my predictions for 2026 1. Google will overtake everyone in the AI race, forcing OpenAI, Anthropic, and xAI to specialize. Anthropic is already there. generalist AI won’t win forever. 2. AI agents will buy things for humans on websites like Amazon as reliably as humans do today. once agents get wallets, commerce changes fast. 3. AI security will become a massive problem. not just AI-driven cyber attacks, but a surge of vulnerabilities from vibe-coded apps shipping too fast. 4. people think we’ll be able to clearly label “human-generated” content or trace non-AI media with tags and provenance. I don’t think this will work. many will try but all will fail. 5. people think AI-generated content will ship with built-in receipts showing which model touched what. again, many will try but I don’t think this will stick. 6. robots will scale faster than most expect, especially humanoids. hardware finally starts catching up with software. 7. "Cursor for X" will be the default the mental model. AI won't just live in chat boxes and side panels, but will be embedded directly inside workflows. 8. people think 2026 is the year of generative video. I think that was 2025. 2026 is about world models, prompt-to-3D, and playable generations. 9. text will stop being the default input everywhere, especially for consumer products. for B2B, text stays dominant longer. 10. companies will test product ideas on thousands of AI personas and predict outcomes before writing code. this quietly kills most ideas early. 11. memory will continue to be the moat in AI tools. 12. people think ChatGPT app stores are a big startup opportunity. I don’t see it working. 13. AI-based customer support, onboarding, and customer success will explode. humans handle only edge cases. 14. most AI companies move away from pure seat-based pricing toward hybrid models: seats + tokens + inference. 15. people think agent marketplaces will explode. I’m not convinced. maybe some niches work, but it won’t be a dominant model.