Welcome to today’s edition of AI News Orb
Meta has launched Muse Glimmer, Google Reportedly Cancels Gemini 3.5 Pro, Microsoft Plans Maia 300 AI Chip, Cut Onboarding Time in Half with Loom + ChatGPT
Latest AI News
Trending AI & SaaS tools
Trending AI & Technology News
AI Art Generation
Today’s Resource
Recommended AI Video
Latest AI News
Meta
Meta has launched Muse Glimmer
Meta has launched Muse Glimmer, a new open-weight AI model designed to handle smaller AI agent tasks on personal computers using a single graphics card.
The Details:
The launch comes as Meta CEO Mark Zuckerberg pushes for wider access to open-weight AI. He argues that open models can be cheaper, easier to customize, and less controlled by a small number of powerful companies.
Zuckerberg also called on the U.S. to reduce barriers for open AI development, saying American developers face stricter limits than competitors in China. His comments come as Chinese AI companies continue to make progress with open models.
Meta also announced plans to release Muse Spark 1.2, which the company describes as its most advanced AI model yet.
Alongside the AI announcements, Meta is creating a $1 billion fund to support communities affected by the expansion of its data centers.
The move highlights Meta’s growing focus on open-weight AI, as the global competition between U.S. and Chinese AI companies continues to intensify.
Google Reportedly Cancels Gemini 3.5 Pro
Google may be abandoning its long-delayed Gemini 3.5 Pro model and shifting its focus to Gemini 4, according to a new report from Geeky Gadgets.
The Details:
However, Google has not officially confirmed that Gemini 3.5 Pro has been canceled, so the claim should be treated as a report rather than a confirmed announcement.
Gemini 3.5 Pro has already faced several delays. In July, Google said the model was still being tested with partners and would arrive “soon,” while also confirming that training for Gemini 4 had started.
The reported change comes as Google reorganizes its AI division. Demis Hassabis has moved from his day-to-day leadership role at Google DeepMind to become its chair and Alphabet’s chief scientist, with Koray Kavukcuoglu taking operational leadership.
If the report is accurate, Google could be skipping Gemini 3.5 Pro entirely and putting more resources behind Gemini 4 as it competes in the fast-moving AI market.
Bottom line: Gemini 3.5 Pro's future remains uncertain, but Google is clearly preparing for its next major AI model.
Trending AI & SaaS tools
Unwrap AI helps businesses turn customer feedback from surveys, reviews, support tickets, and other channels into proactive, actionable insights using AI, without relying on manual tagging or complex dashboards.
xAI has launched Imagine Image 2.0, a new AI image generator with stronger prompt accuracy, sharper text, advanced editing, multi-image references, and smart resizing, now available in Grok’s Quality Mode.
ByteDance’s Seedance 2.5 is a next-generation AI video model that can create 30-second videos in one generation, with stronger reference control, editing, smoother motion, and professional production features.
IdeaBuddy is an AI-powered business planning platform that helps entrepreneurs turn ideas into business plans, financial forecasts, and validated business models with step-by-step guidance and AI assistance.
Microsoft
Microsoft Plans Maia 300 AI Chip
Microsoft is reportedly preparing to unveil its next-generation Maia 300 AI chip this fall, potentially as early as September 2026, according to The Information, cited by Reuters.
The Details:
The chip is part of Microsoft's effort to build more of its own AI hardware and reduce its heavy reliance on Nvidia's processors, which remain widely used for AI workloads.
Microsoft launched its first Maia chip in 2023 and introduced the Maia 200 in January 2026.
The company is now looking to take its custom AI-chip strategy further with Maia 300.
Reports also say Microsoft is discussing production with TSMC, with plans to manufacture more than 300,000 Maia 300 chips by 2027.
Microsoft could eventually seek capacity for more than 1 million chips, although supply and manufacturing limits could affect those plans.
Microsoft also reportedly wants major cloud customers, including Anthropic, to use its custom chips.
The move shows how major technology companies are increasingly developing their own AI processors to control costs, improve performance and compete more directly with Nvidia.
AI & Technology News
Moonshot AI is restructuring its offshore setup under tighter Beijing rules, a move that could delay its planned Hong Kong IPO until 2027 or later.
Higgsfield AI has debuted The Cully Hill Boys, a 110-minute feature made entirely with AI in just four weeks, showing how generative AI could reshape filmmaking and help creators monetize their likeness.
DeepSeek’s V4-Flash is outperforming its own V4-Pro preview on key coding and agent benchmarks while costing far less, making it a strong low-cost option for AI developers.
Alibaba’s Wan3.0 enters beta with 30-second AI video generation from mixed inputs such as text, images, video, audio, webpages, PDFs, and PowerPoint files.
AI Lesson
Cut Onboarding Time in Half with Loom + ChatGPT
Step-by-Step Guide:
Record your process in Loom
Record a short screen-share showing how you complete a common task. Explain each step as you work.Get the transcript
Copy the Loom transcript and paste it into ChatGPT.Create an onboarding guide
Prompt:❝“Turn this transcript into a simple step-by-step onboarding guide. Use clear headings, numbered steps, important tips, and common mistakes. Remove unnecessary details.”
Create a quick checklist
Prompt:❝“Convert this guide into a one-page checklist that a new employee can follow without help.”
Create an FAQ
Prompt:❝“Based only on this transcript, create 10 common questions and simple answers a new employee might need.”
Make it beginner-friendly
Prompt:❝“Rewrite this onboarding material for someone with zero prior knowledge. Keep the language simple and practical.”
Share everything
Give new hires the Loom video + ChatGPT-created guide + checklist + FAQ.They can learn visually and use the written resources when they need a quick reference.
AI Art Generation
Create a similar photo using the Prompt Below
Prompt:
Create a premium 3×2 luxury advertising collage for [PRODUCT NAME], designed as a high-end commercial campaign. Keep the product’s exact design, shape, color, branding, and key features consistent across all six panels. Use a [COLOR PALETTE] aesthetic with cinematic lighting, realistic textures, elegant reflections, dramatic shadows, and a sophisticated atmosphere.
Panel 1: Hero shot of [PRODUCT] on [SURFACE/PROP] with [LIGHTING EFFECT].
Panel 2: [PRODUCT] interacting with [ELEMENT: water/smoke/ice/flowers/etc.].
Panel 3: [PRODUCT] placed in [ENVIRONMENT] with [MOOD].
Panel 4: Detailed close-up highlighting [KEY FEATURE/MATERIAL].
Panel 5: [PRODUCT] with [PACKAGING/ACCESSORIES] in a luxury setting.
Panel 6: Creative lifestyle shot showing [PRODUCT FUNCTION/BENEFIT].
Photorealistic, ultra-detailed, premium product photography, cinematic composition, shallow depth of field, realistic reflections, professional studio lighting, clean framing, no people, no clutter, no unnecessary text, consistent product identity throughout.
Today’s Resource
10 Free GitHub Repos That Turn Claude + Obsidian Into a Real Second Brain
10 free GitHub repos use Claude with Obsidian to create smarter knowledge systems through wiki-based memory, Skills, and MCP connections.
For beginners, AgriciDaniel/claude-obsidian is recommended; power users can choose automation, semantic search, or read-only setups based on their needs.
