In partnership with

Welcome, AI enthusiasts

ChatGPT is facing backlash after announcing a new partnership with the DoD. U.S. uninstalls reportedly spiked 295% in a single day, while rival app Claude shot to the top of the App Store with downloads surging up to 88%. Users are making their feelings known with the only vote they have. Let’s dive in!

In today’s insights:

  • ChatGPT Uninstalls Spike 295% After DoD Deal

  • GPT-5.3 Makes ChatGPT More Direct and Reliable

  • Meta tests AI shopping assistant

Read time: 5 minutes

LATEST DEVELOPMENTS

Source: Appfigures Intelligence

Evolving AI: ChatGPT faces user backlash after DoD deal: uninstalls jumped and 1-star reviews spiking.

Key Points:

  • U.S. ChatGPT uninstalls jumped 295% in one day after its DoD partnership was announced.

  • Claude downloads surged up to 88% day-over-day, hitting No.1 in the U.S. App Store.

  • 1-star reviews for ChatGPT spiked 775% as users reacted publicly.

Details:

Market data from Sensor Tower, Appfigures and Similarweb shows a sharp consumer reaction after OpenAI confirmed a partnership with the U.S. Department of Defense. ChatGPT’s U.S. downloads fell 13% the day after the news and continued declining, while uninstall rates surged far above normal levels. At the same time, Anthropic’s Claude saw download growth between 37% and 88% day-over-day, briefly overtaking ChatGPT in daily U.S. installs and reaching No.1 in multiple countries.

Why It Matters:

This weekend showed how fast “AI ethics” turns into real distribution math. A single government contract can flip installs, rankings, and ratings in days, pushing users to a rival app and forcing public rewrites of contract language. For teams picking an AI assistant at work, this raises a new risk: model quality can stay the same, yet internal buy-in, PR exposure, and even employee sentiment can swing overnight. If your chatbot becomes a political symbol, can you still roll it out company-wide without friction?

Ship Docs Your Team Is Actually Proud Of

Mintlify helps you create fast, beautiful docs that developers actually enjoy using. Write in markdown, sync with your repo, and deploy in minutes. Built-in components handle search, navigation, API references, and interactive examples out of the box, so you can focus on clear content instead of custom infrastructure.

Automatic versioning, analytics, and AI powered search make it easy to scale as your product grows. Your docs stay accurate automatically with AI-powered workflows with every pull request.

Whether you're a dev, technical writer, part of devrel, and beyond, Mintlify fits into the way you already work and helps your documentation keep pace with your product.

Evolving AI: OpenAI rolls out GPT-5.3 Instant as ChatGPT default.

Key Points:

  • GPT-5.3 Instant cuts long safety preambles and gets straight to the answer.

  • Hallucination rates drop up to 30% in internal evaluations.

  • Rollout follows backlash over model swaps and Pentagon deal.

Details:

OpenAI has replaced ChatGPT’s default model with GPT-5.3 Instant, focused on tone and reliability rather than new features. The update trims lengthy disclaimers, reduces overly cautious refusals, and aims for more direct responses. Internal tests show hallucination rates falling by roughly 20 to 30%, with stronger factual performance both with and without web access. GPT-5.2 Instant remains available for paid users during a short transition period.

Why It Matters:

GPT-5.3 Instant is OpenAI trying to fix the part of ChatGPT that annoys people every day: long safety speeches, needless refusals, and answers that feel stiff. If the model really cuts hallucinations by the ~20–27% OpenAI claims, that’s fewer times you need to double-check sources, rewrite copy, or explain to a teammate why the bot confidently made something up. The timing isn’t subtle either: trust is already shaky after the GPT-4o retirement drama and the Pentagon deal backlash, so making the default experience feel more natural and more reliable is a direct attempt to stop churn and keep ChatGPT usable as a daily work tool, not a tool you argue with.

👀 Our tip

Anthropic just dropped a free AI learning platform that teaches you how to actually use Claude, not just talk about it. From “Claude 101” to hands-on API and developer tools training, it’s built for anyone who wants to build smarter with AI. If you’ve been meaning to level up your AI skills, this might be the easiest place to start.

Evolving AI: Meta pilots AI shopping on desktop.

Key Points:

  • Meta is testing an AI shopping tool for select US desktop users.

  • The chatbot shows product carousels with prices, links, and brand info.

  • Recommendations adapt to user data like gender and location.

Details:

Meta has begun testing a shopping feature inside Meta AI on desktop for select US users. A new “Shopping research” button appears in the prompt box. When asked for suggestions, the chatbot displays a carousel of products with images, prices, brand details, and links to external stores. It explains why each item is recommended and can factor in profile data such as gender and location. Checkout happens on retailer sites, not inside Meta.

Why It Matters:

Meta’s test is another sign that “where you shop” is sliding from search bars and social feeds into chat boxes. OpenAI already pushed “shopping research” in ChatGPT ahead of Black Friday, and Google has been stacking Gemini with shoppable listings and even instant checkout via big retail partners, so Meta can’t stay on the sidelines for long. For shoppers, it cuts the messy part: fewer tabs, faster shortlists, clear price anchors, one click out to a store. For brands, the fight shifts from winning ad auctions to being “picked” by the model, which puts product feeds, accurate pricing, shipping coverage, and brand signals front and center.

TOGETHER WITH ELEVENLABS
🤖 Great AI agents start with great prompts

Evolving AI: Learn how to build prompts for AI agents.

This new ElevenLabs guide walks developers through the frameworks, structure, and evaluation methods that make conversational agents reliable, secure, and context-aware.

Learn how to build and iterate on prompts that deliver real-world results, fast.

Download The Prompt Engineering Guide and start building smarter voice AI systems today.

👀 Click on the image you think is real

QUICK HITS

🔋 AI data centers could reduce power draw on demand, study says.

🚀 Gemini 3.1 Flash-Lite: Built for intelligence at scale.

🤖 Tilly Norwood, the fully AI 'actor,' to be part of rapidly expanding 'Tillyverse'.

🗣️ Sam Altman says OpenAI is renegotiating with the Pentagon after an ‘opportunistic and sloppy’ deal.

📈 Trending AI Tools

  • 📝 Granola - AI notetaker that captures the real insights and turns every conversation into ready-to-share, action-driving notes*

  • 🍌 Nano Banana Pro - Google’s AI image generator

  • 🗂️ Tana – A note-taking app built with AI that turns your thoughts into tasks you can track

  • 🎬 Seedance 2.0 - Bytedance’s latest viral video model

 *partner link

Reply

Avatar

or to participate

Keep Reading