Taryn Plumb

One of the assumptions behind today's AI frameworks is that agents require a “boss” at the center; this orchestrator runs the show, routes requests, and makes sure the whole system doesn't descend into chaos.

New benchmark claims Microsoft Defender catches nearly all malicious email on its own, with integrated third-party tools improving detections by less than 1%, but security experts say these numbers may oversimplify the real risk.

By allowing lawsuits to proceed, a federal judge may be opening the door to broader liability for AI tool vendors, and greater governance obligations for enterprises using them.

The first proposed catalog of 'configuration smells' reveals widespread issues like context bloat, skill leakage, and conflicting instructions that can make coding agents less reliable and more expensive.

A new framework, Arbor, they claim, preserves hypotheses, experiments, and lessons learned across long-running research tasks, delivering 2.5x better performance than other models under the same budget.

If technology tends to minituarization over time, on-device AI is the future.

Researchers warn that repeated AI summarization, rewriting, and synthesis can gradually erode the original knowledge, context, and judgment that organizations depend on.

AI isn't coming for your job; it's actually here to team up with you so you can drop the busywork and focus on creative, big-picture decisions.

AI agents are great at coding, but they tend to wander off-track unless you use explicit specs, solid design documents and strict testing to guide them.

Shopify built an LLM proxy and distillation pipeline so its engineers keep working when any model goes away — and often get better results from smaller tuned models.