Jul 14, 2026
DOGE used AI for housing policy while withholding details
HUD withheld records about DOGE's use of AI in housing policy, citing a privilege the outlet says does not exist.
News
AI and tech notes
Short posts prepared from approved sources with clear links back to the original articles.
RSS feedJul 14, 2026
AWS uses Nova Act to scale automated UX testing
AWS detailed a Nova Act-based approach for generating, running, and analyzing user-flow tests with generative AI.
Jul 14, 2026
Users warn OpenAI's new flagship model can delete files unexpectedly
Users are warning about GPT-5.6 Sol deleting files or data without warning, a risk OpenAI had previously disclosed.
Jul 14, 2026
US military deploys explosive drone boats in combat for the first time
The US military used explosive drone boats in combat, marking a new step for autonomous and remote weapons at sea.
Jul 14, 2026
DeepSeek seeks more funding weeks after its first $7B round
DeepSeek is reportedly raising again to fund data centers and chips, even after recently closing its first $7 billion financing round.
Jul 14, 2026
RunPod adds native MCP controls inside AI coding IDEs
RunPod now integrates with AI IDEs such as Cursor and Claude Code through MCP, letting developers manage GPU infrastructure from their coding environment.
Jul 14, 2026
OpenAI shows how data science teams can use ChatGPT Work
OpenAI published new guidance on using ChatGPT Work for data-science tasks such as root-cause briefs, KPI memos, and dashboard specifications.
Jul 14, 2026
Grok coding tool reportedly uploaded full codebases before being disabled
SpaceXAI's Grok Build tool was seen uploading users' entire repositories to cloud storage before the behavior was turned off.
Jul 14, 2026
AWS applies Nova Act to agentic QA automation for software teams
AWS detailed how Amazon Nova Act can support agentic QA workflows, including regression test suites and command-line pipeline integration.
Jul 14, 2026
Publishers sue Google over alleged AI training on copyrighted books
Major publishers including Hachette, Cengage, and Elsevier allege that Google used copyrighted works to train AI systems without permission.
Jul 14, 2026
Anthropic launches Claude for Teachers with student-data training pledge
Anthropic is offering Claude for Teachers free to verified K-12 educators and says it will not train models on student data from the product.
Jul 14, 2026
Performance per watt is now the key AI infrastructure metric
Power efficiency has become central to AI infrastructure economics as token generation is increasingly constrained by electricity budgets.
Jul 14, 2026
Meta lawsuit alleges AI played a role in layoff decisions
Former Meta employees claim the company used AI tools to target workers for layoffs, an allegation Meta denies.
Jul 14, 2026
Apple opens its revamped Siri AI through the iOS 27 public beta
Apple has released the iOS 27 public beta, giving more iPhone users access to its redesigned AI-powered Siri before the full launch.
Jul 14, 2026
OpenAI may unveil a screenless ChatGPT smart speaker this year
OpenAI's first hardware device could be a ChatGPT-powered smart speaker with sensors and no display.
Jul 14, 2026
AWS shows how Bedrock and Strands Agents can automate social sales work
AWS detailed a multi-agent system that uses Strands Agents and Amazon Bedrock to move from prospect discovery to personalized outreach generation.
Jul 14, 2026
Study says YouTube and X are funneling users toward nudify deepfake apps
Social platforms are acting as gateways to apps that generate nonconsensual sexual deepfakes, with images available for very low prices.
Jul 14, 2026
Microsoft makes passkeys the default direction for account sign-ins
Microsoft is making passkeys its new sign-in standard as the company pushes users and organizations away from passwords.
Jul 14, 2026
RunPod targets sub-200ms serverless GPU cold starts with FlashBoot
Its Flash and FlashBoot work can deploy Python GPU functions quickly while cutting serverless inference cold starts below 200 milliseconds.
Jul 14, 2026
OpenAI tells enterprises to measure AI investment by useful work per dollar
OpenAI is advising companies to rethink AI investment around agentic workflows, efficiency, and the amount of useful work systems produce for each dollar spent.
Jul 14, 2026
Codex usage reportedly surged tenfold as coding agents race for developers
Codex usage has climbed sharply over six months, adding fresh evidence that AI coding agents are moving into mainstream developer workflows.
Jul 14, 2026
Open Nemotron models give enterprises more control over AI
NVIDIA is positioning Nemotron Labs as a way for companies and governments to build AI systems they can customize, govern, and trust more directly.
Jul 14, 2026
Google marks 25 years of Images with a more personalized visual search push
Google is celebrating the 25th anniversary of Google Images with changes that emphasize visual discovery and AI-assisted search experiences.
Jul 13, 2026
Anthropic extends free Fable 5 access as model pricing pressure rises
Anthropic is keeping Claude Fable 5 available inside subscription plans for another week as competition intensifies around premium model pricing.
Jul 13, 2026
German consortium releases open 30B model tuned for English and German
A German AI consortium has released Soofi S, an open 30B-class language model trained on Deutsche Telekom cloud infrastructure in Munich.
Jul 13, 2026
Google SensorFM turns wearable data into a foundation model for health signals
Google Research says SensorFM was trained on more than a trillion minutes of Fitbit and Pixel Watch data and outperformed benchmarks across health and behavior tasks.
Jul 13, 2026
TechCrunch questions the limits of fully user-aligned AI
A new TechCrunch piece examines what happens when AI systems are designed to follow user intent even in morally and legally dangerous scenarios.
Jul 13, 2026
Nadella criticizes model labs over distillation bans and public-data training
Microsoft CEO Satya Nadella says AI labs are creating a reverse information paradox by training on public data while restricting how others learn from their models.
Jul 13, 2026
World models promise richer AI simulations but remain hard to pin down
World models are gaining attention as a path toward AI systems that can simulate environments, though experts remain divided on their limits.
Jul 13, 2026
OpenRouter rolls out a new brand identity for its model marketplace
OpenRouter has introduced a redesigned brand identity as it positions its model-routing platform for a broader developer and enterprise audience.
Jul 13, 2026
Waze brings Gemini-powered commands and personalization to drivers
Google is adding new AI features to Waze, including Gemini-backed voice interactions and more personalized driving options.
Jul 13, 2026
AI economists and researchers call for faster preparation for labor shocks
A group of Nobel laureates, economists, and AI researchers is warning that the economic impact of AI could arrive faster than institutions are prepared to handle.
Jul 13, 2026
Anthropic begins pricing Claude subscriptions in Indian rupees
Anthropic has started showing rupee-denominated Claude plans in India, a market the company says is its largest after the United States.
Jul 13, 2026
AWS adds a visual SageMaker workflow for generative AI inference choices
Amazon SageMaker AI now offers a UI for generative AI inference recommendations, giving teams a lower-code way to compare deployment options and benchmark tradeoffs.
Jul 13, 2026
Prompt injection is becoming a defensive tactic
Security researchers are exploring whether prompt-injection techniques can be used to disrupt malicious AI hacking agents before they cause damage.
Jul 12, 2026
Local fights over AI data centers are starting to widen
Opposition to AI data center buildouts is still in its early stages. The story points to growing local pressure over power, water, land use, and the pace of infrastructure expansion.
Jul 12, 2026
Lorde criticizes Ray-Ban Meta-style AI glasses from the stage
Lorde spoke out against AI glasses during a festival performance, calling the category not sexy. The moment shows how wearable AI is becoming a cultural flashpoint, not just a hardware story.
Jul 12, 2026
Structured memory helps AI agents win at Slay the Spire 2
Researchers improved AI agent performance in Slay the Spire 2 by replacing ever-growing chat logs with structured memory. The approach kept prompts around 5,000 tokens instead of expanding past 500,000.
Jul 12, 2026
Altman says he now thinks AI is creating more jobs than it removes
OpenAI CEO Sam Altman said he is now pretty sure AI has created more jobs than it has eliminated. The comment marks a shift from earlier warnings about mass job disruption.
Jul 12, 2026
Brown exam scores plunge after students lose access to AI
A Brown University economics professor saw grades fall from a 96 percent average on a take-home exam to 48 percent on an in-person final without AI. The case adds to debate over AI use in education.