📸 GPT-5.2 is here + Disney’s $1B AI Bet

OpenAI drops GPT-5.2 + Disney's $1B bet, Rivian goes autonomous, and the "AI Slop" backlash begins.

December 14, 2025

My fellow AI explorers

Despite a theoretically calmer week, we’ve got a brand new model from OpenAI, a billion-dollar mouse-eared investment, and a collective internet revolt against "AI slop."

In today’s edition:

  • 🧠 GPT-5.2 is here: Math, science, and a massive context window

  • 🔓 Open Source Explosion: Mistral, Zhipu, and Qwen drop new models

  • 🍔 The "Slop" Backlash: Why everyone hated the McDonald's AI ad

Must See AI Tools

  • 💰 Payman: AI That Pays Humans. Over 10,000+ signed up for the beta

  • 💫 SubMagic: An AI tool that edits short-form content for you! (Get 10% off using code “uncoverai” at checkout)

  • 🎤 11Labs: #1 AI voice generator (Click Here to get 10,000 free credits upon signing up!)

  • 🤖 ManyChat: Automate your responses & conversations on IG, FB and more! (Click Here to get first month for free)

  • 🎙️ Syllaby: The only social media marketing tool you’ll ever need - powered by AI! (Get 25% off the first month or any annual plan with code “UNCOVER” at checkout)

OpenAI’s Big Week: GPT-5.2 & The Disney Deal

OpenAI surprised us (on a Thursday, no less) with the release of GPT-5.2, while simultaneously securing a massive IP partnership with Disney.

The previous model (5.1) was starting to feel a bit "dumb" to power users, but 5.2 seems to be the answer, specifically targeting STEM capabilities.

Here’s what you need to know:

  • Massive Context: The model boasts a 400,000 token context window with a staggering 128,000 max token output (that's roughly 96,000 words of generated text).

  • Coding & Science Gains: It scored 55.6% on Swebench Pro (up from 50.8% on GPT-5.1) and is reportedly blowing other models out of the water on science benchmarks.

  • Pricing: API costs are sitting at $1.75 per million input tokens and $14 per million output tokens.

The Disney Factor: Perhaps just as huge as the model release is the news that Disney is investing $1 billion into OpenAI. This is an IP play. Expect to see Disney characters legitimately integrated into Sora and image generation tools soon. We might be generating our own Moana shorts on Disney+ sooner than we think.

🧠 Prediction: The "Context Window Wars" are heating up again. With 128k output tokens, we are moving away from "chatting" with AI to having AI write entire novellas or codebases in a single shot. Also, the Disney deal signals the end of the "Wild West" of copyright—major IP holders are now becoming the platform partners.

Founders…

🤖 Need an AI Agency to Help Your Business Implement AI Solutions?

Our preferred partner Align AI provides you with an expert AI and Automation implementation team to add 10-40+ hours of increased productivity per employee and achieve your goals faster.

Open Source Intelligence

The Open Weights Ecosystem is Catching Up Fast

While OpenAI grabbed headlines, the open-source community had a massive week with releases from Mistral, Zhipu, and Alibaba.

The gap between proprietary models (like GPT-4/5) and models you can run on your own hardware is closing rapidly.

The Breakdown:

  • Mistral Devstral 2: The French AI lab released a new coding-specific model that rivals the best proprietary models. It scored 72.2% on benchmarks, putting it right next to DeepSeek V3. They also launched the Vibe CLI for local "vibe coding."

  • Zhipu GLM 4.6V: A new native tool-calling vision model. It comes in a 106B parameter size and a flashy 9B size, meaning you can run high-level vision AI on standard consumer GPUs.

  • Alibaba Qwen Updates: The Qwen 3 Omni Flash got a massive upgrade, focusing on indistinguishable-from-human voice interactions and improved multi-turn video understanding.

🧠 Prediction: 2025 will be the year of "Local AI." With models like Devstral 2 and GLM 4.6V, developers will stop paying API fees for sensitive data tasks and start running highly capable agents on their own secure clouds or even laptops.

AI SaaS Founders

🚨Want Millions of Impressions For Your AI SaaS, Done For You?

At uncovernews.co, we specialize in getting AI SaaS products the attention they deserve through strategic influencer marketing campaigns designed to drive millons of impressions at the fraction of the cost!

Get Your AI Startup’s News or Product In Front of Millions Quickly

The Culture Shock

A new AI-generated ad from McDonald's celebrating (or lamenting) Christmas has sparked a massive backlash, highlighting a growing consumer fatigue with low-effort AI content.

The Controversy:

  • The Ad: It featured AI-generated characters meeting misfortunes (falling out of windows, getting stuck in doors) with the uncanny, morphing physics typical of current video models.

  • The Reaction: The internet hated it. The sentiment is clear: Multi-billion dollar corporations have the budget for real actors and VFX artists.

  • The "Slop" Factor: Consumers are tired of questioning reality. When a brand like McDonald's uses AI that looks like the same "slop" floating around social media, it feels cheap and lazy rather than innovative.

Meanwhile in Privacy News: Meta has acquired Limitless, the company behind the "Pendant" that records your entire day. So, the company putting cameras on Ray-Bans now owns the tech to record every conversation you have for "memory recall."

🧠 Prediction: We are entering the "Uncanny Valley" of marketing. Brands will likely pull back on fully AI-generated commercials for a while to avoid the "cheap" label. Expect AI to move back to the background (VFX/Editing) rather than being the star of the show until the quality becomes indistinguishable from reality.

30-Second AI Play

OpenAI has rolled out a new connector that lets you use Adobe Express features without leaving the chat.

Here’s how to do it:

  1. Open ChatGPT and click on "Explore GPTs" or your settings to find Connect Apps.

  2. Add the Adobe Express connection.

  3. Upload a photo or ask ChatGPT to generate a design (e.g., "Create a holiday flyer for my business").

  4. You can then ask it to resize, change formats, or apply templates using Adobe's engine.

Why it matters: While early tests show it can't do everything (like complex background removal on uploads yet), it bridges the gap between text generation and professional design tools.

Other Relevant AI News!

🤝 OpenAI, Anthropic, and Block have formed the Agentic AI Foundation to create a unified standard for how AI agents talk to each other and interact with systems.

🛍️ OpenAI backtracks on "Ads": After users complained about ad-like suggestions, OpenAI clarified they were testing shopping features but have turned them off to improve precision and user control.

📸 Alibaba launches Qwen Image I2L, a wild new tool that lets you generate a style LoRA (a fine-tuned model adapter) from just a single image.

👁️ Zhipu AI releases GLM-4.6V, an open-source model that natively understands screen navigation and tool calling for better computer control.

🎬 New OpenAI Image Model Leaks? Rumors are swirling about "Chestnut" and "Hazelnut," potential new image models that can render perfect text and code inside images.

Golden Nuggets

  • 🧠 GPT-5.2 is a massive leap for STEM and coding, with a 128k output token limit.

  • 🤖 Rivian is betting the farm on vertical integration, building its own chips and voice assistants for cars.

  • 📉 AI Ad Fatigue is real—brands need to be careful not to look "cheap" by overusing generative video.

Would love to hear your thoughts on ChatGPT pulling back on ads! Send me your thoughts by replying to this email (yes, I read them all :)

Until our next AI rendezvous,

Anthony | Founder of Uncover AI