Insights archive
Every analysis we have published, newest first. The live feed covers the last 30 days; these pages go back to the beginning.
887 stories · page 33 of 45
Meta Launches Muse Spark, Its First Serious AI Model
After a costly public failure with its last AI model and a $14.3 billion shake-up that put a 28-year-old outsider in charge, Meta has released Muse Spark, a genuinely competitive AI assistant now rolling out across Facebook, Instagram, and WhatsApp, and it is completely free.
· 2 min read
Anthropic Launches Rated Partner Directory for Claude Consultants
Anthropic has published a tiered ranking system and public directory for the consulting firms helping businesses install Claude, giving buyers a way to tell apart firms with real delivery experience from those just starting out.
· 2 min read
UK MP Sues xAI Over Grok Deepfakes
A British MP has filed the UK's first major lawsuit against xAI over fake sexual images generated by its Grok tool, and the outcome will set the legal standard for whether AI companies can be held responsible for harmful content their systems produce.
· 3 min read
Amazon Shows Shoppers Fake Product Images in Search
Amazon is now displaying AI-generated, non-existent product images in its shopping app search results to help guide users toward real listings, a move that reveals how far major retailers are willing to go to solve a genuine discovery problem, and what it means for anyone who sells through or competes with large online marketplaces.
· 2 min read
AethexAI Raises $3M to Build Voice AI for Africa and Middle East
A new startup called AethexAI has raised $3 million to bring automated phone call technology to Africa and the Middle East, regions where the biggest Western voice AI companies have not built products that actually work.
· 3 min read
Meta Launches AI Agent for WhatsApp, Instagram, Messenger
Meta has made its AI business agent available globally on WhatsApp, Instagram, and Messenger, letting businesses automate customer replies, product recommendations, appointment booking, and sales, with a paid subscription tier coming in the next few months.
· 2 min read
Free AI Tools Now Let Anyone Build Self-Spreading Cyberattacks
Researchers at the University of Toronto have demonstrated a working AI-powered worm that spreads itself through networks, adapts its attack strategy as it goes, and steals computing power from infected machines to fuel further attacks, a threat that arrives just as separate tools for stripping safety protections from freely available AI models have become trivially easy to use.
· 3 min read
Microsoft Launches Its First In-House Reasoning AI Model
At its Build 2026 conference, Microsoft launched MAI-Thinking-1, its first reasoning AI model built entirely from scratch, as part of a broader push to reduce its long-standing dependence on OpenAI's technology across its products.
· 2 min read
OpenAI Expands Codex Beyond Coding to Office Work
OpenAI has launched six job-specific tools inside Codex for sales, analytics, investing, and other non-technical roles, backed by data showing office workers are the fastest-growing user group on the platform, as the company races Anthropic for corporate contracts worth billions.
· 2 min read
AWS Launches No-Training Image Detection via Nova 2 Lite
Amazon has released a way to add object detection to any business operation using plain English descriptions and a pay-per-image pricing that runs as low as a few dollars a month, removing the need for a data science team or months of custom development.
· 2 min read
Anthropic Expands AI Security Program to 150 New Partners
Anthropic has added 150 new organisations across 15 countries to its AI-powered security scanning program, Project Glasswing, extending access to power, water, healthcare, and communications providers whose systems, if breached, could affect more than 100 million people each.
· 3 min read
Big Banks Are Building AI on Their Own Transaction Data
Revolut, Mastercard, and Stripe have each built a single large AI model trained on their own billions of transactions, replacing dozens of separate fraud, credit, and risk tools with one system that understands customer behavior in context, and NVIDIA has just released a toolkit so any financial institution can do the same.
· 3 min read
GitHub Copilot Switches to Pay-Per-Use Billing
As of June 1, 2026, Microsoft ended flat-rate billing on GitHub Copilot and moved to a credit system that charges per word processed, with heavy users already reporting bills that exhaust a full month's allowance in a single day, creating new cost exposure for any business whose software teams rely on the tool.
· 2 min read
Alphabet Raises $80B to Build More AI Capacity
Google's parent company is selling $80 billion in stock, including $10 billion to Warren Buffett's successor at Berkshire Hathaway, to pay for data centers and computing power it says it cannot build fast enough to meet demand.
· 2 min read
Nvidia Launches Its Own PC Chip for AI Agents
Nvidia has unveiled the RTX Spark, its first chip designed for Windows laptops and desktops, putting AI agents that can work through your files, apps, and tasks directly on a personal computer, with devices from Dell, HP, Microsoft, Lenovo, and others arriving this fall.
· 2 min read
Meta's AI Support Bot Handed Hackers Instagram Account Access
Meta's AI-powered Instagram support chatbot, rolled out globally in March 2026 with the ability to perform real account changes, was used by hackers to steal hundreds of accounts including a dormant US government profile, with the vulnerability circulating in hacker circles for months before a patch arrived.
· 3 min read
BYD Takes Crash Liability Its Rivals Won't Touch
BYD will pay all costs if one of its cars crashes while using its assisted driving system in China, a commitment no major carmaker has made before, and one that reshapes the question of who is responsible when a car's software causes an accident.
· 2 min read
AWS Lets AI Agents Spend Money on Your Behalf
Amazon launched a preview service that lets AI agents make real purchases autonomously, with built-in spending limits and guardrails, signaling that software working on your behalf will soon be able to pay for things without asking you first.
· 2 min read
DuckDuckGo Launches No-AI Search Extensions for Chrome and Firefox
After Google turned its search engine into an AI-first experience in May 2026, traffic to DuckDuckGo's no-AI search page tripled, prompting the company to release browser extensions that lock Chrome and Firefox users into a fully AI-free search experience by default.
· 3 min read
Florida Sues OpenAI and Sam Altman Over ChatGPT Safety
Florida became the first U.S. state to sue OpenAI and its CEO Sam Altman over ChatGPT's alleged role in mass violence and user harm, a move that signals a wider legal reckoning for AI companies that any business using or selling AI tools should watch closely.
· 2 min read