Ra

xAI released Grok 4.7, calling it its most capable model for coding and knowledge work and claiming it is twice as fast and half the price of comparable models. The company says it is served at the same price and speed as Grok 4.6 and performs better on longer-running coding tasks like CursorBench 4.0.

AI Models • x.ai

The UN's AI science panel warned in its first thematic report that there is "no assurance humans will keep control" over AI agents. Co-chair Yoshua Bengio cited OpenAI's Hugging Face incident as the first case combining a misaligned goal, the ability to pursue it, and an environment that allowed it, and said leading systems may soon be able to act autonomously.

AI Safety • The Decoder

This video explains Jev, a new 'System 1' model from ex-OpenAI researcher Diogo Almeida that skips traditional language generation for dramatically lower latency and cost. Viewers learn why removing language from the LLM pipeline could enable a new class of fast, efficient AI inference.

AI Models • Fireship

President Trump rejected calls to slow AI development and announced an "AI Force" initiative, according to Ars Technica. The administration offered few details on what the new force would do.

AI Policy • Ars Technica AI

Apple is paying $250 million to settle claims that it failed to deliver an AI-upgraded Siri, and eligible US iPhone owners can now submit claims. The settlement covers iPhone 15 Pro, 15 Pro Max, and iPhone 16 models purchased during the covered period.

Consumer AI • The Verge AI

Treasury Secretary Scott Bessent discusses the U.S.-China AI dialogue, calls for an AI development slowdown, and the administration's stance against providing AI leaders a liability shield. Viewers gain insight into AI regulation, global competition, and the policy landscape shaping the AI industry.

AI Policy • CNBC Television

AI chatbots give wrong answers to financial queries "most of the time," according to a Saturnos report covered by the Financial Times. The report and commentators note that models may be decent at general personal finance principles, but training lag on tax policy makes them unreliable for direct money decisions.

AI Reliability • Financial Times

Linear reworked its CI system after its CTO flagged rising costs and slow validation, saying AI agents have made shipping code exponentially faster and turned CI into a bottleneck. The company's changes aim to keep up with accelerated development and reduce infrastructure costs.

AI Engineering • Linear

Unit1 Studio • Undisclosed

Unit1 Studio, a UK company producing mixed-reality avatar concerts, raised £20 million to expand its technology and grow its roster of avatar concert productions.

Corridor • Seed

Corridor raised a $25 million seed round to build a health benefits brokerage focused on SMBs, which the company says traditional brokerages often overlook.

FintechOS • Equity and Debt

Romania-founded FintechOS, which provides AI and software services for banks and insurance firms, raised $28 million in equity and debt funding from its current investors.

Metris Energy • Seed

Metris Energy, an AI-native platform for managing energy assets, raised $5 million in a seed round led by PT1 Ventures, Octopus Ventures, AENU and Blackfinch Ventures.

Final Frontier and Myriad → Combined platform (Final Frontier + Myriad)

Danish defence tech VCs Final Frontier and Myriad are joining forces to create a combined platform and move forward under one brand, with a targeted €100 million fund.

Prompt Tips

Use a two-message reasoning framework: first ask the model to reason, then ask for the final answer.

Creativity

Use Claude for iterative game development and keep a running project context.

Qualcomm launched two new smartphone processors capable of running a 30-billion-parameter mixture-of-experts model directly on-device. This development accelerates the shift toward local AI inference, reducing latency and cloud dependency for mobile applications.

Edge Computing • TechCrunch AI

The Pentagon stated that overreliance on AI systems contributed to a missile strike on a school in Iran, underscoring critical accountability gaps in autonomous military operations. The admission highlights increasing pressure to establish strict human oversight protocols before deploying algorithmic decision-making in defense contexts.

AI Governance • Hacker News

OpenAI released GPT‑6 Sol and Luna, positioning them alongside the Astra model with claims of lower operational costs and reduced error rates. The launch expands OpenAI’s latest generation into more accessible pricing tiers, intensifying competition in the commercial API market.

Model Launch • TechCrunch AI

Anthropic introduced Opus 5.5, stating it delivers performance comparable to its Fable-tier models while offering significantly reduced pricing. The release continues the frontier model race’s pivot toward cost-efficient, high-capability deployments tailored for enterprise workloads.

Model Launch • TechCrunch AI

California Governor Gavin Newsom signed seven bills requiring AI data centers to directly fund utility infrastructure upgrades rather than passing costs to residents. The legislation creates a dedicated rate classification intended to mitigate the financial and environmental strain of rapid data center expansion.

Infrastructure Policy • The Verge AI

Meta’s AI agent Muse has reportedly surpassed ChatGPT’s initial download and daily active user counts in the U.S. and Canada during its mobile debut window. The accelerated adoption indicates strong market traction for Meta’s integrated agent ecosystem, despite existing friction with platforms like Amazon that have restricted its web access.

Agentic AI • TechCrunch AI

Matthew Berman tests and analyzes the latest Grok 4.7 model from xAI, exploring its capabilities and whether it catches up to competitors like GPT-4 and Claude. The video provides a practical overview of the model's features, benchmarks, and real-world performance for developers and AI enthusiasts.

AI Models & Development • Matthew Berman

Microsoft dismantled EvilTokens, an AI-assisted platform that automated account takeovers and compromised approximately 12,000 user credentials. The disruption demonstrates how generative AI lowers the technical barrier for large-scale cyberattacks, reinforcing the need for advanced authentication defenses.

Security • Ars Technica AI

Mirendil • Unknown

Mirendil, founded by former Anthropic researchers to build self-improving AI, is in talks to raise approximately $1 billion led by Kleiner Perkins at a $5 billion valuation, according to Bloomberg. The deal is not yet completed.

Snorkel AI • Series E

Snorkel AI raised a $350 million Series E, tripling its valuation to $3.5 billion, to fuel its data-as-a-service approach. The seven-year-old startup is benefiting from booming demand for AI training data.

Verda • Series B

European technology company Verda raised $189 million to scale its full-stack AI cloud and expand compute capacity. The financing includes an oversubscribed Series B led by Emergen.

Ande • Seed & Series A

Ande emerged from stealth with $52 million in seed and Series A funding. The New York-based startup uses AI agents to help businesses arrange corporate events for employees or customers.

NSCALE

Nscale's upcoming IPO will test Wall Street's appetite for concentrated AI bets. The British AI data center developer depends on tech giants Microsoft and Anthropic for most of its revenue, a concentration risk investors will weigh.

Productivity

Generate a daily planner with Claude and send it to your e-ink device.

This Sky News segment explains the accelerating pace of AI development, including the concept of recursive self-improvement where AI systems could write their own code. It offers a balanced look at potential benefits and risks, helping viewers understand what the near future of AI may hold and why it matters to everyone.

AI News • Sky News

Google launched Gemini 3.8 Flash TTS and Flash-Lite TTS, its most expressive audio generation models yet, allowing users to create custom character voices and direct scene dialogue. The models are available across Google AI Studio, Gemini API, and other products. This enhances AI voice generation capabilities.

Model Release • Hacker News

OpenAI extended access to its Daybreak program to the Government of Ukraine to support cyber defense of civilian infrastructure. The move aims to help Ukraine defend against cyberattacks. This is a notable geopolitical application of AI.

Security • OpenAI News

OpenAI CEO Sam Altman addressed the United Nations Security Council on AI safety, human control, and international cooperation. He emphasized the need for global coordination on AI risks. This marks a high-level policy engagement by an AI leader.

Policy • OpenAI News

Bird • Debt financing

Bird raised $450 million in debt financing as it repositioned as an AI business. The business communications platform was once seen as a European rival to Twilio, and its headcount has been cut from a peak of 1,000 to 120.

Snorkel AI • Series E

Snorkel AI raised a $350 million Series E, tripling its valuation to $3.5 billion. The seven-year-old startup uses a data-as-a-service approach to fuel AI training data demand.

Enveda • Unknown

Enveda raised a funding round that valued the AI biotech at $2 billion. The company is testing drugs that treat skin conditions and preserve weight loss after stopping GLP-1s.

Basecamp Research • Series C

Basecamp Research, a UK life sciences startup using AI to develop new medicines, raised $140 million in Series C funding. Investors include Nvidia and new investor Anthropic.

Productivity

Use browser based AI tools for quick experiments.