Sign Up | Advertise with Us

A big robotics push from NVIDIA headlines today, while Adobe rolls out AI that helps composites look real instead of fake. Micron lifts its forecast on relentless AI memory demand, Midjourney fires back at Disney and Universal in a closely watched copyright fight, GPT-5’s rollout gets mixed reviews, and Seoul’s Datumo raises new cash to take on model evaluation at scale. Here is the fast, clear version.

Imagine if ChatGPT and Zapier had a baby. That's {{Lindy}}. Build AI agents in minutes to automate workflows, save time, and grow your business.

Let Lindy's agents handle customer support, data entry, lead enrichment, appointment scheduling, and more while you focus on what matters most - growing your business.

Learn More About Lindy

NVIDIA announced new Omniverse libraries and “Cosmos” physical-AI models that let developers reconstruct real spaces, generate synthetic training data, and run physics-aware simulations for robots. Highlights include NuRec 3D Gaussian splatting, Isaac Sim 5.0 and Isaac Lab 2.2 on GitHub, and Cosmos Reason, a 7-billion-parameter vision-language reasoning model for planning and data curation. Early users range from Boston Dynamics and Figure AI to Amazon Devices and Foretellix, and the stack runs on RTX PRO Blackwell servers or DGX Cloud. The aim is faster, more realistic robot training that transfers better to the real world.

Adobe’s update adds Harmonize, which automatically matches lighting, color and shadows so pasted objects blend convincingly, plus improved Remove and a Generative Upscale that boosts images up to 8 MP. The company pairs these with Content Credentials that record edits for provenance, addressing authenticity concerns as powerful tools get easier to use. The features are in beta on desktop and web with early access on iOS, making complex compositing far more approachable for non-experts.

Micron raised guidance to about 11.2 billion dollars in revenue and 2.85 dollars adjusted EPS for the quarter, citing stronger pricing and demand for high-bandwidth memory used in AI data centers. The company also nudged its gross margin view to roughly 44.5 percent and saw shares rise about 5 percent pre-market after the update. The forecast underscores how HBM supply and pricing remain tight as hyperscalers scale up GPU clusters.

In a formal response, Midjourney argues its image generator is protected by fair use and says Hollywood studios and vendors have themselves used generative tools, calling it inconsistent to benefit from AI while trying to restrict it. The case centers on whether outputs that resemble famous characters constitute infringement and could set important precedents for training data and filters. The filing follows the June complaint that accused Midjourney of enabling unlicensed derivatives.

OpenAI’s new flagship model is rolling out broadly, but early users reported basic mistakes and slower responses in some tasks, prompting a wave of criticism alongside strong benchmark claims. Sam Altman told users the team is addressing issues and reiterated that longer-thinking methods and multi-agent search are core to future quality gains. The episode shows how public expectations have risen as labs promise step-function improvements.

Seoul-based Datumo, once a data-labeling startup, secured 15.5 million dollars led by Salesforce Ventures to expand no-code test generation and safety evaluation for enterprise AI. The company claims more than 300 clients in Korea, including Samsung and LG units, and is building out a Silicon Valley presence. Its platform automates checks for unsafe, biased or incorrect answers so non-engineers in policy or compliance can assess models before deployment.

Get access to all of my tools, guides, prompts and automations. Let AI and Automations do the work for you.

What you’ll get:

Unlock your unfair advantage

Stay tuned for more updates, and have a fantastic day!

Cheers,
Zephyr