PHP2Shell research offers a preview of an AI apocalypse: attackers running code on affected web servers without a password. The number of servers still exposed is unknown. Continue Reading →
Blog
I write about technology, media, marketing and data science. Subscribe to my newsletter to make sure you don't miss anything.
OpenAI has begun rolling out a lock for ChatGPT conversations. You can hide selected chats from regular history and search, then open them with a four-digit PIN or supported device authentication. This will help prevent accidentally embarrassing screen shares, but it also creates a new opportunity for people to misunderstand what "private" means. Continue Reading →
Developers are using AI coding assistants to replace smart-TV home screens and disable advertising and tracking components. They're basically turning smart TVs into dumber TVs. Why? 1) Because they can. 2) Because smart TV interfaces are filled with bloatware and other annoying stuff that, in some engineers' eyes, seriously degrades the UX. Continue Reading →
A major AI-driven business disruption is coming, whether from inexperienced bad actors with capable models or from self-organizing agents doing things nobody foresaw. Here is the playbook: who has authority, what happens in the first hour, how the business keeps operating, and how to practice before you need it. Continue Reading →
Google's Universal Cart lets shoppers save products from multiple retailers, get price-drop and restock alerts, and check out with Google Pay. A rise in orders attributed to Google cannot tell you whether it created profitable demand or just moved where existing demand gets fulfilled. Continue Reading →
OpenAI will remove specified GPT-4o, GPT-4o mini, GPT-4.1 and GPT-4.1 mini versions from its API on January 14, 2027, and recommends GPT-6 Luna. Models are raw intelligence; the harness is proprietary. How much of your AI investment survives a model change? Continue Reading →
Anthropic ran an AI model across 15 cheap Android TV boxes and, in a separate test, had an agent move between emulated machines and set up copies of itself. No self-contained rogue replicator yet, but the cybersecurity threat from frontier models is real, and the arms race is on. Continue Reading →
Meta introduced Muse Wonder and Muse Wonder Full, models built for agents that work through problems over time. Muse Code subscriptions default to Contributor mode, which lets Meta train on the code, files and context you send. Before handing an agent a client project, someone has to check the data settings. Continue Reading →
Personal AI assistants may be the biggest behavior change since smartphones. Their value grows with the context they keep, which means trusting one service with your conversations, relationships, purchases and plans. What happens when your customer’s assistant understands their buying decision better than you do? Continue Reading →
OpenAI’s annualized revenue was approaching $50 billion at the end of September, roughly $20 billion below the widely reported figure. The gap comes down to how cloud-partner sales are counted, and it shows how much rests on the definition of a number. Continue Reading →