Skip to content
Tech News
clear
Topics: Today This Week This Month This Year

YouTuber builds road-legal Gasolini AR1 sports car for $7,000 using motorcycle engine

British fabricator Robert Ely built a single-seat, road-legal sports car called the Gasolini AR1 in his home shed for roughly $7,000. The car uses a stripped-down 2017 Honda CBR500 motorcycle engine producing 47 horsepower, Ford Ka wheels, and a fiberglass body designed without complex curves for easier fabrication. Ely says the lightweight design allows about 72 mph top speed and 41 mpg fuel economy.

Calendly unveils rebrand and adds AI email assistant, meeting notetaker

Calendly has rolled out a new logo, wordmark, color palette, and brand identity designed by studio Smith & Diction, moving away from its image as a simple scheduling calendar. Alongside the visual overhaul, the company announced new AI-powered tools, including an email assistant and a meeting notetaker, aimed at reducing scheduling friction and administrative busywork.

Identity Verification Gaps at Hiring and Account Recovery Emerge as Major Attack Vector

A joint alert from the US State Department, Japan, Canada and the UK warns that North Korean operatives are using falsified identity documents to get hired as remote IT workers at foreign companies, letting them enter corporate networks with legitimate-looking access. Separately, groups like Scattered Spider have used social engineering against help desks to trick staff into resetting passwords or MFA for accounts they don't own, bypassing strong authentication entirely.

Google Pixel 11 Pro units reported bricked after factory reset within days of launch

An Android Authority reviewer says his Pixel 11 Pro became unusable just five days after purchase, after a factory reset triggered a shutdown loop during the setup process. The device reportedly went blank mid-reset and later kept powering off every few seconds whenever the user tried to complete initial setup, leaving the phone effectively dead.

Ollobot pushes AI companion robots toward 'gentle intelligence' to fight loneliness

Ollobot argues that early companion robots from 2017 relied on simple voice commands and novelty, causing many to be abandoned once their appeal faded and some manufacturers shut down. The company is now framing a new generation of robots built with psychological research and clinical insight, aiming to shift from reactive gadgets to proactive, ever-present companions for isolated groups like elderly adults living alone and children separated from migrant parents.

OpenAI claims custom Jalapeño chip beats Nvidia GPUs on AI inference speed

OpenAI detailed benchmark results for its Jalapeño chip, an inference-focused ASIC built with Broadcom, claiming it outperforms Nvidia's GB200 and GB300 chips on efficiency and response speed. Using its InferenceX benchmark across models like GPT-OSS 120B, DeepSeek R1, and Kimi K2.5, OpenAI reported 1.5-1.9x more work per watt and up to 3.6x lower latency. Small-scale deployment is planned by year-end, with volume scaling into 2027.

Border wall construction set to cut through Texas ranchland archaeological sites

Federal contractors are moving forward with border wall and road construction across private ranchland in West Texas, threatening thousands of years' worth of Native American artifacts and rock art. While public pressure halted a planned steel bollard wall inside Big Bend National Park, DHS and Customs and Border Protection still intend to build barriers and roads through mountains and canyons on private land nearby. Ranchers like Jeff Fort, whose property hosts ongoing excavations led by archaeologist Bryon Schroeder, worry the construction will split their land and destroy sites that haven't even been fully studied yet.

Study finds infants outpace AI chatbots in language-learning efficiency

MIT Technology Review spoke with cognitive scientists comparing how babies and large language models acquire language, finding that infants learn to speak proficiently after hearing only 10 to 30 million words, while LLMs require vastly more data to approach similar fluency. Stanford's Michael C. Frank noted that training a model like GPT-2 on the same word count a toddler hears produces a 'nonsense generator,' not coherent speech.

Today's top topics: openai muse charm mark zuckerberg meta connect meta vr glasses ai agent ray-ban meta gen 3 ray-ban meta audio muse ai dolby vision
Browse all topics →  ·  Today's trending topics →