Just scored 76.11% on ARC-AGI 2 — beating public GPT-5.2 and Gemini-3-Pro baselines by >20%, and (as far as I know) the best publicly reported result so far. Approach: what I’d call Multi-Model Reflective Reasoning - Using GPT-5.2, Gemini-3, Opus 4.5 - Long-horizon/multi-step reasoning (~6hrs/problem) - Agentic codegen (>100,000 python calls) - Visual reasoning - Council of judges Fun fact: all solver code was written by Gemini-3-CLI. Does this count as AI generating a new AI that beats the prior SOTA? 🤔 Full run code (open source): kaggle.com/code/johanland/jo… @GregKamradt , holiday break is over 🙂 semi-private when? #ARCAGI #AIResearch

J Wales: "The majority of edits... are from... 600-1000 people". Waat? Not in line with #csn12 data? 10:30 @ alturl.com/ortjz