According to @gdb, GPT-4.5 is state-of-the-art on a forecasting benchmark and, at current trends, models are about one year away from matching the performance of superforecasters. According to @gdb, this assessment references benchmarking shared by @Research_FRI on X.

According to @StanfordAILab, the presentation at ICLR will explore the integration of Large Language Models (LLM) in scientific research, which could significantly influence cryptocurrency trading strategies by enhancing data analysis and prediction accuracy.

According to Andrej Karpathy, the integration of AI in cryptocurrency trading could significantly enhance market efficiency and prediction accuracy, as AI algorithms can process vast amounts of data faster than humans (source: Twitter @karpathy, February 27, 2025). This could lead to more informed trading decisions and potentially higher returns for traders using AI-driven strategies.

According to AI at Meta, a new paper has been released detailing advanced AI models for cryptocurrency trading. These models reportedly enhance prediction accuracy by leveraging deep learning techniques, which could significantly impact trading strategies (AI at Meta).