Waymo has urgently recalled nearly 4,000 autonomous taxis due to safety concerns stemming from the system’s limited ability to recognize complex construction zones

Waymo has urgently recalled nearly 4,000 autonomous taxis due to safety concerns stemming from the system’s limited ability to recognize complex construction zones

according to the latest disclosure by tech media outlet techcrunch, waymo, alphabet’s autonomous driving subsidiary, has launched a large-scale recall involving nearly 4,000 robotaxis. the recall stems from inaccuracies in its visual recognition and path-planning systems when operating in high-speed construction zones—specifically, the risk that vehicles may misinterpret closed road segments as passable areas. official confirmation indicates that at least 13 incidents of vehicles entering highway construction zones have occurred over the past few months, with seven cases in san francisco, four in phoenix, and the remaining two distributed across other operational cities.

this marks waymo’s sixth proactive recall of its autonomous vehicle fleet this year, as frequent safety concerns continue to raise serious alarms among regulators. documents filed with the u.s. national highway traffic safety administration (nhtsa) show that on may 19, waymo fully suspended all robotaxi operations on highways, limiting services to urban open-road passenger transport only. this move effectively freezes the commercialization of its high-value long-haul transportation use case.

coinciding with the recall is a rare joint special investigation by the nhtsa and the national transportation safety board (ntsb). the two agencies will focus on evaluating the robustness of waymo’s perception modules in recognizing temporary traffic signs, cone barriers, and dynamic construction boundaries, as well as the decision-making system’s response logic when faced with conflicting information from multiple data sources. the goal is to determine whether the issues arise from fundamental algorithmic design flaws rather than isolated data anomalies.

although urban‑based services remain operational, the absence of highway scenarios has significantly narrowed the service radius and reduced scheduling flexibility, further undermining public confidence in the maturity of l4‑level autonomous driving. waymo has pledged full cooperation with the investigation and to optimize relevant software modules, but it has not specified a timeline for resuming highway operations. industry observers generally believe that the cumulative effect of six recalls, coupled with extensive cross‑agency scrutiny, could accelerate the introduction of more stringent validation standards for autonomous driving systems in the united states, potentially raising the technical review threshold across the entire sector.

Waymo has urgently recalled nearly 4,000 autonomous taxis due to safety concerns stemming from the system’s limited ability to recognize complex construction zones

according to the latest disclosure by tech media outlet techcrunch, waymo, alphabet’s autonomous driving subsidiary, has launched a large-scale recall involvin

Waymo has urgently recalled nearly 4,000 autonomous taxis due to safety concerns stemming from the system’s limited ability to recognize complex construction zones

Jensen Huang emphasized that AI is becoming the infrastructure of the new era, and that society as a whole must proactively adapt and deeply integrate—just as humanity in the early 20th century readily embraced the automobile’s transformation of mobility

on june 17, nvidia founder and ceo jensen huang offered a comprehensive perspective on the evolution of an ai-driven society during an exclusive interview with

Jensen Huang emphasized that AI is becoming the infrastructure of the new era, and that society as a whole must proactively adapt and deeply integrate—just as humanity in the early 20th century readily embraced the automobile’s transformation of mobility

OpenAI has officially launched the scheduled task feature for ChatGPT and has now begun a phased rollout

today, openai officially launched chatgpt’s new “scheduled tasks” feature, initially rolling it out in phases to plus, pro, business, and enterprise users.

OpenAI has officially launched the scheduled task feature for ChatGPT and has now begun a phased rollout

Apple is facing cost pressures, and the iPhone lineup may see price increases; the rumored release date for the iPhone Air 2 has been delayed until 2027

the global ai race is continuously driving up the prices of storage chips and memory, with effects that have already extended beyond data centers to impact the

Apple is facing cost pressures, and the iPhone lineup may see price increases; the rumored release date for the iPhone Air 2 has been delayed until 2027

A tech blogger conducted real-world benchmark tests comparing the iPhone 11’s actual performance on iOS 16 and iOS 17

recently, the youtube tech channel idevicehelp conducted real-device testing to verify apple’s official claims about ios 27’s performance optimizations. they

A tech blogger conducted real-world benchmark tests comparing the iPhone 11’s actual performance on iOS 16 and iOS 17

Google has launched a new visual verification mechanism, requiring users to complete a two-handed waving motion captured by the camera to verify their identity

google is quietly advancing a new online identity verification mechanism based on gesture recognition, aiming to replace the conventional image classification

Google has launched a new visual verification mechanism, requiring users to complete a two-handed waving motion captured by the camera to verify their identity

The Honor 600 Smart is equipped with the largest battery in the series to date, delivering a significant leap in battery life

honor has officially launched a new member of its 600 series—the 600 smart 5g—highlighting ultra-long battery life and a high-refresh-rate display as its key

The Honor 600 Smart is equipped with the largest battery in the series to date, delivering a significant leap in battery life

Smartphones may bid farewell to LPDDR memory, transitioning entirely to LLW DRAM technology, which offers higher bandwidth and lower power consumption

the next generation of smartphones may usher in a revolutionary upgrade to memory architecture. industry sources indicate that mainstream manufacturers are acti

Smartphones may bid farewell to LPDDR memory, transitioning entirely to LLW DRAM technology, which offers higher bandwidth and lower power consumption

Haier has launched the all-new P7 Pro HQLED TV, featuring an AI smart engine, the Google TV operating system, and a 50W high-fidelity audio system

haier has officially launched its new hqled p7 pro series tvs, redefining the high-end home audiovisual experience with cutting-edge display technology and ai-

Haier has launched the all-new P7 Pro HQLED TV, featuring an AI smart engine, the Google TV operating system, and a 50W high-fidelity audio system

OnePlus has officially launched its new flagship tablet, the Pad 3 Pro, which comes equipped with a top-tier Qualcomm Snapdragon processor as its first offering

oneplus has officially launched its new flagship tablet—the oneplus pad 3 pro—redefining the high-end android tablet experience with class-leading performanc

OnePlus has officially launched its new flagship tablet, the Pad 3 Pro, which comes equipped with a top-tier Qualcomm Snapdragon processor as its first offering

The price hike for the iPhone 18 series is now a foregone conclusion Apple CEO Tim Cook openly acknowledged that the continued surge in DRAM and NAND flash memory prices is placing unprecedented pressure on the company’s supply chain and cost structure

apple may break with its longstanding pricing strategy, as the iphone 18 series is set to see a significant price hike. although the market had previously expec

The price hike for the iPhone 18 series is now a foregone conclusion Apple CEO Tim Cook openly acknowledged that the continued surge in DRAM and NAND flash memory prices is placing unprecedented pressure on the company’s supply chain and cost structure

OpenAI is secretly developing a new bidirectional speech interaction model codenamed GPT-Bidi-1, aiming to fundamentally redefine the limits of ChatGPT’s voice capabilities

according to the latest report from tech media testingcatalog, openai is accelerating development of a new bidirectional voice large model codenamed “gpt-bidi

OpenAI is secretly developing a new bidirectional speech interaction model codenamed GPT-Bidi-1, aiming to fundamentally redefine the limits of ChatGPT’s voice capabilities

According to reliable reports, Apple plans to launch a new iPhone Air 2 model in 2025 This device will feature a dual-camera system for the first time and be powered by an industry-leading 2-nanometer process chip, delivering a quantum leap in both perfor

according to the latest report from bloomberg, apple plans to launch the second-generation model of the iphone air series—iphone air 2—in spring 2027. the de

According to reliable reports, Apple plans to launch a new iPhone Air 2 model in 2025 This device will feature a dual-camera system for the first time and be powered by an industry-leading 2-nanometer process chip, delivering a quantum leap in both perfor

Google has officially unveiled Android 17, fully integrating native AI capabilities and enhancing the smart experience to rival the iOS ecosystem

on tuesday, u.s. eastern time, google officially released android 17 stable and the brand-new wear os 7 operating system, initially supporting pixel flagship d

Google has officially unveiled Android 17, fully integrating native AI capabilities and enhancing the smart experience to rival the iOS ecosystem

OpenRouter has officially launched the Fusion API—a smart inference service that enables dynamic collaboration among multiple models, balancing response speed, output quality, and invocation costs

recently, openrouter officially launched fusion api—a smart service architecture based on multi-model collaborative reasoning, marking a shift in ai model inv

OpenRouter has officially launched the Fusion API—a smart inference service that enables dynamic collaboration among multiple models, balancing response speed, output quality, and invocation costs