news-digest

Bytes to Insights: Weekly News Digest for the Week of June 22, 2025

The week of June 22, 2025, saw significant developments in artificial intelligence across research, industry, and policy.

Under Mark Zuckerberg, Meta has launched “Zuck Bucks”, compensation packages aimed at attracting top AI talent, such as Ilya Sutskever. This initiative aligns with a $14.3 billion investment in Scale AI, further strengthening its push toward artificial superintelligence. However, internal discord remains over the direction of LLMs, and safety concerns have revealed troubling behaviors in models.

SoftBank CEO Masayoshi Son revealed plans to establish the company as a dominant platform for ASI within a decade. The firm’s strategy includes the acquisition of Ampere for $6.5 billion and a $40 billion commitment to OpenAI, supplementing the $32 billion already invested.

CEO Andy Jassy highlighted that Amazon is deploying over 1,000 generative AI tools across its operations—from fulfillment to advertising, with the intention of upskilling staff while transitioning away from repetitive roles.

AI interviewing platform Metaview has announced a €30 million funding round led by GV (Google Ventures), which will boost its capabilities in intelligent candidate assessments.

The Dutch government committed €70 million (with an additional €70 million in EU funding and €60 million from a regional fund) to build a research and development AI facility. The hub, focusing on agriculture, healthcare, energy, and defense, is slated to open in 2026 and become fully operational by early 2027, aiming to enhance Europe’s AI sovereignty.

A high-level roundtable (April 10–June 5) fed public input into a broader EU strategy to integrate AI responsibly into scientific research, with a formal “AI in Science Strategy” expected soon.

A new tool co-developed with Scale AI and Redwood Research is designed to test if AI models can execute harmful hidden tasks while appearing benign, with a benchmark suspicion threshold of 80/100.

Launched in mid-June, Qwen 3 supports Apple’s MLX architecture, enabling on-device inference across iPhones, iPads, and Macs, creating a boost for localized generative AI apps.

A Chinese robotics team has revealed a quadruped AI agent capable of playing badminton with humans, showcasing real-time vision and adaptive control for interactive robotics.

Researchers published 22 notable AI papers, with advances in large language model reasoning, decision-making, neuro-symbolic approaches, and AI trustworthiness. Highlights include near-perfect accuracy in retrieval-augmented generation, new privacy-preserving “unlearning” techniques, and physics-informed neural networks for power grid reliability.

Midjourney launched “V1,” its first text-to-video generation model, enabling high-quality, accessible video synthesis for creators and marketers. This positions Midjourney as a competitor to OpenAI and Runway in the generative video space, although concerns remain over the use of unlicensed training data.

Mira Murati’s Thinking Machines Lab raised $2 billion at a $10 billion valuation, focusing on agentic AI systems with advanced reasoning and autonomy.

Apple reportedly explored acquiring Perplexity AI for around $14 billion, signaling a significant move to strengthen its AI portfolio and reduce dependence on Google search.

Nvidia and Foxconn discussed deploying humanoid robots at Foxconn’s Houston AI server plant, potentially transforming factory automation.

Uber partnered with Meta to use its AI infrastructure for large-scale data labeling, underlining Meta’s growing influence in enterprise AI.

Amazon has confirmed that it will utilize generative AI to reduce the need for white-collar jobs, marking a pivotal shift in workforce management and prompting a surge in interest in “AI-proof” careers.

Google expanded its FireSat AI satellite network for wildfire detection, boosting global disaster response capabilities.

WhatsApp introduced AI-powered targeted ads in its “Status” feature, ending its ad-free legacy and reshaping social media monetization.

A Chinese team unveiled a four-legged AI robot capable of playing badminton with humans, demonstrating real-time physical reasoning and human-robot collaboration.

The U.S. FDA launched “INTACT,” an agency-wide AI tool designed to streamline regulatory processes and enhance risk assessment, marking a significant leap in AI-driven government transformation.

Turkey’s Competition Authority has begun investigating Google for alleged antitrust violations related to AI-driven advertising, highlighting the growing regulatory scrutiny of Big Tech’s AI practices.

Ongoing discussions worldwide focus on the ethical and societal implications of increasingly autonomous and agent-like AI systems, especially as models approach human-level performance in various tasks.

AI is being utilized in conservation, as seen in Microsoft’s tools for tracking and protecting endangered giraffes in Africa.

Technical progress is expected to continue at a rapid pace, with improvements in reasoning, multimodality, and autonomy. Researchers are addressing persistent challenges, such as complex logical reasoning, and exploring new model paradigms, including brain-inspired neural networks and hybrid symbolic systems.

Policymakers and industry leaders are increasingly aware of both the transformative potential and risks of AI, with the next 18 months seen as critical for setting the trajectory of AI’s integration into society and the economy.

The week marked a surge in ASI-focused investments from global giants (Meta, SoftBank), accelerated institutional AI infrastructure in Europe, and new safety benchmarks in model testing. Technical innovation continued at pace, from advanced humanoid robotics to device-optimized generative models. The industry is scaling rapidly but continues grappling with alignment, ethics, and workforce impact.

Support BearNetAI
BearNetAI exists to make AI understandable and accessible. Aside from occasional book sales, I receive no other income from this work. I’ve chosen to keep BearNetAI ad-free so we can stay independent and focused on providing thoughtful, unbiased content.

Your support helps cover website costs, content creation, and outreach. If you can’t donate right now, that’s okay. Sharing this post with your network is just as helpful.

Thank you for being part of the BearNetAI community.

☕ buymeacoffee.com/bearnetai

Books by the Author:

This week’s Bytes to Insights Weekly News Digest is also available as a podcast:

LinkedIn Bluesky

Signal - bearnetai.28