The afternoon briefing.
Google's AI Search faces immediate user issues, OpenAI expands its enterprise tools, and a key OpenAI co-founder joins Anthropic.
Google AI Search. Google's new AI Overviews are experiencing significant issues, with reports of the system failing to provide standard search results for certain queries. Users searching for terms like "disregard" have received generic chatbot responses instead of summaries, indicating a breakdown in the AI's ability to process information. This comes shortly after Google's I/O event heavily promoted AI integration. The immediate user impact highlights challenges of deploying large-scale AI features in critical services.
AI Enterprise & Talent. The enterprise AI landscape is seeing rapid movement, with Anthropic and OpenAI launching enterprise deployment arms and securing major financial services partnerships. This signals a strong push to monetize advanced AI models in corporate environments. OpenAI co-founder Andrej Karpathy joined Anthropic to lead its pre-training research, intensifying the competitive race for AI leadership. Chinese AI startup Deepseek is also reportedly prioritizing AGI research over quick profits, backed by billions in new funding.
AI Tooling & Policy. OpenAI continues to expand its ecosystem with new productivity tools, including a ChatGPT PowerPoint plugin and Appshots for Mac, allowing users to feed any app window content to Codex. These tools aim to integrate AI more deeply into daily workflows, though the PowerPoint plugin comes with a warning about potential content deletion. On the policy front, a planned executive order on AI safety testing was abruptly canceled by President Trump after top AI firm CEOs declined to attend.
Data Security & Autonomy. Data security remains a critical concern, as Trump Mobile exposed customer personal data, including home addresses and phone numbers, before its official launch. This incident underscores challenges in safeguarding sensitive user information. Google API keys were found usable for over 20 minutes after deletion, creating a significant security vulnerability. In autonomous tech, Waymo suspended freeway driving across all US markets due to construction zones and paused service in some cities due to flooded roads.
Google's AI search is so broken it can 'disregard' what you're looking for
Google's AI Overview section is reportedly failing for certain queries, returning generic chatbot responses instead of summaries. This issue, observed when searching for terms like "disregard," highlights significant problems with the new AI search integration.
OpenAI launches a ChatGPT Powerpoint plugin and warns it might accidentally delete your content
OpenAI has introduced a new ChatGPT plugin for PowerPoint, enabling users to create and edit presentations from notes or images. The company advises saving important decks before use, as the beta plugin carries a warning about potential accidental content deletion.
OpenAI Appshots turn any Mac window into context for Codex
OpenAI's new Codex feature, Appshots, allows Mac users to send the contents of any application window directly to the coding assistant. This provides Codex with immediate context for tasks, streamlining development workflows.
A founding member of OpenAI has joined Anthropic to boost Claude's research capabilities
Andrej Karpathy, a co-founder of OpenAI, has joined Anthropic's AI team to lead pre-training research for its Claude models. This move is expected to significantly enhance Anthropic's research capabilities in the competitive AI landscape.
What Anthropic and OpenAI launched in 72 hours has Wall Street paying attention
Anthropic and OpenAI both launched enterprise deployment arms and announced major financial services partnerships within a 72-hour period. These strategic moves are attracting significant attention from Wall Street, indicating a growing focus on enterprise AI solutions.
Deepseek reportedly prioritizes AGI research over quick profits despite billions in funding
Chinese AI startup Deepseek is reportedly set to raise around $10 billion, valuing the company at approximately $45 billion. Founder Liang Wenfeng has indicated a strategy to prioritize AGI research over immediate profitability.
Trump abruptly cancels EO signing event after top AI firm CEOs declined to go
President Trump canceled an executive order signing event focused on AI safety testing after several top AI firm CEOs declined to attend. The administration cited concerns that the proposed order could act as an innovation "blocker."
Cloudflare CEO Prince says builders and sellers are safe but AI is coming for the measurers
Cloudflare CEO Matthew Prince stated that AI is replacing middle management and compliance roles, leading to over 20% workforce reduction despite record revenue. Critics suggest the layoffs are more a classic efficiency program than direct AI replacement.
Trump Mobile has exposed customers' personal data, including home addresses and phone numbers
Trump Mobile has admitted to a data leak that exposed customers' personal information, including home addresses and phone numbers. The incident occurred before the phone's official release, raising significant privacy concerns.
Experts find Google API keys are still usable, even after you delete them
Security experts have discovered that some Google API keys remain usable for more than 20 minutes after being deleted. This creates a significant security vulnerability, as users cannot quickly revoke access or confirm when keys cease functioning.
Meta’s Forum is part Reddit, part Facebook, and part Google AI Overview
Meta has launched Forum, a new standalone app for iPhones that integrates Facebook Groups with an AI chatbot, similar to Google's AI Overview. The app allows users to view, search, and post advice within their groups, with optional AI assistance.
Waymo suspends freeway driving amid safety concerns
Waymo has suspended freeway driving across all its US markets due to concerns over construction zones and paused service in Atlanta and San Antonio due to flooded roads. This decision follows incidents involving its robotaxis.
Flipper Zero creators reveal massive new Flipper One swiss-army knife PC
The creators of Flipper Zero have unveiled Flipper One, a new "swiss-army knife PC" offering a Linux kernel, an open-source platform, and greater modularity. They are seeking community assistance for its development.
Intel Wants to Make the Ultimate Laptop Chip, and AMD Should Be Worried
Intel is reportedly aiming to develop the "ultimate laptop chip," leveraging lessons from its recent chip designs to challenge AMD's market position. This strategic focus indicates intensified competition in the laptop processor segment.