/AI
An Anthropic researcher just gave us a peek at self‑improving AI
Anthropic’s Chen Yueh‑Han reveals a system that automatically improves alignment benchmarks, a step toward recursive self‑improvement.
← All topics Category
AI
Artificial intelligence, automation, and machine learning news, including applied AI strategy and industry trends
435 stories
More in AI
/AI
Accelerated Understanding Inc unveils physics‑focused AI model handling 5 trillion data points
Accelerated Understanding’s new physics‑centric AI, built on neural operators, can process 5 trillion data points in a single prompt, far outpacing language models.
/AI
London Startup Inherent Says Small AI Agent Faraday Outperformed Larger Rivals On Scientific Replication
Inherent’s new AI agent Faraday beat larger systems from Anthropic and OpenAI at reproducing scientific findings, despite its smaller size.
/AI
Ex‑Meta scientists want to bring visual AI to the factory floor
Former Meta researchers launch Perceptron to develop vision models that let robots perceive, reason and act in industrial settings.
/AI
a16z creates a $1.1B ‘Machine Age’ fund to ‘accelerate the physical buildout of AI’
Andreessen Horowitz launches a new fund focused on AI infrastructure, including chips, memory, data centers and robots.
/AI
Deep Cogito raises $43M for self‑improving AI models
Deep Cogito Inc. secured $43 million in a Series A round to expand its enterprise business and release new open‑source self‑improving AI models.
/AI
Anthropic was illegally blacklisted by the Trump administration, court rules
A federal judge ruled that the Pentagon’s blacklisting of Anthropic was unconstitutional, marking a significant legal win for the AI lab.
/AI
Anthropic Unveils Model Hardware Standard for AI Agents to Operate Physical Devices
Anthropic introduces a framework that lets AI agents safely control lab and manufacturing hardware, marking a step toward embodied AI.
/AI
EU AI Act Enforcement Begins: Bans Take Effect Immediately, While High‑Risk Compliance Quietly Delayed to End of 2027
The EU AI Act went into enforcement on August 2, 2026, imposing immediate bans on certain high‑risk AI uses while delaying other compliance obligations to December 2027.
/AI
Runable hits $21M to bet AI agents can go from building businesses to growing them
Indian AI startup Runable raises $21 million to expand its agent platform beyond building sites, aiming to help small businesses acquire customers and scale.
/AI
Bill Gates says we’ve passed AI’s danger thresholds. Now what?
Bill Gates warns that AI has crossed multiple danger thresholds, from bio‑capabilities to job‑market disruption, and calls for urgent safeguards.
/AI
Caddi Launches an AI Agent That Builds and Governs a Firm's Back‑Office Agents
Caddi’s agent automates and governs back‑office workflows for wealth, legal and insurance firms, using deterministic code and AI for decision points.
/AI
Accel-backed Keenable is indexing the web for AI agents
Keenable is building a web search index of over 100 billion documents to power AI agents, with a focus on AI‑friendly infrastructure.
/AI
Microsoft introduces ThinkingBox to assess AI agent reliability
Microsoft released an open‑source sandbox framework that tests whether AI agents can reliably perform business tasks, revealing a 25% pass rate across 20 trials.
/AI
Google Cloud Launches Gemini Enterprise for Financial Services Accessibility Statement Skip Navigation
Google Cloud unveiled Gemini Enterprise for Financial Services, a purpose‑built agentic AI solution engineered for financial professionals, announced on August 25 2026.
/AI
Advancing price‑performance for developers with GPT‑5.6 in Kiro
OpenAI rolls out its latest GPT‑5.6 model family to the Kiro software‑development agent, enabling teams to plan, build, review, and test code with higher quality and lower cost per token.
/AI
Who’s behind the new ‘stealth model’ Ox Alpha?
A mysterious new AI model called Ox Alpha has driven a frenzy of speculation about who actually built it.
/AI
Hugging Face reportedly in talks to be acquired for $13B
TechCrunch reports that Hugging Face is in acquisition discussions, potentially valuing the open‑source AI platform at $13 billion, amid growing interest in AI infrastructure services.
/AI
Is it legal to train AI models on copyrighted books? It’s complicated
A new court ruling clarified that training AI on copyrighted works can be lawful if the use is transformative, reshaping the legal landscape for AI developers.
/AI
OpenAI cuts developer pricing for frontier GPT‑5.6 Sol model by more than 20%
OpenAI announced a price cut of over 20% for its GPT‑5.6 Sol model, easing costs for developers amid rising competition.