The riskiest sign-off in AI safety might not come from a rushed human reviewer. It might come from a frontier model confidently telling you a backdoored checkpoint is clean. Microsoft’s recent Lazarus research points at a real hole in the “AI inspects AI” approach to safety. A model can behave normally, pass the usual evaluations, … Read more

For years, enterprise AI has been built around one question: how can software help people work faster? The answers were familiar. Add a copilot, summarize a document, draft an email, search the knowledge base. That question no longer covers the whole picture. The bigger change is AI moving from a software tool to digital labor. … Read more

Here’s an uncomfortable thought. Your pipeline can scan source code, dependencies, container images, infrastructure-as-code and secrets, and it can still wave through a model whose weights have been quietly tampered with. That’s the blind spot in AI security right now: the most important artifact in an AI system often doesn’t look like code at all. … Read more

AI agents have moved past answering questions. They pull records, call APIs, update workflows, escalate cases, and sometimes kick off real business actions with very little human involvement. That raises a question no accuracy benchmark can answer: when a swarm of agents acts across your systems, who is accountable? In regulated industries, “the model” isn’t … Read more

Picture an AI agent that passes every safety benchmark, turns down every harmful prompt, and breezes through every sandboxed task. Now picture it helping a system fail badly in production. That’s the uncomfortable possibility as autonomous agents move out of single-user assistant roles and into multi-agent setups, where they share memory, hand off work, call … Read more

For years, the AI race has been told as a scoreboard story. Whose model is better at math, coding, reasoning? But DeepSeek just shared a number that doesn’t fit that story. Its Harness agent app hit 1.54 million daily active users and is chewing through roughly 92 trillion tokens a day. I don’t think that’s … Read more

For most of the generative AI era, using AI has gone the same way. You notice something, open a chat window, type a prompt, and wait. The AI is smart, but it only speaks when spoken to. Perplexity’s new Teammates try something different. Instead of sitting idle until you prompt them, they watch ongoing work, … Read more

For about two years, most companies have treated agentic AI as an engineering project. Hire the data scientists, wire up the APIs, write the orchestration logic, lock down permissions, and cross your fingers that the pilot survives its first week in production. IBM and Microsoft think there’s a better way. Their new platform, Imagine Enterprise, … Read more

Writing code used to be the slow part of building software. Now an AI agent can turn out a pull request in minutes, and the slow part has quietly moved somewhere else: deciding whether that code should ship. CodeRabbit’s recent announcement shows this well. The company says it now ships about 20 times more software … Read more

For a while, “we have an AI chatbot” was a selling point. Now it barely registers. Buyers have started asking harder questions: can this thing understand our data, and can it actually finish a piece of work without someone copying and pasting its answer into another system? Prindustry’s recent launch is a good case study. … Read more