|
In this newsletter:
- xAI launches new AI video model with lower costs and better quality
- Anthropic's Claude completes robotics tasks 20x faster than humans
- Adobe expands Firefly AI to build brand kits, videos, and storyboards
Plus, you’ll find new AI tools and this week’s top AI news headlines!
🎬 xAI Launches New AI Video Model With Lower Costs and Better Results
xAI launched Grok Imagine Video 1.5 to full availability, taking the top spot on an independent video quality leaderboard while pricing 86% below Sora 2 Pro's equivalent tier.
What's new:
- Available now on Imagine API, grok.com, and iOS and Android apps
- Tops the Image-to-Video Arena leaderboard, ranking above Sora 2, Veo 3.1, Seedance 2.0, and Kling
- Generates video clips up to 15 seconds at 480p or 720p with synchronized audio in the same pass
- Sound effects, ambience, dialogue, and lip-synced speech generated alongside video, no separate audio step needed
- Built on the Aurora engine, which generates each frame in sequence based on everything that came before it, for smoother camera movement and stable subjects
- 720p resolution ceiling, no 1080p option yet, though xAI says a higher-resolution mode is coming
Pricing:
- $4.20 per minute at 720p via API, compared to Sora 2 Pro's $30 per minute and Google Veo 3.1 Quality at $24 per minute
- A studio generating 100 minutes of video monthly saves roughly $2,580 compared to Sora 2 Pro
- Free tier available at grok.com with limited generation quotas
- SuperGrok subscription at $30 per month for higher limits
Speed:
- New Fast variant generates a 6-second 720p clip in about 25 seconds, down from 40-plus seconds previously
- Built for real-time and consumer-facing uses where speed matters more than maximum quality
New workflow tools rolling out:
- Projects organize related generations in a sidebar
- Multi-agent execution runs multiple prompts in parallel instead of one at a time
- Library search makes old generations easy to find
Limitations:
- Clips top out at 15 seconds, with longer videos built by chaining clips together, though quality drops after two or three extensions
- Frame rate fixed at 24fps, lower than the 60fps used in gaming and sports content
Why it matters:
xAI is stepping into the gap OpenAI left behind by killing Sora, undercutting Google on price while topping a major quality leaderboard at the same time. That combination is hard to ignore for any team currently paying premium rates for AI video, though the lower price comes with a real resolution tradeoff that rules Grok out for any project that needs full HD or broadcast-quality output.
👀 Read more about Grok Imagine Video 1.5 upgrades!
🐕 Anthropic's Claude Now Completes Robotics Tasks 20x Faster Than Humans
Anthropic tested whether Claude Opus 4.7 could operate a robotic quadruped on its own, finding it completed tasks up to 20 times faster than the fastest human team did just months earlier.
What's new:
- Claude Opus 4.7 operated the robot dog entirely on its own, with a researcher only plugging in a laptop, approving commands, and moving to the next task
- Completed every task at least 10 times faster than any human team from the original 2025 experiment
- On the four tasks completed by all groups, Opus 4.7 finished in under 10 minutes versus 181 minutes for the human team using Claude and 361 minutes for the team working without it
- Wrote roughly 10 times less code than the human team that used Claude, while matching or beating their results
- Tasks included connecting to the robot's camera and sensors, writing code to control its movement, and detecting a beach ball
Where it still struggled:
- Could not reliably move the robot to nudge a beach ball back to its starting point, the one task requiring constant real-time adjustment
- Defaulted to an outdated object detection method at one point, though it worked around the issue
- A more experienced human researcher was still able to complete the ball-retrieval task that Claude could not
Why it matters:
This builds on a pattern Anthropic has tracked before in cybersecurity: AI starts by helping people do hard technical work, then gets good enough that people barely need to help anymore.
Seeing that same shift happen with a physical robot suggests AI's usefulness is no longer limited to software, and the next wave of progress could show up in warehouses, factories, and anywhere else physical tools currently need a skilled human operator.
👀 Read more about Claude robotics operations upgrades!
🎨 Adobe's AI Assistant Now Handles Your Brand Kits, Videos, and Storyboards
Adobe expanded its Firefly AI assistant into Premiere, Illustrator, InDesign, and Frame io, adding the ability to create brand kits, product videos, and storyboards.
What's new:
- Available now in Premiere, Illustrator, InDesign, and Frame.io, already in Express, Photoshop, and Acrobat
- Works through ChatGPT, Claude, and Copilot, with Google Gemini and Slack support coming soon
- In Premiere, sorts assets into bins, batch-renames clips, identifies interview questions, and adds markers
- In Illustrator, it reorganizes layers across a document and checks for missing fonts
- Describe a brand and its style, or upload existing materials, to generate a brand kit with logos, identity, and color palettes
- Generate product videos directly from photos
- Create storyboards to plan out videos
New features in private beta:
- Elements saves AI-generated characters, objects, and locations so you can reuse them across projects
- Project stores existing assets in one place and shares context, useful for teams working on a video series or brand campaign
Why it matters:
Adobe's tools have always been hard to learn, and that difficulty kept people locked in once they learned it - nobody wants to relearn a whole new program.
AI assistants like this one remove that barrier by letting you just describe what you want instead of learning menus and shortcuts, which means the real test for Adobe isn't whether Canva can catch up anymore, but whether being "hard to learn" was the only thing keeping people from leaving in the first place.
👀 Read more about Firefly AI assistant updates!
My Latest LinkedIn & X/Twitter Posts:
- ChatGPT, Perplexity, Grok, Gemini: When to use each AI (view post)
- 3 levels of prompting to go from basic to expert-level (view post)
- 15 Perplexity prompts to supercharge your business (view post)
- 12 AI skills to learn to future-proof your career (view post)
- 10 prompts to replace expensive business consultants (view post)
In partnership with Kit:
As a creator, your time should be spent doing what you love, not juggling a dozen tools just to run your business.
Kit gives you everything you need in one place:
✅ Build and grow your email list (I use Kit for my newsletter)
✅ Easily monetize with paid newsletters and digital products
✅ Automate your emails with triggers and custom workflows
✅ Track what’s working and optimize with powerful insights
It’s not just email. It’s your entire creator business, simplified.
Join a thriving community of successful creators.
Use Kit for Free — and Start Building Smarter
New AI Tools to Boost Your Productivity:
- Teno: Creates campaign-ready image and video ads from a chat.
- Lynx: Automates business workflows with AI agents.
- GetMoreInterviews: Tailors resumes and cover letters for job applications.
- AuditAE: Tracks your brand's visibility across AI search engines.
- ChatGBot: Chats with multiple AI models in one app.
- VidBGRemover: Removes backgrounds from videos.
- Outify: Automates lead generation and sales outreach.
- AnyVids: Accesses multiple AI video models from one platform.
- 2PR: Grows LinkedIn with ideas, drafts, visuals, replies, and analytics.
- Denoisr: Removes noise from audio recordings.
- ClaviSay: Turns content into language-learning lessons.
- MAi: Creates branded Instagram grids from brand assets.
- Wery: Automates customer support with AI agents.
- FutureAGI: Tests and monitors AI agents for reliability and performance.
- Whacka: Builds working apps from plain-language descriptions.
- ProofWrite: Creates researched, cited, fact-checked SEO articles.
- CastReader: Turns articles, PDFs, and documents into audio.
This Week's Top AI News Headlines:
- Pinterest Launches Ask Pinterest, an Experimental AI Shopping App that Helps Users Discover Products Through Conversational Recommendations (View Article)
- Snap Moves its AI Video Technology into New Startup Dotmo After High Development Costs Make Independent Growth More Practical (View Article)
- Google Launches DATALAND, an AI-Powered Digital Art Museum Showcasing Cultural Collections Through Interactive and Generative Experiences (View Article)
- Anthropic Launches Claude Design to Help Teams Create On-Brand Content, Presentations, and Marketing Materials Using Company Guidelines (View Article)
- Character AI, the Platform for Creating and Chatting With AI Characters and Virtual Personalities, Launches Creator Bundle to Help Creators Build and Monetize Experiences (View Article)
- Perplexity Introduces Self-Improving Memory for AI Agents, Allowing them to Learn From Past Tasks, Refine their Behavior, and Improve Performance Over Time (View Article)
- Google Launches Ask Ad Manager, an AI Assistant that Helps Publishers Analyze Revenue, Troubleshoot Issues, and Manage Advertising Operations (View Article)
- Palmier, an Open-Source AI Video Editor Built for Claude, Lets Creators Edit, Search, and Generate Video Directly From the Timeline (View Article)
- OpenAI Launches Record & Replay for Codex to Turn Demonstrated Workplace Tasks into Reusable AI Automation Skills (View Article)
- Google DeepMind Outlines Security Framework for AI Agents to Protect Users From Prompt Injection, Data Theft, and Malicious Actions (View Article)
- Anthropic’s Project Fetch Shows Claude AI Agents Improving Web Research Performance Through Multi-Agent Collaboration and Self-Correction (View Article)
- Midjourney Launches Medical AI Initiative to Explore How Generative Image Models Can Support Healthcare, Medical Research, and Clinical Imaging (View Article)
- OpenAI Launches LifeSciBench, a Benchmark Designed to Measure How Well AI Models Can Solve Complex Biology and Life Sciences Research Problems (View Article)
- OpenAI Adds Spend Controls to ChatGPT Enterprise to Help Organizations Set Budgets and Manage Employee AI Usage Costs (View Article)
Work With Me:
If you enjoy this newsletter, please forward it to your friends and colleagues.
Follow me on LinkedIn and X/Twitter to see my latest posts.
Have a wonderful week!
Andrew Bolis
|