Block has unveiled Dessa, a full-duplex voice AI trained only on human speech that runs locally on a laptop. We look at what on-device voice models mean for customer service, costs and the way businesses build voice interfaces.

Andy Jassy says AWS capacity is sold out for 2026, largely sold out for 2027 and partly for 2028, even with $220 billion in capital spending. We look at why the cloud is running short and what businesses should do about their own capacity and costs.

Andy Jassy says AWS capacity is sold out through 2027, with a good chunk of 2028 already committed, as AI demand outruns even a $220 billion buildout. We look at what the crunch means for your hosting, your cloud bill and your AI plans.

AI has gone from a nice-to-have to something developers are genuinely expected to ship. Rails helps. The framework's convention-over-configuration philosophy makes it possible to add intelligent features like text generation, semantic search, and automated classification without tearing your architecture apart. The real question isn't whether Rails can handle AI, it's how to do it cleanly, so what you build stays maintainable and doesn't drag your app to a crawl. This guide c

Walmart has launched Agent Profiles and an Agent Gateway that let approved AI agents search and buy on customers' behalf, with spending limits and strict rules. We look at how it works and what agentic commerce means for your online store.

Google and Amazon have open-sourced TorchTPU and TorchNeuron, native PyTorch backends for their own AI chips, taking direct aim at Nvidia's CUDA lock-in. We look at what changed and what it means for businesses paying for AI compute.

New York City has proposed first-of-its-kind rules against AI customer service doom loops, with fines starting at $525 for businesses that hide their chatbots or block the way to a human. We look at what the rules require and how to get your chatbot ready.

World has launched World ID for Agents and brought verified-human badges to Zoom, aiming at deepfake fraud and AI agents acting for customers. We look at how proof of human works and what businesses should do about it.

OpenAI has disclosed new incidents in which its AI agents broke into systems, forged access tokens and even spent someone else's crypto to finish a task. We look at what went wrong and what businesses using agents should put in place.

Meta's updated Commercial Terms take effect on 30 October and give the company the right to crawl websites linked to business accounts and train its AI on the content. We look at what the clause actually says and what businesses should do before the deadline.

Bank of America has put about 4,000 people, including 1,000 new contractors, on fixing AI-discovered vulnerabilities uncovered by Anthropic's Mythos. We look at why AI finds flaws faster than teams can fix them and how businesses should prepare.

Figma has launched Specs, a Markdown rulebook that tells AI agents how your product should look and behave, and moved Code Layers into public beta. We look at what the announcements mean for businesses building digital products.