👋 Hello hello,
Grok 4.6 landed this week and it's already undercutting rivals on price, which is either great news for your API bill or a small headache for OpenAI's pricing team. Claude finally figured out how to follow you across devices instead of forgetting you the second you close a tab. And Gemini decided it wants to book your dinner reservations too.
Everyone shipped something.
🚨🚨 Tomorrow we will be hosting our weekly live session where we will be building AI workflows from scratch, in real-time. If you haven’t attended any of these sessions, you are missing out! 🚨🚨
More details on this later. Now, let’s get into it.
🔥🔥🔥 Three Curated AI Updates
1.Grok 4.6 Lands, and It's Coming for the Price War
xAI released Grok 4.6 this week, its third major model release in under four months. Musk called it "smart, fast, and high value" in the announcement, and the benchmarks back up some of that. On the Artificial Analysis Intelligence Index, Grok 4.6 scores 61, tying GPT-5.6 Sol Max and jumping five points over Grok 4.5. It also leads on GDPval-AA v2 and AA-Briefcase, both built around real knowledge work rather than coding puzzles.
The bigger story is price. Independent reporting pegs Grok 4.6 as a 1.5-trillion-parameter model built on the same base as Grok 4.5, just with a longer, more curated training run stacked on top. That let xAI close the gap without a full rebuild, and it's using the savings to undercut rivals while running a 2x usage promo inside Cursor to pull developers over.
Claude's Chrome extension just got a memory upgrade. Sessions you start in the browser now carry over to desktop, web, and mobile, so a conversation you kick off mid-scroll doesn't die the moment you close the tab.
Your skills and connectors travel with you too, so whatever workflows you've already built work the same way inside Chrome as they do everywhere else.
Google used its Made by Google event to announce a fresh batch of connected apps for Gemini, rolling out over the coming weeks. The lineup spans categories: Granola, Otter.ai, and Wix for productivity and creativity; Fever, GetYourGuide, Localiza, OpenTable, and Ticketmaster for bookings and events; iHeartRadio and Pandora for music; and Angi, Thumbtack, and Zocdoc for home and health tasks.
The pitch is a single place to plan, create, and clear your to-do list instead of bouncing between a dozen apps. Worth watching if Gemini already handles your daily admin.
🔥🔥 Two Pro AI Tips To Bookmark Today
1.📊 Metricool + your LLM
Without performance data, your AI is just guessing at what content will work. The move here is pulling your operational data, things like social analytics or video transcripts, straight into Claude or ChatGPT so you can ask for specific, evidence-based script improvements instead of generic advice. Link scheduling platforms like Metricool to your LLM, review what actually performed well, and let the AI work from real numbers. Best for beginners who have content but no real feedback loop yet.
2. Microsoft Copilot + PowerPoint
If your company already pays for a Copilot license, this one's for anyone who freezes up at a blank slide. Point Copilot Chat at a doc that already has all your info and it'll build the whole presentation for you, questions, charts, and all. Best for anyone who needs a polished deck fast and would rather edit than start from zero.
🔥 One Claude Skill To Install
Days after Anthropic confirmed Claude now embeds invisible watermarks in everything it generates, someone already built a skill to counter it. It's not a magic eraser (more on that below), but it's live and open-source now.
Clone the repo: git clone https://github.com/guillaumemeyer/watermarks-remover.git
Create a skills folder and copy the remove-ai-marks skill into it: mkdir -p .claude/skills then cp -R watermarks-remover/skills/remove-ai-marks .claude/skills/
Point the rewrite hook at a non-Claude model, since rewriting Claude text with Claude defeats the purpose.
Invoke it with /remove-ai-marks, or just ask your agent to "strip AI watermarks" on the file or text in question.
Treat the results as best-effort. The tool can verify the Unicode and metadata cleanup, but it can't confirm the statistical rewrite actually beats Anthropic's private detector.
🎤 Join Us Live Tomorrow
Are you looking to automate your business with AI? We’re hosting live build sessions with Practicaly AI for Entrepreneurs members every Friday.
Not a member? Join now to get 20% off one month
Join us live, build in real time, and get your questions answered on the spot.
Can’t make it live? Don’t worry, we’ll send out the recording after the session to everyone who RSVP’ed.
Event details:
📅When: Friday at 10AM PT/1PM ET
📍Where: Virtual, Google Meet
After you RSVP, you’ll be sent a link with the event details. Hope to see you there!
Did you find today's post useful?
Until next time,
Team @PracticalyAI
