Sadie St. Lawrence's 2026 AI Predictions: Nested Learning, Spatial Intelligence & More

2025 AI Predictions Recap

• 🎯 Agentic AI was the dominant trend in 2025, so much so that it's now considered passé for 2026 discussions.
• 💡 AI integration into everyday devices accelerated, with real-time translation in AirPods being a highlight, though many other integrations felt unnecessary.
• 🔬 AI-driven scientific research saw mixed results, with some promising discoveries but also a notable paper redaction.
• 💰 Enterprise AI monetization remained crucial, with many projects stuck in MVP stages due to implementation challenges and unclear ROI.
• 🚀 Demand for AI engineering skills surpassed traditional data science skills, reflecting a shift in the job market.

2025 AI Awards

• ✨ Wow Moment: The ability of AI agents to build a functional stock trading platform in 90 minutes, and the creation of the HMCI.ai website with a single prompt using Lovable.
• 🏆 Comeback of the Year: Google, for its seamless integration of AI into search and G Suite with models like Gemini 3 Pro and Nanabanana Pro.
• ⚠️ Disappointment of the Year: The gap between the hype of agentic AI and its practical implementation in enterprise settings, and Apple's perceived slow progress in AI integration.
• 🥇 Overall Winner: Nvidia, for its foundational role in AI hardware, and the concept of 'Vibes' as the overarching theme of 2025 AI developments.

2026 AI Predictions

• 🧩 Specialized Industry Models: Expect more domain-specific AI models, akin to AlphaFold, due to limitations in data for training massive general-purpose models.
• 🧠 Continual and Nested Learning: Models will increasingly be able to update and learn continuously, bridging a key gap between AI and human intelligence.
• 💡 Return to Research: Labs will refocus on fundamental research for the next AI breakthroughs, as scaling alone is no longer yielding significant advancements.
• 🌍 Spatial Intelligence & Physical AI: Momentum will build in exploring 3D environments and simulations as new data sets to bring AI into the physical world and improve robotics.
• 🛠️ AI Ops Emergence: A new job category, AI Operations, will arise, focusing on GPU management, model orchestration, and agent reliability, similar to DevOps in its early days.