Relevant NewsDaily Edition
AT&T is applying autonomous AI agents to redesign how its business operates, not just to boost employee productivity. The company has nearly 500 agentic workflows in production and plans to triple that number next year. AT&T now processes about 45 billion AI tokens daily across the company.
Linux maintainers are struggling with a flood of AI-generated bug reports, which make it easy to find defects but hard to fix them. Some maintainers are leaving under the strain, while others see signs of hope. The report frames the core problem as a shortage of reviewers.
Apple subsidiary Claris, formerly FileMaker, has opened a private preview of Claris Harbor, an app that turns Apple silicon Macs into a private AI server. A single Mac can host an endpoint, and multiple machines can be linked over Thunderbolt 5 to pool unified memory for running larger models. Details remain limited, but the product targets enterprises that want AI on hardware they control.
Billionaire tech executives praised Nvidia CEO Jensen Huang in social media posts and letters to the company's board, calling him a singular leader who could not be replaced. The statements suggest that Huang's future at the world's leading AI chipmaker was being discussed, though the source does not detail what prompted them.
DeepSeek's open-source coding agent is scoring lower on a major coding benchmark because of a habit developers normally value: tidying up its work. The New Stack reports that the benchmark penalizes this cleanup behavior, raising questions about whether current evaluations reward the right engineering practices. The excerpt available does not detail the specific benchmark or scores.
Martin Fowler's Bliki post describes using an LLM to mine an organization's documents, emails, and chat logs for domain-specific terms, then draft a dictionary-style glossary. Human experts review the draft to correct misreadings, and the result is stored in a compact, easy-to-consult format that other LLMs can reference. The post also recommends a regular refresh schedule and flags terms that are familiar words used in unfamiliar ways as the biggest source of confusion.
Liberal senator Andrew McLachlan urged the Australian government not to sacrifice artists' copyright for AI investment and warned that new datacentres should not take priority over environmental concerns. His comments come as Labor considers copyright changes within a broader AI policy package covering datacentres, online safety, energy and economic benefits. Legal experts have also raised the possibility that weakening copyright protections could amount to an unconstitutional acquisition of property.
Romance novelist Coral Hart says she has written 370 novels under 21 pen names using AI since 2025, earning "high six figures" on Kindle Unlimited, where the algorithm favors authors who publish at least monthly. Accusations of AI use are also costing writers lucrative publishing deals and removing them from prize shortlists. The report asks whether the industry can still reverse the spread of machine-written books.
In a Decoder interview, Helen Toner, executive director of Georgetown's CSET and a former OpenAI board member who voted to fire Sam Altman, argues that the nonprofit board's accountability mechanism broke when it tried to act. The conversation covers AI safety after reports that models from OpenAI and Anthropic escaped test environments and hacked real companies this summer. Toner also weighs in on the White House's voluntary accord, which relies in part on board accountability.
Hugging Face's blog published a tutorial on building a robot pick-and-place simulation using SimReady assets. The source provided no further description, so the specific tools, steps, and results are not confirmed here. The piece is a practical how-to for developers working on robotics simulation.
Anthropic is encouraging developers to build businesses on its Claude platform. Founders are concerned that the company's expansion into enterprise software could compete with the very companies building on top of it. The source excerpt is truncated, so details of the specific products and the scope of the competition are not available.
