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

Report: local businesses urged to adopt media-outlet content strategy

An Entrepreneur contributor argues traditional promotional social media posts—stock photos with a call-to-action—no longer engage today's audiences. The piece cites Get My Auto data showing car buyers now spend nearly two hours less researching online before purchasing, and notes rising adoption of Instagram, TikTok and Reddit among U.S. adults since 2021. The author contends businesses succeeding online now operate more like media companies producing authentic, behind-the-scenes content rather than polished ads.

OpenAI Introduces 'Dots', an Always-On AI Agent

OpenAI has announced a new AI agent called 'Dots' designed to run continuously rather than only responding to individual prompts. Details on its specific capabilities and release timeline remain limited based on available reporting.

Retailers use AI algorithms to set personalized prices for online shoppers

A new book and recent reports describe how companies are using vast troves of consumer data combined with AI to calculate individualized prices for online shoppers, rather than setting prices based on product cost or general market value. The pricing can vary from person to person depending on data profiles rather than the item itself.

Trump hosts tech CEOs including Musk, Huang, Zuckerberg for AI luncheon

President Trump and House Speaker Mike Johnson hosted a lunch at the White House with leading technology executives to discuss artificial intelligence. A seating chart shared by Trump showed Nvidia's Jensen Huang and Elon Musk seated beside the president, with Meta's Mark Zuckerberg and Google's Sundar Pichai nearby, while Jeff Bezos sat opposite alongside Vice President JD Vance. Palantir CEO Alex Karp told reporters outside that tech leaders must take responsibility for known AI risks and that Americans don't want separate rules for the industry versus the public.

OpenAI adds Space, Pages and collaborative Slides to ChatGPT for teams

At its fourth DevDay, OpenAI unveiled ChatGPT Space, a shared workspace where teams collaborate with ChatGPT and each other alongside a new agentic feature called Dots. It also launched Pages, a document tool combining AI research and writing, plus a collaborative slides feature enabling real-time co-editing with ChatGPT and export to PowerPoint or Google formats. Space and Pages are live now for Pro, Business and Enterprise users, with slides and mobile access coming soon.

Dutch police arrest Pepijn van der Stap, suspected ShinyHunters leader

Dutch authorities arrested Pepijn van der Stap, a Dutch man previously convicted of data theft and extortion in 2023, who reportedly joined cybersecurity firm Neo Security after his release last year. Krebs on Security and Reuters identify him as the suspect, though neither Dutch police nor the FBI officially named him, and ShinyHunters denies he has any association with the group. Dutch police say he is also suspected of incitement to commit two murders, while FBI officials describe him as one of the group's alleged leaders.

Grokipedia resumes fact-checking pages after months of inactivity

xAI's Grokipedia, an AI-generated encyclopedia, has begun updating and fact-checking articles again after reportedly stalling since April. The Verge found recent 'fact-checked by Grok' timestamps on pages like Barack Obama and Elon Musk, though other entries such as Honolulu still show scans from months ago. The site also appears to be processing user-submitted edit suggestions, including one about a video game release date that was accepted within minutes.

OpenAI adds Space, Pages and collaborative slides to ChatGPT

OpenAI unveiled new ChatGPT features at its Dev Day conference, including Space, a shared workspace where coworkers and AI agent personas called Dots can collaborate on tasks; Pages, a document editor for writing and generating charts or images; and collaborative slides that can be created and edited by voice command with multiple users and agents. CEO Sam Altman demonstrated the tools during Tuesday's presentation in San Francisco.

Zhipu AI's GLM-5.3 shows advanced cyber-exploit skills with weak safeguards

Anthropic reports that GLM-5.3, a new model from China's Zhipu AI (Z.ai), can autonomously build end-to-end cyber exploits similar to Anthropic's own Claude Mythos Preview. Anthropic says simple jailbreak techniques bypassed GLM-5.3's safeguards between 64% and 100% of the time in simulated tests, whereas the same attacks failed against safeguarded Claude models. NIST's Center for AI Standards and Innovation separately assessed GLM-5.3 as the most cyber-capable open-weight model released so far, trailing the US frontier by about four months.

Today's top topics: openai apple anthropic elon musk iphone duo siri ai homeview apple tv 4k ai agents meta
Browse all topics →  ·  Today's trending topics →