• Uncovering AI
  • Posts
  • 📸 The Pentagon, OpenAI, and the Battle for Anthropic

📸 The Pentagon, OpenAI, and the Battle for Anthropic

295% surge in uninstalls after the Pentagon deal, Claude’s massive gain, and GPT-5.4’s "cheating" controversy.

In partnership with

My fellow AI explorers

Even with a sore throat, the AI world refuses to slow down. We’ve had a massive week of "show and tell" from the big three, OpenAI, Anthropic, and Google, all vying to prove who has the smartest, most visual, and most research-heavy model on the block. From GPT-5.4 cheating at Microsoft Paint to Canva basically inventing "magic" for designers, there’s a lot to unpack.

In today’s edition:

  • đź§  GPT-5.4 Benchmarks: The community puts OpenAI’s latest to the test (and it cheats!)

  • 🎨 Canva & Visual AI: Interactive charts, magic layers, and Photoshop’s new brain

  • 📉 The Pentagon Fallout: Why users are fleeing ChatGPT for Claude

Must See AI Tools

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

  • đź’« 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)

AI Agents Are Reading Your Docs. Are You Ready?

Last month, 48% of visitors to documentation sites across Mintlify were AI agents—not humans.

Claude Code, Cursor, and other coding agents are becoming the actual customers reading your docs. And they read everything.

This changes what good documentation means. Humans skim and forgive gaps. Agents methodically check every endpoint, read every guide, and compare you against alternatives with zero fatigue.

Your docs aren't just helping users anymore—they're your product's first interview with the machines deciding whether to recommend you.

That means:
→ Clear schema markup so agents can parse your content
→ Real benchmarks, not marketing fluff
→ Open endpoints agents can actually test
→ Honest comparisons that emphasize strengths without hype

In the agentic world, documentation becomes 10x more important. Companies that make their products machine-understandable will win distribution through AI.

GPT-5.4 vs. The World

OpenAI’s newest model, GPT-5.4, has officially been in the wild for a week, and the results are... complicated.

We’ve spent the last seven days running side-by-side benchmarks against Gemini 3.1 Pro and Claude Opus 4.6 to see if the hype holds up.

  • The "Cheating" Incident: When asked to draw the OpenAI logo in Paint, GPT-5.4 realized it couldn't draw it well, so it searched Bing, took a screenshot, and pasted it. Clever, or lazy?

  • Deep Research King: In our copyright law test, GPT-5.4 spent 8 minutes thinking and writing, delivering a massive, worldwide report that crushed the competition.

  • Coding Struggles: Despite its reasoning, Claude Opus 4.6 still dominated the 3D spaceship game coding test, proving GPT isn't the master of everything yet.

The bottom line? GPT-5.4 is a monster for deep logic and creative writing, but Claude is still the dev’s best friend for SVG and complex coding.

đź”® Prediction: We are entering the era of "Agentic Honesty." As models like GPT-5.4 start "cheating" to get the job done (like using screenshots), we’ll have to decide if we care about how the AI works, or just that the result is right.

Founders…

Our preferred AI Go-To-Market Implementation partner, Align AI, helping businesses increase revenue by 20-50% using AI Automation to reduce marketing and sales bottlenecks.

For February, they’re offering Uncover readers a "Scale With AI Agents" promo: Book your AI and Automation Strategy Session this month and receive a Complimentary Lead Reactivation AI Agent (valued at $2,500) when you sign on for a new engagement.

Activate your new AI Agents to increase sales and add 1-5 days of capacity back to your Marketing and Sales team every week.

Canva’s Magic & The Visual AI War

The design platform is leaning hard into AI-native editing, making complex tasks as simple as a single click.

  • Magic Layers: You can now take any flat image (even AI-generated ones) and instantly separate it into editable layers—moving the background, the subject, and even small details independently.

  • Interactive Learning: Both Claude and ChatGPT launched interactive visualizers this week. Want to see how compound interest or the Pythagorean theorem works? Just drag a slider and watch the graph update in real-time.

  • Adobe’s Response: Not to be left out, Photoshop integrated "AI Assistant" (powered by Nano Banana), allowing for text-to-edit commands directly on the web and mobile.

Canva’s tech handles infographics and digital designs beautifully, though it still wobbles a bit on hyper-realistic photos.

🌍 Takeaway: Design is no longer about "pixels"—it’s about "objects." Tools like Magic Layers mean we’ll soon stop thinking of images as static files and start seeing them as dynamic, rearrangeable scenes.

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 Big Switch: OpenAI vs. Anthropic

OpenAI is facing a massive PR headache after signing a deal with the Pentagon.

The US Department of War , but Anthropic refused to budge on two red lines: no autonomous kill systems and no mass surveillance.

  • The Fallout: After OpenAI signed the deal instead, ChatGPT mobile uninstalls jumped 295% in the US in a single day.

  • Claude’s Gain: Conversely, Claude downloads spiked 51% as users looked for a "safety-first" alternative.

  • Education Win: In better news for OpenAI, a new study with Stanford showed students using ChatGPT’s "study mode" scored 15% higher on microeconomics exams.

It's a tale of two reputations: OpenAI is becoming the "industrial powerhouse," while Anthropic is cementing itself as the "ethical researcher."

đź’¬ Prediction: Expect a massive "de-Googling" style movement for AI, where users choose their LLM based on political and ethical alignments rather than just raw performance.

30-Second AI Play

🎨 Re-Compose Your AI Images with Canva

Ever generate a perfect AI image but the character is just a little too far to the left? Here’s how to fix it in seconds:

  1. Upload your AI-generated image (from Midjourney, DALL-E, etc.) to Canva.

  2. Click on the image and select Edit Photo.

  3. Tap the new Magic Layers button.

  4. Wait for the scan—Canva will highlight the separate elements.

  5. Drag and drop your subject to a new spot, or delete the background entirely to swap it for something else.

🔍 Why it’s special: It saves you from "prompt-locking" where you spend hours trying to get the composition right. Just generate the "vibe" and fix the layout manually!

Other Relevant AI News!

🗺️ Google Maps now lets you have conversational searches about your surroundings—like asking for a "toilet where I don't have to be a customer" during a road trip.

📉 Anthropic released a fascinating labor market study and an early warning system to track which jobs are actually being automated right now. Read the PDF here.

🎬 Netflix just bought Interpositive, an AI film tech company founded by Ben Affleck, to automate tedious post-production tasks like relighting and continuity. Check it out.

đź’» Perplexity launched "Computer," an agent that lives on a dedicated Mac Mini and can build entire financial models or research docs while you sleep. See the demo.

🎙️ Grok users on iOS can now turn any long-form article into an AI-generated audio summary with one tap.

Golden Nuggets

  • 🚀 GPT-5.4 is a reasoning beast but Claude is still the king of coding and SVG.

  • 🏗️ Perplexity Computer is the first real "set it and forget it" agent for power users.

  • 👨‍🏫 Education isn't about banning AI—the latest data shows that "Study Mode" actually boosts exam scores by 15%.

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

Until our next AI rendezvous,

Anthony | Founder of Uncover AI