Claude Cowork is rolling out on mobile and web, which means you can start work on one device, let Claude keep going in the background, and jump back in when it needs your input. A few things that stood out to me: - AI agents moving beyond coding - More than 90% of Cowork usage not software development - Big use cases in business ops and content creation - Tasks like client briefs, contract trackers, variance memos, and decks - Background work that continues after you close your laptop - Scheduled tasks that run without a device online - Human approval still required before anything ships - Mobile notifications when Claude needs your decision - A bigger shift from asking AI questions to delegating real work Read the full article here: https://claude.com/blog/cowork-web-mobile

Ideogram 4.0 just dropped, and it's their first ever open weights release. This is HUGE!! If you do any image work in your builds (thumbnails, ad creative, lead magnets, social graphics), pay attention to this one. What's actually new: - First Ideogram model you can download and run yourself. Every version before this was closed API only. - Native 2K output with text that renders correctly, baked in. No patching it in after the fact. - Ranked #1 among all open weight models on the DesignArena leaderboard. The only models scoring higher are closed ones from OpenAI and Google. - Don't want to self host? Run it through fal.ai. Pricing starts around $0.03 per megapixel in Turbo mode, and anything you generate through the API is yours to use commercially. The text rendering is the real story. That's the thing a lot of image models still botch, and it's exactly what matters for design and ad work. Hugging Face download links, API pricing, and more: https://docs.google.com/document/d/1vveMIlHaM12fNERyfSzAtvQiyFSfG4cYvGqgXgxtWlk/edit?usp=sharing

Looks like Google is testing a new Gemini video generation tool called Omni, and it might be a step toward a single model that can handle both images and video. - Omni showing up in Gemini UI text, not just hidden code, so it could be a real product name soon - Screenshot line: “Powered by Omni” inside the video generation tab, which is a pretty loud hint - Today’s setup is split: Veo 3.1 for video, Nano Banana 2 and Nano Banana Pro for images - Omni might be a wrapper on top of Veo, a brand new Gemini video model, or the start of a true omni-model that outputs both images and video - If Google unifies image plus video generation, automation workflows get way cleaner, one API, one set of prompts, fewer handoffs - Competitive pressure is real in AI video right now, with ByteDance’s Seedance 2.0 leading some benchmarks - Google I/O 2026 (May 19–20) looks like the most likely stage for an official reveal If Omni ends up being one model for both image and video output, what would you guys automate first, ad creatives, product demos, or short-form content pipelines? Read the full article here: https://www.testingcatalog.com/google-is-testing-new-omni-model-for-video-generation-ahead-of-i-o/

Last week I got a new lead. I made their demo in less than 5 minutes (thanks to Trillet), had the sales call and closed the deal in less than 48 hours! Auto Repair Shop Closed at $159/mo - Handles FAQs - Sends booking link - Handles call transfers - 150 complementary call minutes each month - Website chatbot Very basic setup, no integration with their calendar, no integration with their CRM. Just my starter plan essentially. I have a new video up on my YouTube: https://youtu.be/wPOQG5h8_3s?si=NM62C473uJJIOTcx

This is an overlooked goldmine where local businesses actively show up looking for help, and almost no AI experts are there. Ming Xu, CIO of Trillet, explains how he positions himself as the only AI service in the room and turns simple conversations into paying clients. If you’re trying to land clients without competing online, this changes the game.