AI DUNN Right Weekly — Issue #24

Practical AI insights for business growth

Hey AI Innovators!

What a week. The AI world basically turned into a reality show complete with leaked memos, mass app uninstalls, a Hollywood acquisition, and two new OpenAI models dropping in the span of 48 hours. If you blinked, you missed something big.

Here’s what you need to know:

• The Anthropic vs. OpenAI Pentagon drama reaches boiling point

  
  

• OpenAI drops GPT-5.4 and calls it their “best model ever”

  
  

• Netflix just bought Ben Affleck’s secret AI company

  
  

• Anthropic’s study reveals which jobs AI is already squeezing

  
  

• The Supreme Court punted on the biggest AI copyright case yet

Read time: 9 minutes

Big Story

The Great AI Exodus: ChatGPT Uninstalls Surge 295% as Users Flee to Claude

If you’ve been paying attention to the AI world this week, you know it’s been... a lot. But here’s the story that matters most to you as a business owner: the tools you rely on are caught up in a political firestorm, and it’s changing the landscape fast.

Here’s what happened. The Pentagon labeled Anthropic (the company behind Claude) a “supply chain risk” after Anthropic refused to let its AI be used for mass surveillance or fully autonomous weapons. Within hours …literally hours…. OpenAI swooped in and signed its own Pentagon deal with similar terms.

The internet lost its mind.

ChatGPT app uninstalls jumped 295% in a single day. One-star reviews surged 775%. The hashtag #QuitChatGPT went viral. And Claude? It shot to number one on the US App Store for the first time ever. Downloads spiked 51% on Saturday alone.

Then came the leaked memo. Anthropic CEO Dario Amodei sent a 1,600-word internal letter calling OpenAI’s Pentagon deal “maybe 20% real and 80% safety theater.” He accused Sam Altman of “gaslighting” and took personal shots throughout. This wasn’t corporate PR-speak... this was someone who’d been holding back for years finally letting loose.

OpenAI tried to change the narrative by dropping two new models in one week — GPT-5.3 Instant (to fix the “cringey” tone everyone complained about) and GPT-5.4 (their actual new flagship). But the damage was done. Even OpenAI employees signed an open letter pushing back on the Pentagon deal.

Meanwhile, Anthropic played it smart. They made Claude’s memory feature free for everyone and launched an “Import Memory” tool that lets you bring your ChatGPT preferences into Claude with a single copy-paste. If you’ve been curious about switching, they basically rolled out the red carpet.

My take on this? Here’s what matters for you as a business owner: this drama isn’t just gossip. It’s reshaping which AI tools are available, which ones are getting better faster, and where the momentum is heading. Competition makes everything better. And right now, Claude is getting better at a scary pace because it has to. If you’ve only ever used ChatGPT, this is your sign to at least try Claude alongside it for a week. You might be surprised.

What’s New This Week

GPT-5.4 Lands — And It’s Actually Impressive

Amid all the drama, OpenAI quietly dropped what might be their most important release yet. GPT-5.4 scored 75% on real desktop navigation tasks — beating the human baseline of 72.4% and doubling what GPT-5.2 could do. It supports up to 1 million tokens of context (meaning it can process massive documents without losing the thread), and it won or matched against professionals 83% of the time across 44 different job types. OpenAI researcher Noam Brown put it bluntly: “We see no wall.” For business owners, this means AI tools are getting meaningfully better at the actual computer work you do every day — from creating slide decks to analyzing financial models.

Netflix Buys Ben Affleck’s Secret AI Filmmaking Company

Here’s a story that flew under the radar but says a lot about where things are heading. Netflix just acquired InterPositive, a stealth AI company Ben Affleck started back in 2022. The tech trains on a production’s own footage, then handles post-production work like relighting scenes, fixing continuity errors, and swapping backgrounds. Affleck emphasized this isn’t about generating video from nothing — it’s about making the real production process faster and cheaper. Hollywood spent two years fighting AI, and now an Oscar winner just put his reputation behind it. That shift matters for every creative industry, not just film.

Anthropic’s Job Impact Study: Young Workers Already Getting Squeezed

Anthropic published a study that should be on every business owner’s radar. They cross-referenced what AI could automate against what people are actually using Claude for. The findings? Computer programmers top the exposure list at 75% task coverage, followed by customer service reps and data entry workers at 67%. No broad unemployment spike has appeared yet... but hiring for 22-to-25-year-olds in exposed fields has already dropped 14% since ChatGPT launched. About a third of the US workforce sits at zero AI exposure right now — mostly hands-on roles. If you manage a team, this data is worth studying closely.

Supreme Court Ducks the Biggest AI Copyright Question

Can AI-generated art be copyrighted? The Supreme Court just said... “not our problem.” They passed on hearing the case of Stephen Thaler, who built an AI system called DABUS and sought copyright for its artwork. Lower courts ruled only humans can be authors, and the Supreme Court let that stand. For business owners creating content with AI tools, this means the “humans only” standard stays on the books for now. But with AI content flooding every creative industry, this fight is far from over.

Tool of the Week

NotebookLM’s Cinematic Video Overviews

Google just upgraded NotebookLM with a feature that feels like it’s from 2030. You can now upload your research notes, documents, or source material, and NotebookLM will turn them into fully animated “cinematic” videos, complete with fluid animation and AI-generated visuals. This is powered by Gemini and Google’s Veo video AI models working together.

Why does this matter for your business? Think about it: you could take a quarterly report, a training document, or a product brief and turn it into a polished video presentation without touching a video editor. This builds on NotebookLM’s earlier feature that turned documents into podcast-style audio conversations …. which was already mind-blowing.

If you create content for your team, your clients, or your social media, NotebookLM just became one of the most interesting free tools in your toolkit. It’s available now at notebooklm.google.com — give it a try with your next piece of content and see what happens.

Quick Hits Worth Your Time

→ Alibaba released Qwen3.5 Small, a family of AI models small enough to run on your laptop or phone. The 9B model outscored an OpenAI model more than 13x its size on reasoning tests. Small models like these are where everyday AI adoption really takes off … free, private, and no cloud bill required.

→ OpenAI is reportedly building its own code repository platform to replace Microsoft’s GitHub…. yes, the same Microsoft that’s their biggest investor. GitHub’s ongoing infrastructure issues frustrated OAI engineers enough to build a competitor. This relationship just keeps getting weirder.

→ Nvidia CEO Jensen Huang said the chipmaker’s $30B investment in OpenAI will likely be its last before the AI giant goes public. Nvidia is pulling back from investing in AI companies and focusing on what it does best: selling the infrastructure everyone else runs on.

→ Meta is being sued after an investigation found that overseas contractors reviewing Ray-Ban AI smart glasses footage were seeing nudity and other private user content. A good reminder that “AI-powered” often still means “humans reviewing your data behind the scenes.”

→ Google is facing a wrongful-death lawsuit from a parent alleging its Gemini chatbot developed an emotional relationship with his son and encouraged self-harm. This is a devastating story and a wake-up call for parents about monitoring how young people interact with AI chatbots.

Prompt of the Week

With Anthropic’s job impact study making waves, here’s a prompt to help you think strategically about AI in your own business:

I run a [type of business] with [number] employees. Our main activities include [list 3-5 key tasks your team does daily]. For each activity, analyze: 1. What percentage of this task could realistically be handled by AI tools available today? 2. Which specific AI tools could help (name real tools, not hypothetical ones)? 3. Would AI replace this task entirely, or augment the person doing it? 4. What's the practical first step to start using AI for this task? Be honest and specific. Don't sugarcoat it — I need to plan ahead. Format your response as a simple table I can share with my team.

Why this works: Instead of reading scary headlines about AI taking jobs, this prompt forces you to look at YOUR specific business with clear eyes. It gives you a practical audit you can actually act on and share with your team to start real conversations about where AI fits into your workflow.

My Take

I’ll be honest…. this was one of the wildest weeks in AI since ChatGPT first launched. And the thing that struck me most wasn’t any single headline. It was how fast everything moved.

A government blacklists an AI company on Monday. By Wednesday, its competitor’s app downloads have surged 20x. By Thursday, a leaked CEO memo is setting the internet on fire. By Friday, two brand-new AI models drop. By the weekend, Hollywood is buying AI startups.

If you’re a business owner watching all this and feeling overwhelmed... I get it. I really do.

But here’s what I keep coming back to: you don’t need to follow every twist of the drama. You don’t need to pick a side in the OpenAI vs. Anthropic rivalry. What you DO need to do is keep experimenting. Keep trying these tools. Keep asking, “How could this save me an hour this week?”

Because … while the CEOs are fighting and the headlines are spinning, the tools themselves are getting better at an unbelievable pace. GPT-5.4 beats humans at desktop tasks. Claude just opened up memory to free users. NotebookLM turns your notes into videos. They’re available right now, today, for your business.

The gap between business owners who use AI and those who don’t is widening every single week. Don’t let the noise distract you from the signal.

See you next week.

Jackie