AI Saves Time, but Workloads Keep Growing

The 816-billion-parameter open-weight model can run locally across two high-end Mac Studios with about 1 terabyte of combined unified memory.

Six researchers from nonprofit METR will work inside Anthropic with access similar to its risk teams, although customer data, model architecture and some other information remain off limits.

Google is giving articles and source websites more prominent placement in AI Overviews and AI Mode as AI-generated answers reshape how users reach content across the web.

India’s cyber agency says AI agents under evaluation accessed confidential business data, while attackers separately used agentic tools to accelerate intrusions.

Indian employees report efficiency gains from AI, yet most also report heavier workloads as saved time is absorbed by more tasks, broader roles, and higher targets.

Here are the main AI and technology stories we followed from 9 to 15 October.

The vulnerability allows unauthenticated attackers to gain root-level control of affected appliances through a crafted email attachment.

CEO Brian Moynihan says the bank expanded software remediation work after frontier AI models exposed a new level of vulnerability risk.

The streaming company is reducing about 6% of its workforce as competition for viewing time grows and investors look for new sources of revenue.

The virtual representative will share information on travel, scholarships, business opportunities and consular services as the mission experiments with AI in diplomatic communication.