity issues and damages reputations

AI-generated 'workslop'—polished but unhelpful content—is wasting employee time and harming workplace trust. A recent study found 40% of desk workers encountered workslop, costing companies millions annually and damaging senders' reputations. Experts advise investing in AI literacy training, establishing organizational guardrails, and modeling responsible AI use to combat this growing problem.

Kennedys and Spellbook launch AI training for lawyers

Law firm Kennedys is partnering with AI platform Spellbook to create a new training program for junior lawyers. This initiative addresses concerns that AI and automation will reduce entry-level tasks traditionally used for training. The program will use AI-assisted drafting and simulated scenarios to equip lawyers with skills for an AI-driven legal field, aiming to ensure they can work effectively alongside AI tools.

Pope says AI cannot replace human touch in medicine

Pope Leo XIV told medical professionals that while artificial intelligence can assist in healthcare, it can never replace the essential human element of a doctor's care. He emphasized the importance of personal contact, dialogue, and compassion in the patient-doctor relationship. The Pope stated that algorithms cannot provide the comfort and closeness that human caregivers offer.

VUMC symposium explores AI's role in healthcare

Vanderbilt University Medical Center hosted its third annual symposium exploring the growing role of artificial intelligence in healthcare. Experts discussed AI's impact on health care, innovation, and entrepreneurship. Topics included AI data security, ethical considerations for chatbots, and insights from founders using AI to transform the medical field.

NVIDIA AI team shares practical LLM security advice

The NVIDIA AI Red Team (AIRT) has identified key security vulnerabilities in AI systems, particularly Large Language Models (LLMs). They advise against executing LLM-generated code directly due to risks of remote code execution and recommend secure sandboxing. AIRT also highlights issues with insecure access controls in data sources used by retrieval-augmented generation (RAG) systems and warns about data exfiltration through active content rendering.

BrandPilot AI expands AdAi to Google Shopping Ads

BrandPilot AI has expanded its AdAi tool to audit Google Shopping Ads, helping retailers reduce wasted ad spend. The tool identifies instances where brands bid against themselves or pay for unnecessary clicks on product listings. This expansion aims to improve return on ad spend (ROAS) for Shopping campaigns, which represent a significant portion of retail search ad expenditure.

Ocala high school receives $260,000 for AI program

South Marion High School in Ocala, Florida, will receive nearly $260,000 in funding to launch an Artificial Intelligence (AI) program. The grant from the Florida Department of Education will support specialized programs in AI, avionics, digital cinema, entrepreneurship, and agribusiness. This funding aims to equip students with skills for high-demand industries and support economic growth.

Sources

• Apple sidelines lighter Vision Pro to prioritize smart glasses
• Apple tables Vision Pro sequel to pursue smart glasses: report (AAPL:NASDAQ)
• Apple halts Vision Pro overhaul to focus on AI glasses, Bloomberg News reports
• Apple Plans AI Glasses, Challenging Meta’s Dominance: First Product Could Be Announced Next Year, Report Says
• Trump's AI videos of top Democrat were a joke, not racist, Vance says
• Newsom mocks Trump with AI-generated Marie Antoinette image
• OpenAI Benchmark Puts AI Productivity to the Test as CFOs Demand ROI
• AI ‘Workslop’ Could Be The Biggest Threat To Productivity
• Kennedys partners with Spellbook to launch groundbreaking legal training programme for the AI era
• Pope: AI can never replace the human touch of the doctor
• Symposium explores AI’s role in health care
• Practical LLM Security Advice from the NVIDIA AI Red Team
• BrandPilot AI Expands Product Offering with AdAI for Google Shopping Ads
• New Ocala high school gets $260,000 grant for Artificial Intelligence (AI) program