Use multiple models - by Nathan Lambert - Interconnects

kept by eddie jan 12

To get the most out of AI in 2026, the author recommends using multiple models for different tasks as each model has jagged capabilities.

Google Antigravity
Google released Antigravity, an agentic IDE and VS Code fork with three surfaces—agent dashboard, editor, and browser extension—for building apps with Gemini models.
also on Gemini 3 Pro, #generative-ai, #ai-coding
What (I think) makes Gemini 3 Flash so good and fast
Gemini 3 Flash likely uses a massive 1.2 trillion-parameter ultra-sparse MoE architecture, activating only 5–30 billion parameters per inference to deliver high intelligence at low cost, though it suffers from token bloat and a 91% hallucination rate on refusals.
also on Gemini 3 Pro, #generative-ai, #ai
Better AI models enable more ambitious work
A University of Chicago study of 500 companies found AI usage rose 44% as models improved, with developers shifting to more complex and cross-system tasks.
also on GPT-5.2, #generative-ai, #ai-coding, #productivity
GPT-5.2 | Hacker News
OpenAI introduces GPT-5.2, a model series for professional knowledge work that sets new state-of-the-art on GDPval and other benchmarks, with gains in coding and vision.
also on GPT-5.2, #generative-ai, #ai-coding, #ai
18 Predictions for 2026 - Jakob Nielsen on UX
Accelerating AI capabilities will shift focus to autonomous agents and Generative UI, making UX the primary business moat.
also on Nano Banana Pro, #generative-ai, #ai