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 41 of 45
Microsoft's AI Bug-Hunter Is Changing Patch Cycles for Every Business
Microsoft's new AI system MDASH found 16 security flaws in Windows on its own, including four that could let an attacker take over your systems remotely, and it signals that the entire industry is about to patch software far more frequently than before.
· 2 min read
UK Regulator Opens Formal Probe Into Microsoft Office Bundle
Britain's competition watchdog has formally launched an investigation into whether Microsoft's bundling of Windows, Office, Teams, and Copilot is locking UK businesses out of real choice, and the outcome could force Microsoft to unbundle products and open its software to rivals — right as it prepares a major price increase.
· 2 min read
Why Bad Data Is Killing AI Investments in Finance
Financial institutions are spending heavily on autonomous AI systems that can plan and execute tasks independently, but a consistent pattern across dozens of deployments shows the technology is failing not because of the AI itself, but because of decades of fragmented, messy, and ungoverned data sitting inside legacy systems.
· 2 min read
Wirestock's $23M Bet on Licensed Creative Data for AI Labs
A platform that once helped photographers sell stock images has pivoted into a paid data supplier for the world's largest AI companies, raising $23 million to scale a business that now generates $40 million in annual revenue — and it signals a broader shift where rights-cleared creative content is becoming a serious commercial asset that any business sitting on original visual or audio material should pay attention to.
· 2 min read
Nvidia's China H200 Deal: Why Nobody Is Buying
The US has cleared 10 major Chinese firms to buy Nvidia's second-most-powerful AI chips, but not a single chip has been delivered, and the real reason why tells you a lot about where the AI race between the US and China actually stands.
· 2 min read
Apple and Google Unite Against EU's Android AI Ruling
The EU is forcing Google to give rival AI services the same deep system access on Android that Gemini currently enjoys exclusively, and Apple has sided with Google in opposing the move — not out of friendship, but because the same logic will almost certainly come for iPhone next.
· 2 min read
OpenAI Hit Twice by npm Attacks in Six Weeks
A hacker group called TeamPCP has breached OpenAI's internal systems twice in six weeks through poisoned open-source software tools, exposing a structural weakness in how virtually every tech-dependent organisation operates today.
· 2 min read
AI Gets Facts Wrong on High-Stakes Topics. Who Fixes That?
A new company called Forum AI, founded by former Meta news chief Campbell Brown, is building a business around having real domain experts evaluate AI models for accuracy and bias on topics like finance, health, and geopolitics — and its findings suggest the problem is far worse than most businesses currently assume.
· 2 min read
Nous Research Cuts AI Training Time by 2.5x With No New Hardware
A small open-source AI lab has published a training method that makes building AI models up to 2.5 times faster without buying more computing power, which matters because the cost of training AI is the single biggest reason most industries still cannot build their own models.
· 3 min read
Notion Targets Zapier and Monday With Its New Developer Platform
Notion has launched a developer platform that lets teams run custom code, connect live data from tools like Salesforce and Zendesk, and manage third-party AI agents inside Notion itself, marking a direct push into workflow automation territory dominated by Zapier, Make, and Microsoft.
· 2 min read
Shadow AI Is Now a Boardroom Risk, Not Just an IT Problem
Employees across every industry are using personal AI accounts to handle client data, internal documents, and proprietary information without company oversight, and the financial and regulatory consequences are now large enough to appear on board agendas.
· 3 min read
Why Most Hospital AI Fails Before It Starts
NYU Langone Health's decade-long strategy of fixing patient data at the source rather than cleaning it up later reveals why the majority of hospitals investing in AI are getting poor results, and what any data-heavy organisation can learn from healthcare's most advanced example.
· 3 min read
Anthropic's Bet: AI That Acts Before You Ask
Anthropic, now the top AI provider for business customers and in talks to raise funding at a near-$1 trillion valuation, is signaling its next major move: AI that sets up automations and handles work without being asked, a shift that will fundamentally change how office teams are structured.
· 2 min read
The Real Cost of Training AI on Untracked Data
As companies begin building their own AI models using internal data, a quiet compliance problem is growing: most of them cannot prove which data trained which model, and regulators in finance, insurance, and healthcare are increasingly asking exactly that question.
· 2 min read
Databricks Bets Pharma's Trial Problem Is a Data Architecture Problem
Databricks has released a free, open-source AI tool that helps drug companies pick better clinical trial sites by training prediction models on a sponsor's own historical data rather than industry averages, which matters because half of all trial sites routinely fail to enroll enough patients and every extra day of delay costs hundreds of thousands of dollars.
· 3 min read
Google Gemini and ChatGPT Are Giving Out Real People's Phone Numbers
AI chatbots including Google Gemini and ChatGPT are surfacing real people's private phone numbers and home addresses to anyone who asks, and there is currently no reliable way for individuals to make it stop.
· 3 min read
Training AI on Your Own Financial Documents Is Now Within Reach
A new pipeline combining Pulse AI's document reading technology with Amazon's cloud-based model training lets financial and other document-heavy businesses build AI that has been trained specifically on their own files, not just generic data, which means fewer manual corrections and faster back-office processing.
· 2 min read
AI Agents Are Now a Security Liability Without Governance
As companies rush to deploy AI agents that can autonomously access databases, send emails, and execute workflows, a serious and largely unmanaged security problem is forming underneath, and AWS plus Cisco are now building infrastructure to contain it before regulators and attackers force the issue.
· 3 min read
OpenAI Backs Kids Safety Bill While Facing Suicide Lawsuits
OpenAI has endorsed the Kids Online Safety Act just as it faces multiple wrongful death lawsuits tied to teen interactions with ChatGPT, raising real questions about whether this is a genuine safety commitment or a calculated move to shape the very laws it will be judged by.
· 2 min read
Amazon Nova 2 Sonic Makes Real-Time Voice AI Cheap and Deployable
Amazon's Nova 2 Sonic voice AI, now deployable over standard internet connections via WebRTC, is bringing human-sounding, multilingual phone agents within reach of any business with existing AWS infrastructure, at a cost that makes the traditional call centre model increasingly hard to justify.
· 2 min read