SpaceXAI takes on GPT-5.5

Welcome back to The Shift. Let’s get straight to what matters in AI today…

Today we have:

🚀 SpaceXAI Launches Grok 4.5

😍How to Remix Any Video Into a Cinematic Ad With Gemini Omni Flash

🎙️ OpenAI Launches GPT-Live For ChatGPT Voice

🔨Tools and Shifts you cannot miss

🚀 SpaceXAI Launches Grok 4.5

SpaceXAI unveiled Grok 4.5, its latest flagship model, bringing faster performance, stronger coding capabilities, and lower pricing as it pushes deeper into the frontier AI race.

The Shift:

Built For Engineering Work - Grok 4.5 was trained for coding, science, engineering, and agentic tasks, delivering benchmark results that place it alongside leading models like GPT-5.5 and Claude Opus 4.8.

Faster And More Efficient - The model runs at up to 80 tokens per second while claiming roughly four times the efficiency of Opus 4.8, aiming to deliver frontier performance with lower compute requirements.

Lower Frontier Pricing - SpaceXAI priced Grok 4.5 at $2 per million input tokens and $6 per million output tokens, significantly below Claude Opus 4.8's $5 and $25 pricing.

More Models On The Way - The company is offering Grok 4.5 free inside Cursor and Grok Build for now, while Elon Musk says an even larger flagship model is planned for release next month.

Competition among frontier AI models is increasingly shifting beyond raw capability. Speed, efficiency, and pricing are becoming just as important as benchmark scores, making deployment costs a bigger differentiator for developers and enterprises.

Together with Fin

Own AI deployment, grow your career

Making AI actually work day to day is becoming its own job.

On July 16, hear from three people doing it: Simone Santiago Broad (Yoco), Yelva Espinoza (Zumba Fitness), and Fin's Dave Lynch. They'll share what the role really looks like, how it came to exist, the skills worth hiring for, and the challenges they're tackling right now. Bring your questions, since the best moments happen live.

Register to save your spot.

Reserve my spot

😍How to Remix Any Video Into a Cinematic Ad With Gemini Omni Flash

Your existing footage is the starting point. Gemini Omni Flash reshapes everything on top of it.

Step 1: Open Higgsfield - Head to higgsfield.ai/gemini-omni-flash. Runs in the browser, nothing to install.

Step 2: Upload your input - Drop in a source video, reference image, or both. Omni accepts any combination of video, image, and text in one prompt.

Step 3: Describe the transformation - Type what you want changed. Swap the aesthetic, shift the mood, change the lighting, replace objects. The source video stays the foundation, Omni rewrites everything on top.

Step 4: Steer with a reference image - Drop in a reference image alongside your video. Omni applies its composition, palette, and character likeness across the entire clip automatically.

Step 5: Iterate and export - Each prompt builds on the last. Adjust, refine, and export as MP4 when it’s ready to run.

No timeline. No keyframes. Just describe what you want. You can try Gemini Omni Flash here.

🎙️ OpenAI Launches GPT-Live For ChatGPT Voice

OpenAI introduced GPT-Live, a new voice system for ChatGPT that enables continuous conversations while handing search, reasoning, and agentic work to GPT-5.5 behind the scenes.

The Shift:

Voice No Longer Waits - GPT-Live replaces turn-based conversations with full-duplex interaction, allowing ChatGPT to listen, speak, pause, interrupt, and continue conversations naturally instead of waiting for users to finish talking.

GPT-5.5 Handles Harder Work - When conversations require search, deeper reasoning, or agentic capabilities, GPT-Live delegates those tasks to GPT-5.5 while keeping the conversation flowing without noticeable interruptions.

Rolling Out Across ChatGPT - GPT-Live-1 is now the default voice model for Go, Plus, and Pro users, while free users receive GPT-Live-1 mini across iOS, Android, and ChatGPT.com.

More Platforms Coming Soon - Business, Enterprise, Edu, desktop, Temporary Chats, Work, Codex, and custom GPTs aren't supported at launch, while OpenAI says GPT-Live will arrive in the API soon.

OpenAI is moving voice beyond a simple speech interface. The new architecture separates conversation from reasoning, allowing ChatGPT to stay responsive while more capable models work behind the scenes, making voice interactions feel much closer to natural conversations.

Together with Kalshi

Turn Your Opinions Into Profit

Join millions of traders putting their knowledge to work on real-world events—from inflation to elections. Buy “Yes” or “No” shares and earn if you’re right.

No house. Peer-to-peer. Cash out anytime.

Get a free $10 to start. Claim it and start trading now.

Start Predicting Now

_{Trade responsibly.}

🔨AI Tools for the Shift

📰 LetterBucket – Helps creators and businesses send newsletters while maintaining full ownership of their audience and subscriber list.

🔌 Hostinger Connector – Connects your AI coding assistant directly to your infrastructure for deployment, management, and monitoring without leaving your editor.

🔍 Are You Found by AI? – Checks whether ChatGPT, Claude, Perplexity, and Google AI recommend your business and reveals what's limiting your AI visibility.

🎬 VidFlux – Transforms static photos into cinematic AI-generated videos with realistic motion and depth in seconds.

🛠️ SophMate – Assists WordPress and WooCommerce users by generating support replies, workflows, design updates, and marketing content with approval controls.

🚀Quick Shifts

🎥 Google Photos launched AI Video Remix, using Gemini Omni to transform personal videos with cinematic relighting, custom backgrounds, and artistic styles like watercolor and sketchbook for AI Plus, Pro, and Ultra subscribers.

💰 Lovable is reportedly raising $300M at a $13.2B valuation, doubling its value in seven months as enterprise demand for AI-powered vibe coding accelerates, with the startup reaching a $500M annual revenue run rate.

🛡️ Google’s SynthID watermark helped debunk a viral AI deepfake, with fact-checkers confirming a fake image of Senator Mitch McConnell using the invisible watermark embedded in AI-generated images from participating models like Gemini and OpenAI. 

🚀 MiniMax is reportedly planning a Q3 launch for a 2.7-trillion-parameter AI model, which would be over six times larger than its current flagship and surpass every existing open-weight model in parameter count.

📱 Udemy co-founder Eren Bali's startup Monogram emerged from stealth with a $40M seed, introducing an iOS app that replaces traditional text interfaces with AI-generated interactive UIs built instantly for each task.

🧩 Prompt of the Day

Choose the Right AI Chatbot for the Job

Most people pick AI tools based on hype instead of fit. The best chatbot depends on what you do, the tasks you perform, your budget, and how technical you are.

Use the right AI, not just the most popular one.

Paste the prompt: Drop this into ChatGPT, then fill in your work details.

Prompt to paste

Recommend the best AI chatbot or AI assistant for my needs based on my [Insert role or use case], [Insert task to accomplish], [Insert budget], and [Insert technical comfort]. Compare the most relevant options and recommend the single best choice, along with strong alternatives if appropriate. Explain why each tool fits my workflow, where it performs best, its limitations, and how I can get the most value from it. Include pricing recommendations and practical usage tips instead of generic feature lists.

🤳AI Shift of the Day

That’s all for today’s edition. See you tomorrow as we track down and get you all that matters in the daily AI Shift!

If you loved this edition, let us know how much:

Forward it to your pal to give them a daily dose of the shift so they can 👇