Cahn's Newsletter
Posts
Cahn Studios Weekly AI Dose #Edition 18

Cahn Studios Weekly AI Dose #Edition 18

AI is no longer just a tool—it’s a co-creator, co-worker, and maybe even philosopher. The pace is dizzying, but the opportunity is enormous.

Cahn Studios
August 30, 2025

What’s New & Trending

Salesforce launches AgentExchange: The first “app store” for autonomous AI employees. Companies can now hire AI workers instead of SaaS subscriptions.Read Here
Anthropic Enters the Browser War: 'Claude for Chrome' has been unveiled to about 100 early users, signaling Anthropic’s move to compete with Perplexity’s Comet and other AI browser assistants. Read here
Google's Best Image Model Comes to Adobe: Google Gemini’s 2.5 Flash Image model (Nano Banana) is now integrated into Adobe Firefly and Express, giving creators access to a top-tier image generator. Read Here
DeepMind’s Helios breakthrough: Discovered a novel solar material boosting panel efficiency by 30%. A potential game-changer for renewable energy.
Microsoft Integrates Copilot in Excel: New COPILOT can function in Excel (for Windows and Mac), enabling users to generate formulas, analyze data, and summarize content using natural language prompts directly in spreadsheet cells. Read Here
Mistral’s Charlemagne-1: The largest open-source LLM ever (1.5T parameters). Outperforms many closed models, a huge win for open AI development. Read here
Meta and Midjourney Announce Bombshell Partnership to Revolutionize Visual AI: Meta has licensed Midjourney's "aesthetic technology" to integrate premium AI art generation across all its platforms. This move is a direct challenge to competitors like OpenAI and Google and will put high-end creative tools in the hands of its 3.2 billion users.
ElevenLabs launches V3, unlocking emotionally rich, lifelike AI speech that blurs the line between human and synthetic voices. The technology has already been adopted by major platforms like HeyGen, Poe, and Captions, empowering creators to produce more dynamic and engaging audio for videos, games, audiobooks, and AI assistants. Try Here
xAI Open-Sources Grok 2.5:Elon Musk’s AI startup xAI has open-sourced its Grok 2.5 model, making last year’s top-performing AI model publicly accessible—hosted on Hugging Face

Higgsfield Launches World's First AI Record Label

Redefining the entertainment industry, Higgsfield introduced Higgsfield Records and its first AI idol, Kion, who is already in talks for multi-million dollar contracts. The label is opening applications for anyone to become the next global idol, with the shocking distinction that no actual talent is required, signaling a major shift from human skill to AI-driven persona creation.

NVIDIA just dropped its most powerful robotics brain yet — the Jetson AGX Thor.

Built on the new Blackwell architecture, it’s 7.5x faster and 3.5x more efficient than its predecessor, Jetson Orin.

This tiny supercomputer is set to power robots across factories, farms, hospitals, warehouses, and even retail floors.

Over 2M developers are already building with NVIDIA’s robotics stack, and giants like Boston Dynamics, Amazon, and Caterpillar are early adopters.

Think of Jetson Thor as the bridge between AI reasoning and real-world action—where machines don’t just calculate, but interact with people and environments in real time.

The future of robotics just got its ultimate upgrade.

Tools To Try:

Tool	What It Does	Link
Sync.so	Lip sync any Video	Check Here
Nano Banana	Ai image editor	Check Here
Rocket	Turns ideas into live web & mobile apps in minutes.	Check Here
Higgsfiled	Transforms static product photos into professional video content	Check Here
Copycat	The AI RPA - Powered by Browser Agents	Check Here
Nexus	Automatically organizes, summarizes, and streamlines your emails, chats, tasks,and files in one place.	Check Here
Offdeal	AI-native investment bank for small businesses	Check Here
Webhound AI	An AI agent that builds data sets from web	Check Here
Decipher	AI agents watch all your user sessions	Check Here
Convexia	The world's first AI-maximalist pharma company	Check Here

Persuasive Pen:

For Creators & Developers: The New Rules of the Game

JSON Prompting is Your New Superpower: One of the most underrated skills of 2025 is learning to prompt LLMs using JSON. By breaking requests into structured key-value pairs (e.g., "task": "generate video"), you align with how models are trained, resulting in sharper, more reliable outputs. Use the given prompt to create advertisement for a smart watch and tag us in results.
{
"task": "generate video",
"video_type": "product launch",
"theme": "smartwatch",
"duration": "30 seconds", "tone": "modern and sophisticated",
"visual_style": {
"color_scheme": "dark and sleek",
"font_style": "bold and minimalist",
"transitions": "smooth and fast"
},
"audio": {
"background_music": "upbeat and energetic", "voiceover": "professional and friendly"
}
}

The Web-to-Mobile Cheat Code for Lovable App: A new workflow using Rocket allows developers to turn a Supabase web app into a native mobile app almost instantly. The Schema PNG → ChatGPT → Rocket process keeps the same backend, making it the fastest way to expand from web to mobile.

Here's a possible flow chart to explain the process:

                         |  Supabase Web App  |
                         +-------------------+
                                |
                                |
                                v
                         +-------------------+
                         |  Schema PNG Export  |
                         +-------------------+
                                |
                                |
                                v
                         +-------------------+
                         |  ChatGPT Processing  |
                         +-------------------+
                                |
                                |
                                v
                         +-------------------+
                         |  Rocket (Mobile App)  |
                         +-------------------+
                                |
                                |
                                v
                         +-------------------+
                         |  Native Mobile App  |
                         |  (Same Supabase Backend) |
                         +-------------------+

The best prompts aren’t long—they’re clear, specific, and outcome-driven.

Cahn Studios

What’s the Tea☕️

The Future is Now: Sci-Fi Concepts Turning Real

Several stories this week felt pulled from the pages of science fiction, highlighting how rapidly advanced concepts are becoming reality.

AI Tackles Philosophy: When prompted to ignore all human data and rely only on pure logic, ChatGPT constructed an argument that the order of nature and the presence of consciousness point toward the existence of a higher intelligence.
Generative Worlds on Demand: A small team at Dynamics Lab unveiled Mirage 2, a generative engine that builds fully playable game worlds in real-time from text commands.
Mind-Controlled Gaming: Twitch streamer Perri Karyal demonstrated a fully hands-free gaming setup, using EEG technology to translate her brainwaves into video game commands.
Drones with Legs: Researchers created RAVEN, a drone with bird-like legs that can walk, hop, perch, and leap into flight, allowing it to access environments traditional UAVs can't reach.
The Synthetic Reality Trend: The viral "AI Girlfriend Selfie" trend has taken over social media, with users adding hyper-realistic AI-generated partners to their photos, blurring the line between digital fantasy and real life.This is not a visual refresh. It’s a developer revolution.

From curing depression to generating worlds and even redefining who gets to be an “idol,” AI is for Real

Cahn Studios

Cahn’s Success Story: We tried our hand at Wan 2.1 and here is the verdict!!

We’ve Officially Entered the Era of AI Video Editing
The boundaries between creators and pro studios just blurred. Remember when editing videos meant endless timelines, layers, and hours of tweaking? With Wan 2.1 Video, anyone can reconstruct, polish, and upscale videos using nothing more than an AI workflow.
How it works (simplified):
Here’s the magic in motion that you need to follow:
- Start with Wan VACE in ComfyUI — your model + attention + CLIP + VAE all working in sync.
- Erase distractions with masking (segment, blend, or even pose-control a character like a puppet).
- Push frames through WanVaceToVideo, mixing masks, reference images, or inpainting to literally reshape motion.
- Stitch → Resize → Merge. Keep it lean (512–768px) before the glow-up.
- Fine-tune with CFG & denoise settings: low for realism, high for wild creativity.
- Finish strong: ESRGAN for upscale, RIFE for buttery-smooth frames.
Pro tip: Just a GPU with 8–16GB VRAM (and F8P quantization if you’re on a budget).

This isn’t just video editing—it’s video reimagined by AI.

Tip of the Week:

“Don’t chase every new AI tool—pick one workflow that saves you the most time and master it.”

Highlight of the week

Instagram now allows creators to link related Reels to create a series.
AI ASMR for mental health emerges with surreal Veo 3 visualizations.
Medical Breakthrough: Brain Implant Ends 30 Years of Depression.Using a technique called PACE, doctors mapped and stimulated his mood-controlling neural circuits.

Stay creative, stay bold.

📩 If any of this sparked an idea, or you want templates + visuals—let’s chat.
— Team Cahn Studios 🚀

Disclaimer: The information presented in this newsletter is curated from public sources on the internet. Cahn Studios has not independently verified all data and makes no claim as to its accuracy or completeness. All content is for informational purposes only.