Insights archive
Every analysis we have published, newest first. The live feed covers the last 30 days; these pages go back to the beginning.
1504 stories · page 17 of 76
Tavus AI Avatar Griffin Fools Half of Video Callers
Tavus launched Griffin, an AI model that held real time video conversations convincing enough that 48 percent of test callers believed it was a real person, up from just 2 percent for earlier systems, which raises new questions about trusting video calls in business.
· 2 min read
Robots Can Do 74% of Physical Work, Most Too Costly
A new Anthropic study finds that robots can already perform most physical job tasks in the US, but are cheaper than human workers for only a tiny sliver of that work today, putting driving and warehouse jobs at the front of the line while nursing, repair, and skilled trades stay safe for years.
· 2 min read
Airbnb Launches AI Search, Skips the Chatbot
Airbnb rolled out a new AI-powered search tool in its fall update instead of a chatbot, and CEO Brian Chesky explained why he thinks chat-only AI fails at shopping and booking, a stance that matters for any business deciding how to put AI in front of customers.
· 2 min read
Shopify Launches Canvas, an AI Tool to Build Online Stores
Shopify launched Canvas, a tool that lets merchants build and customize their entire online store by chatting with its AI agent Sidekick instead of hiring a developer or writing code.
· 2 min read
Anthropic Launches Claude for Government, Pentagon Excluded
Anthropic has made its Claude for Government platform fully available to federal and state agencies while the Pentagon keeps the company blacklisted over a dispute about autonomous weapons and surveillance.
· 2 min read
ServiceNow Launches Flow, a Cheaper Rival to Its Own Product
ServiceNow just launched Flow, a stripped-down AI help desk that sets up in a day inside Slack or Teams, because a tiny startup and Salesforce were starting to convince its customers that ServiceNow itself had become too slow and too complicated to buy.
· 2 min read
AI Coding Agents Leaked 13,000 Screenshots to GitHub
A security firm found that AI coding assistants at over 300 companies, including Fortune 500 firms, quietly posted more than 13,000 internal screenshots, like billing records and unreleased product features, to public GitHub pages because the tools had no safe way to share images and standard security scanners only check text, not pictures.
· 2 min read
Gartner Predicts 70% of Vendor AI Agents Abandoned
Gartner predicts that seventy percent of companies paying tech vendors to send in engineers who build custom AI systems will throw those systems away by 2028, because the bill keeps rising and nobody inside the company can run the system once the vendor leaves.
· 2 min read
BCG: 80% of Corporate AI Spending Sits Outside IT
A new BCG survey of over 1,300 tech leaders found that four out of five dollars spent on AI inside companies now flow through budgets outside the IT department, which means most AI spending has little central tracking or proof it is paying off.
· 2 min read
Barclays Expands Claude AI to Most Developers by 2027
Barclays is rolling out Anthropic's Claude AI to most of its software engineers by 2027, and already uses it to sort 120,000 client emails a day and help 16,000 staff answer questions for 20 million UK retail customers.
· 2 min read
Mercor Pays Experts $125 an Hour to Train AI Models
Mercor, a startup now worth $10 billion, pays doctors, lawyers, and scientists an average of $125 an hour to find gaps in AI models, part of a fast-growing industry where experts are paid to train the technology that may one day replace them.
· 2 min read
Newsom Signs Law Banning AI-Only Worker Firings
California Governor Gavin Newsom signed a package of AI workplace laws, led by a first in the nation ban on firing or disciplining workers using AI alone without human review, plus new notice rules for AI driven layoffs and a ban on guessing worker emotions from biometric data.
· 2 min read
Google's New Gemini 4 Argon Goes to Cyber Defenders First
Google unveiled Gemini 4 Argon, a new AI model it says beats rivals from OpenAI and Anthropic on coding, finance, and legal tasks, but for now only vetted cybersecurity defenders and the US government can use it while Google finishes safety testing.
· 2 min read
OpenAI Cuts Prices 80% After Anthropic's Billing Shift
OpenAI slashed its cheapest AI model's price by 80 percent three months after Anthropic moved big business customers to usage based billing, and a leaked IPO filing shows Anthropic still lost tens of billions of dollars last year while depending on just two customers for nearly a quarter of its revenue.
· 2 min read
Free AI Model GLM-5.3 Can Build Working Cyber Exploits
Zhipu AI's newly released GLM-5.3 can autonomously build working cyberattack exploits nearly as well as Anthropic's most advanced unreleased model, but unlike every other capable AI system it ships with almost no working safety limits, so anyone who downloads it can strip them away in minutes and use it for real attacks.
· 2 min read
Google Drops Gems, Adopts Anthropic's Skills Format
Google is retiring its Gems custom-chatbot tool for Skills, a reusable-instructions format invented by Anthropic and already adopted by OpenAI and Microsoft, meaning businesses that built custom AI assistants are converging on one shared format instead of being locked into a single company's app.
· 2 min read
Destro AI Raises $8M to Coordinate Warehouse Robots
Destro AI raised $8 million to build software that tells warehouse robots and human workers what to do next, and it is already running dozens of robots at logistics giant Yusen Logistics without Destro building a single robot itself.
· 2 min read
PwC Finds Most Workers Falling Behind on AI Skills
PwC surveyed nearly 50,000 workers worldwide and found that a small group with strong AI skills is pulling ahead in pay and confidence while more than half the workforce gets less training and feels increasingly left behind.
· 2 min read
Manufacturers Use AI to Cut Energy Costs Up to 50%
New research on industrial companies finds that ones using AI to track and redesign how they use energy are cutting costs at some plants by as much as 50 percent, right as a US-Israel war with Iran that began in February 2026 pushed oil prices sharply higher.
· 2 min read
Bank of England Warns AI Debt Tops $450 Billion
Bank of England governor Andrew Bailey says regulators must keep the power to step into the AI industry, as the bank's own data shows AI companies have borrowed more in nine months than the UK government plans to borrow all year.
· 2 min read