🦅 OpenAI’s $6.5B Gadget Stuck In Slides, 🤖 Sora 2 Pays Creators, 🎓 ChatGPT Opens Its App Store

✍️ Essentials

OpenAI bought Jony Ive’s Io Studio for $6.5B, hired 20 engineers from Apple and a team from Meta Quest to create a screenless, always-on AI device that could replace the smartphone. The idea was to kill the screen, not the conversation.

But in 2025 there’s still no final design, no interface, and no launch date. AI assistants talk too much - no one has figured out how to make them useful without being exhausting. The main blocker is infrastructure: Sora consumes enormous compute, ChatGPT is near capacity, and this device would need to stay online and listen continuously.

Google and Amazon already have the advantage of cloud capacity, custom chips, and over 100 million home devices. OpenAI, for now, only has presentation slides.

ContextHumane AI Pin raised $240M and spent seven years developing a wearable - only to ship an overheating badge with a 3-second lag and fewer than 10K sold units. Reviews were brutal.Rabbit R1 went viral on TikTok but turned out to be a repackaged API with a button. Out of 10K devices, up to 60% were returned.Meta’s Ray-Ban smart glasses work as a camera, but the AI assistant still struggles with conversations and language support.Amazon’s Alexa sits on a billion devices but still relies on scripts, not generation.Google killed Project Iris (its AR glasses), and Project Astra is still in demo.Apple’s Vision Pro is a visual masterpiece - but lacks AI entirely.Nobody has managed full-stack performance: recognition → generation → action, in real time, on the device itself.

🐻 Bear’s Take

The market for AI gadgets promises to be massive, but no one has cracked a seamless experience yet. Products like AI Pin and R1 make users skeptical: they expect magic, but get lag and glitches. Even major players can’t yet handle real-time generation. Until hardware supports local inference and better user scenarios, there’s no point in building new platforms around it.

🚨 Bear In Mind: Who’s At Risk

✍️ Essentials

Sora 2 is overflowing with AI-generated videos featuring Mario, Michael Jackson, Bob Ross, and Pikachu - no authors, no rights, no regulation. Yet the traffic is massive: Sora hit #1 in the App Store within 24 hours, surpassing ChatGPT and Gemini even with invite-only access.

Now OpenAI introduces monetization: it will share revenue with rights holders and add manual control. Owners of likenesses can now decide - allow face use, allow voice use, or prohibit it entirely. The model shifts from “opt-out” to “opt-in”: creators must explicitly permit their content to be used.

The payout system works via revenue share. Example: an AI video featuring Jackson goes viral, and OpenAI sends a check to the holder of his likeness rights.

ContextTikTok has 1.5 billion MAU and faces a long-term threat of U.S. bans. Meta pushes Reels and now Vibes with generative storytelling.Google’s DreamTrack and YouTube Music AI offer limited features but no real creator monetization.Sora moves one step further - toward legal AI UGC, where real brands, celebrities, and fan creators coexist in one ecosystem.

🐻 Bear’s Take

For brands, Sora enables campaigns using known characters legally and at low cost. For creators, it opens a new earning model. OpenAI is quietly building a hybrid of TikTok and YouTube with IP-driven monetization.

🚨 Bear In Mind: Who’s At Risk

✍️ Essentials

Everyone expected Dev Day 2025 to unveil GPT-6 with permanent memory and personality - a voice assistant that makes calls, full Sora-2 rollout, and the debut of the Ive hardware.Instead, OpenAI delivered something else entirely: ChatGPT as a platform.

Inside it now run apps like Canva, Figma, Spotify, and Zillow. Users can launch them directly inside the chat. Developers can build their own apps through the new Apps SDK and monetize them in the upcoming App Store.

Alongside that, OpenAI introduced AgentKit - a visual no-code editor to build AI agents with logic, API calls, and automated testing, all inside a browser.Sora works inside ChatGPT for Plus users - just type a prompt and get video output.GPT-5 Pro adds a reasoning model, a low-cost mini voice version, and GPT-5-Codex for Slack integrations and workflow SDKs.

All of this signals the start of a new environment where browsing, automation, and generation happen in one interface.

🐻 Bear’s Take

For businesses, it means creating and monetizing AI apps or agents in hours, not months. For users, it removes switching friction between tools. OpenAI just merged Chrome, Zapier, and the App Store into one chat window.

🚨 Bear In Mind: Who’s At Risk