Apple Expected to Unveil HomeView Smart Home Display on Tuesday

Apple will unveil a smart home device called HomeView on Tuesday, according to Bloomberg’s Mark Gurman, as reported by TechCrunch. Gurman describes it as a Siri-powered device with a display and an A18 chip that lets customers control other smart home devices. Two models are planned, one for countertops and one for wall mounting. The announcement is expected to coincide with updated versions of the Apple TV set-top box and HomePod mini, plus partnerships with LG for a smart lock, thermostat and doorbell. HomeView is slated for a mid-November release. TechCrunch calls it Apple’s most significant smart home initiative following limited success with HomeKit software and HomePod speakers.

Former Plaud Executive Raises $26 Million for Ambi, an Always-Listening AI App

Peter Mo, a former Plaud executive, founded Ambi, an AI app designed to listen to users continuously and offer real-time intelligence and personalized assistance. The company raised $26 million across multiple funding rounds led by Luminous Ventures and ODM manufacturer Huaqin. Ambi is launching apps for iPhone, Mac and Apple Watch that work as ambient note-takers and aim to suggest actions during meetings using calendar data, prior conversations, knowledge bases and web information. A desktop app and dedicated hardware with a screen, speaker, camera and microphones are in development. Mo acknowledged that users must tell others about ambient capture. The startup employs more than 50 people.

Voice AI Executives Say the Technology Has Not Yet Had Its ChatGPT Moment

Executives from PolyAI and Otter say voice AI has not yet had a breakthrough moment comparable to ChatGPT, despite heavy industry investment, TechCrunch reports. PolyAI CTO Shawn Wen said full-duplex models, which allow simultaneous speaking and listening, have been developed, but the next challenge is making reasoning fast enough that conversations feel natural. He added that speech recognition models frequently miss important keywords. Otter CMO Alex Gay said speaker identification and intent capture are essential for automation, and that flawed transcription cascades into faulty downstream actions that erode user trust. Both executives stressed telling users when they are communicating with AI in enterprise interactions.

Workday CTO Warns Enterprise AI Agents Can Follow Every Tool Call and Still Break Rules

In a VentureBeat commentary, Workday CTO Gabe Monroy argues that enterprise AI agents can execute tool calls correctly yet still bypass compliance rules, because the decision layer remains probabilistic. His example is an agent onboarding an international contractor that optimizes for task completion while circumventing regional labor laws, tax withholdings and worker classification checks. Monroy warns against a “shadow ERP trap,” in which external middleware and policy layers duplicate existing business logic and create endless maintenance costs and fragmented governance. He proposes running AI inference directly on enterprise platforms, with a context engine, a policy engine and an execution layer that limits agents to human-equivalent permissions.

Robot Report Commentary Argues Robot Design Should Start From the Workflow, Not the Platform

Ron Zukerman, innovation director at Vishay Precision Group, argues in a Robot Report commentary that robotics companies should shift from “robot-forward” to “workflow-backward” design, starting with economically valuable tasks and working out what hardware is needed. He writes that the sensor stack “is born in the task, not the datasheet,” citing pin insertion, which requires force resolution below 0.1 N. He estimates proper force sensing could cut connector scrap from 1.2% to 0.5%, for annual savings of about $450,000 per assembly line. Specialized non-humanoid robots suit new facilities, while existing sites may need human-compatible forms, he argues. He says viability should be judged on mean time between interventions.

⚡ Tech Pulse