🧠 OpenAI launches GPT-5.4

Good Morning! OpenAI just released GPT-5.4, a new model designed to take AI beyond chat, giving it the ability to control computers and complete tasks across apps automatically. Plus, I’ll show you how to create flawless AI face swaps and digital influencers.

Plus, in today’s AI newsletter:

• OpenAI Launches GPT-5.4 With Computer Control

• LTX-2.3 Brings Local AI Video Generation

• Anthropic Launches AI Job Impact Tracker

• How to Create Flawless AI Face Swaps and Digital Influencers

• 4 new AI tools worth trying

AI MODELS

🖥️ OpenAI Launches GPT-5.4 With Computer Control

OpenAI’s latest model combines stronger reasoning, coding, and professional productivity skills with native computer use, allowing AI to interact with software and perform multi-step tasks across your device.

• Native computer use lets the model issue keyboard and mouse commands and operate apps from screenshots

• Improved reasoning helps it search across multiple sources and synthesize answers more reliably

• 33% fewer false claims compared to GPT-5.2, making it OpenAI’s most factual model yet

• New GPT-5.4 Thinking mode in ChatGPT outlines its reasoning and allows users to adjust requests mid-response

GPT-5.4 marks a major step toward autonomous AI agents that can actually do work on your computer, not just answer questions. If widely adopted, this could shift AI from a chat interface to a true digital operator for everyday tasks.

TOGETHER WITH LIGHTRICKS

🎬 LTX-2.3 Brings Local AI Video Generation

Lightricks just launched LTX-2.3, a production-grade AI video model with built-in audio generation that runs entirely on consumer GPUs. Instead of relying on cloud GPUs like Sora or Runway, creators can now generate high-quality AI videos locally on their own machines.

• Runs locally on NVIDIA RTX 30/40/50 GPUs (as low as 8GB VRAM) and MacBooks

• Generates up to 4K video at 50 FPS, with clips up to 20 seconds and native portrait formats for TikTok/Reels

• Two modes: Fast for rapid iteration and Pro for final high-quality renders

• Fully open weights with free commercial use for companies under $10M revenue

Most AI video tools lock creators into cloud APIs, high GPU costs, and privacy trade-offs. By running locally, LTX-2.3 gives creators full control over their footage, removes ongoing compute fees, and opens the door for studios to fine-tune models on their own proprietary content.

AI IMPACT

📊 Anthropic Launches AI Job Impact Tracker

Anthropic introduced a new detection tool that measures how artificial intelligence is affecting different professions. The system compares what AI models can theoretically automate with what tasks large language models are already performing in the real world.

• Tracks AI exposure across jobs by combining task automation potential with real LLM usage data

• Finds computer programming (75%), customer service (70%), and data entry (67%) most exposed to AI disruption

• Physical and service jobs like cooks, mechanics, bartenders, and lifeguards show near-zero AI exposure

• Early data shows no clear rise in unemployment yet, though hiring for some entry-level roles may be slowing

AI job disruption is widely debated but poorly measured. By tracking real-world AI usage across occupations, Anthropic’s tool could help policymakers and companies spot labor market shifts early instead of reacting after jobs disappear.

HOW TO AI

🗂️ How to Create Flawless AI Face Swaps and Digital Influencers

In this tutorial, you will learn how to create photorealistic face swaps and entirely new digital identities using Arcads. By combining Nano Banana Pro for image generation and Kling for motion control, you can rebuild videos from scratch to ensure your character stays sharp, seamlessly lit, and consistent without that uncanny, pasted-on look.

🧰 Who is This For

• Content creators wanting to scale their brand without constantly filming themselves

• Marketers building permanent, free-to-use digital influencers for social media ads

• Video editors looking for the most realistic, artifact-free face swap workflow available

• Anyone who wants to place themselves (or an AI avatar) directly into iconic movie scenes or viral trends

STEP 1: Generate Your Character Reference

Head over to the image section inside Arcads and select the Nano Banana Pro model, which provides the sharpest details for this workflow. Upload the very first frame of your target motion video (the video whose movements you want to copy) into the reference field. Next, type a prompt describing your new character (e.g., "a woman in her late 20s with dark curly hair").

Crucial Tip: You must end your prompt by telling the AI to "Keep the exact same posture, angle, and background." This locks in the environment so the face swap transitions cleanly.

STEP 2: Apply Motion with Kling

Traditional face swaps just paste a 2D face over a moving head, which ruins the lighting. From the Arcads dashboard, click "See More" and select Kling Motion Control. This engine actually rebuilds the entire video from scratch based on your character.

Upload your original moving video into the "Motion Reference" field, and upload the static AI character image you just generated into the "Character Reference" field. Set it to the highest quality and hit generate. Your new character will now perform the exact movements, expressions, and gestures from the original video.

STEP 3: Edit and Repurpose in One Workspace

Normally, you would have to download your new video and take it into CapCut or Premiere. Arcads lets you do this natively. Hover over your newly generated face-swap video and click the three dots.

From here, you can instantly click "Add Captions" to auto-transcribe the video for social media. You can also select "Take a Snapshot" to grab a perfectly crisp still frame to use as a thumbnail, or remove the background completely, all without ever leaving the platform or losing pixel quality.

STEP 4: Scale with Viral Digital Influencers

To completely remove your own time and acting from the equation, you can create a character that has never existed. Generate an original face from a text prompt in Arcads. Then, grab the first frame of a trending or viral video (like a TikTok dance), and use Nano Banana Pro to place your newly created character into that scene's exact starting pose. Feed the full viral video into Kling Motion Control as the motion reference, and your custom character as the image reference.

You now have a recognizable digital influencer participating in viral trends, slowly building a permanent audience that you can use for future ads and scripts.

Google has quietly released an unsupported Workspace CLI, making it easier for agentic AI tools to access Gmail, Calendar, Drive, Docs, and other apps.

Dario Amodei says Anthropic plans to fight the DOD's risk designation in court, claims the DOD's letter has a “narrow scope”, and apologizes for the leaked memo.

Microsoft's new gaming CEO, Asha Sharma, teases the next-gen Xbox, codenamed Project Helix, saying it “will lead in performance and play your Xbox and PC games”.

Oracle is planning to cut thousands of jobs as soon as March, among its moves to handle a cash crunch from a massive AI data center expansion effort.

⚡ GPT 5.4: OpenAI’s new model designed to take AI beyond chat

📚 NotebookLM: New custom styles for infographics to turn info into visuals

🧠 GPT-5.3 Instant: The latest default ChatGPT model with fewer refusals and fewer hallucinations.

🎥 Replit Animation: Turn text prompts into animated videos

THAT’S IT FOR TODAY

Thanks for making it to the end! I put my heart into every email I send, I hope you are enjoying it. Let me know your thoughts so I can make the next one even better!

See you tomorrow :)

- Dr. Alvaro Cintas