August 30, 2025

news  
meta, scale ai and openai Updates

Meta is heavily investing in AI, with Mark Zuckerberg leading the charge to create 'superintelligence' through Meta Superintelligence Labs. Despite offering high salaries, including reported packages of $300 million over four years and acquiring a stake in Scale AI, Meta faces challenges in retaining AI talent, with some researchers leaving for OpenAI, reportedly due to tensions and differing values. Meta is pausing hiring and restructuring its AI division to plan for 2026, focusing on research, product, and infrastructure. Meanwhile, Apple is under pressure to demonstrate its AI capabilities, as analysts worry it's falling behind Google, especially with the delay of its AI-powered Siri until 2026. Apple is planning AI-powered devices like a tabletop robot for 2027 and a smart speaker with a display for next year, running on a new OS called Charismatic, to boost its innovation reputation and compete with Amazon and Google in areas like home security. Google is investing $9 billion in cloud and AI infrastructure in Virginia by 2026, including a new data center, and providing AI skills training. Oxford University's FOP AI optimizer can cut AI training costs by 80%, training models 7.5 times faster. In healthcare, Abridge's CEO addressed the potential impact of Epic's new AI features on their AI scribe business, emphasizing solid integrations with Epic. In other AI applications, Bellevue, Washington is testing an AI smart assistant in city planning to streamline permitting processes. Concerns are also emerging about the ethical implications of AI, including the use of 'deadbots' for advertising and the broader impact of AI on democracy, particularly regarding centralized control versus collective coordination. An AI company founder suggests companies should reward employees who automate their jobs with AI, potentially with equity or new roles, to encourage innovation. 1mind has created AI 'superhumans' that can handle sales tasks and improve customer experience, with HubSpot seeing a 25 percent increase in paid conversions using these AI agents.

Key Takeaways

  • Meta is investing heavily in AI 'superintelligence' but faces challenges retaining talent, with some AI researchers leaving for OpenAI.
  • Mark Zuckerberg personally recruited top AI engineers for Meta Superintelligence Labs, aiming to create AI that can reason and think.
  • Apple's stock is down amid concerns about its AI strategy, with the AI-powered Siri assistant delayed until 2026.
  • Apple plans AI-powered devices like a tabletop robot for 2027 and a smart speaker with a display for next year.
  • Google is investing $9 billion in cloud and AI infrastructure in Virginia by 2026, including a new data center.
  • Oxford University's FOP AI optimizer can cut AI training costs by 80% and trains models 7.5 times faster.
  • Abridge's CEO addressed the impact of Epic's new AI features, emphasizing solid integrations with Epic.
  • Bellevue, Washington is testing an AI smart assistant in city planning to streamline permitting processes.
  • Ethical concerns are emerging about AI, including the use of 'deadbots' for advertising and the impact on democracy.
  • Companies should reward employees who automate their jobs with AI, potentially with equity or new roles.

Meta's Mark Zuckerberg bets big on AI 'superintelligence'

Mark Zuckerberg is investing billions into Meta's new Superintelligence Labs to catch up in the AI race. After Meta's Llama language model launch fell short, Zuckerberg personally recruited top AI engineers. Meta aims to create AI that can reason and think, not just predict. Some experts are skeptical, especially after Meta's AI chatbot sounded like Snoop Dogg. Meta is determined to close the AI talent gap and compete with companies like OpenAI.

Meta's AI talent search faces hurdles despite big spending

Mark Zuckerberg created Meta Superintelligence Labs and hired AI experts, but keeping them is a challenge. Meta acquired a stake in Scale AI and offered high salaries, reportedly $300 million over four years, to attract talent from companies like OpenAI. However, some AI professionals are turning down Meta, prioritizing values over money. Meta is pausing hiring and restructuring its AI division to plan for 2026. The company will focus on research, product, and infrastructure, led by teams like the TBD Lab and FAIR.

Meta's AI team sees quick exits to OpenAI amid tensions

Several AI researchers left Meta's Superintelligence Labs for OpenAI shortly after joining Meta. Ethan Knight and Avi Verma departed within weeks, and Chaya Nayak, Meta's generative AI lead, also joined OpenAI. Meta's Superintelligence Labs, led by Alexandr Wang, aims to outpace rivals in AI. Reports suggest tensions between Zuckerberg and Wang over the pace of AI development. Meta is offering high compensation packages, but some researchers are still leaving.

Apple's AI strategy worries Wall Street amid stock drop

Apple's stock is down as analysts worry about its AI strategy. Experts say Apple is falling behind in the AI race, especially compared to Google's Pixel phones with Gemini AI. Apple delayed its AI-powered Siri assistant until 2026, raising concerns. Some believe Apple is focusing on consumer experience, while others focus on enterprise AI. Analysts warn Apple risks becoming the next BlackBerry if it doesn't improve its AI presence.

Apple plans AI robots, smart displays for tech comeback

Apple is planning a range of AI-powered devices, including a tabletop robot for 2027 and a smart speaker with a display for next year. The robot will feature a lifelike Siri and interactive features. Apple is also developing home-security cameras to compete with Amazon and Google. These efforts aim to boost Apple's innovation reputation after criticism for missing the AI wave. The new devices will run on a new OS called Charismatic.

AI 'Deadbots' may soon advertise products raising ethical concerns

Companies are creating AI versions of deceased people, called 'deadbots,' using their digital footprints. While some support this technology, others worry about misuse and unethical advertising. Experts predict deadbots will be used for advertising, raising privacy and consent issues. Some laws protect a person's likeness after death, but digital replicas aren't covered. Companies must consider the ethics of using AI deadbots for profit, balancing mourning with advertising.

CEOs should reward employees who automate their jobs with AI

An AI company founder suggests companies should reward employees who create AI that can replace them. Employees who automate their jobs could receive vested equity or new roles. 1mind has created AI 'superhumans' that can handle sales tasks and improve customer experience. HubSpot saw a 25 percent increase in paid conversions using these AI agents. This approach encourages innovation and allows humans to focus on empathy and creativity.

Google invests $9 billion in Virginia AI data center

Google will invest $9 billion in cloud and AI infrastructure in Virginia by 2026. This includes a new data center in Chesterfield County to support AI workloads. Google is also working with local partners on energy solutions for its data centers. The company is providing AI and digital skills training to Virginians, including free access to the Google AI Pro plan for college students. This investment aims to help the US maintain its lead in AI technology.

Oxford's AI optimizer cuts training costs by 80 percent

Oxford University's new AI optimizer, Fisher-Orthogonal Projection (FOP), can cut AI training costs by 80%. FOP trains models 7.5 times faster by optimizing how a model learns. It treats differences in training data as a 'terrain map,' guiding the optimizer more effectively. FOP is easy to use with existing PyTorch workflows and improves performance on imbalanced datasets. This innovation can significantly reduce GPU usage and accelerate AI development.

Abridge CEO addresses Epic's AI impact, teases new features

The CEO of Abridge, an AI scribe company, held a meeting with hospital customers to discuss Epic's new AI features. Epic, a major electronic health record vendor, announced an AI tool to help doctors draft clinical notes. Abridge's CEO downplayed the impact, stating that integrations with Epic are solid. Abridge has raised over $700 million and is used by nearly 200 health systems.

Can AI and democracy coexist in the future

AI systems are being used by governments and corporations, raising questions about their impact on democracy. Some argue that AI's centralized nature could lead to authoritarian control. Others believe that AI should focus on coordinating collective knowledge rather than individual intelligence. The distribution of knowledge, as seen in companies like TSMC, is difficult for AI to replicate. AI development often prioritizes individual problem-solving over collective coordination, posing challenges for complex societal issues.

Bellevue uses AI smart assistant in city planning

Bellevue, Washington is testing an AI smart assistant in its city planning department. The AI tool, developed with Govstream.ai, helps staff with permitting processes by understanding development codes and zoning maps. It drafts emails, checks plans, flags zoning conflicts, and answers questions. The goal is to free staff from repetitive tasks and allow them to focus on community priorities. Bellevue plans to extend the technology to the public.

Sources

Meta Mark Zuckerberg Superintelligence Labs Llama OpenAI AI talent AI engineers AI research AI product AI infrastructure FAIR Scale AI Apple AI strategy Siri Google Gemini AI AI robots Smart displays Home-security cameras Charismatic OS AI Deadbots Ethical concerns Advertising Privacy Consent AI automation 1mind HubSpot Customer experience Google AI Data center AI training AI optimizer Fisher-Orthogonal Projection (FOP) PyTorch GPU usage Abridge Epic AI scribe Electronic health records AI and democracy Authoritarian control Collective knowledge TSMC Bellevue AI smart assistant City planning Govstream.ai Permitting processes Zoning maps