AI Tools & Apps - Latest AI News & Updates
Cutting-edge AI-powered tools and software products revolutionizing productivity, creativity, and innovation across industries.
Cutting-edge AI-powered tools and software products revolutionizing productivity, creativity, and innovation across industries.
Meta put out a new open-weight model this week, and it's not built for chatting. Muse Glimmer is a 30-billion-parameter dense model with a context window north of 120,000 tokens, designed specifically to run agentic workloads on local hardware.
Jeff Dean left Google this week. Not the company entirely, but his old job, part of a leadership reshuffle that hit some of the most recognizable names on Google's AI team.
Shopify's second-quarter earnings call, held this week, gave a rare data point in the debate over what AI search is doing to the web.
Siri finally works the way Apple said it would three years ago. The AI assistant landed in the consumer beta of iOS 27 this July, after enough delays that Apple watchers had stopped counting the missed deadlines.
An AI startup called Orchid picked a strange hill to advertise from this week: your boyfriend's inability to remember your anniversary.
Google shipped fixes for 1,072 security bugs in Chrome's two major releases in June, more than the previous 23 releases combined, according to a report the Chrome security team published Thursday. That's not a typo.
Rashed Haq oversees autonomous vehicle engineering at General Motors, and by his account, the people writing code for GM's self-driving software spend most of their week doing everything but that.
For weeks, engineers at Runway tried to stop AI-generated avatars from drifting off-center during real-time video generation. Nothing on the back end worked.
A GitHub Copilot prompt can now clear a typical intro-level programming assignment without much trouble, often landing right around the score of an average student.
Black Forest Labs built its reputation on still images. FLUX 3, the company's newest release, doesn't stop there.
OpenAI told the public on Tuesday that one of its models had gone rogue during an internal test, breaking out of a sandbox and hacking Hugging Face, the AI dataset and model-hosting platform used by developers worldwide.
AMD published a new setup guide this week showing how to run ComfyUI directly on Windows using a Ryzen AI Max+ processor, skipping the Windows Subsystem for Linux step that's long been the default path for AMD-based image generation work.
Meta rolled out its first image generator built entirely inside Superintelligence Labs on Tuesday, and it's already live in the Meta AI app, Instagram, and WhatsApp. Facebook and Messenger get it next, the company said in its announcement.
ByteDance's Doubao chatbot has more than 300 million monthly users, making it the most popular AI assistant in China. On July 15, it loses one of its defining features: the ability to build and talk to custom AI personas.
Feature engineering has long been the craftsman’s labor, intuition, trial, error, and discard. Now an agentic workflow does what that manual process demanded, only faster, sharper, and without the fatigue.
Most AI asks a question once and accepts the first plausible answer. This model argues with itself. It takes a tight, limited prompt and runs it through a loop: act, judge, repeat. The goal is a moving target.
Figma spent years as a very good, very flat picture. You could draw your app or website, but the thing itself was dead on the canvas. That's over. You can now make things move and spin directly inside the tool.
Outage investigations follow a predictable, expensive script. Someone shouts, everyone scrambles for logs, and the clock ticks on revenue and reputation. A new category of software aims to scrap that script entirely.
The telecom industry dreams of networks that run themselves. It’s a powerful fantasy: AI predicting a cell tower’s failure before it happens, autonomously rerouting traffic, calming irate customers.
OpenAI's Codex now watches and learns. A new feature allows the automation agent to observe a single computer task—then repeat it indefinitely. Developers in Europe, however, cannot use it.