Recent advancements in machine learning and computational chemistry have led to several noteworthy developments. Sakana AI introduced Transformer², a self-adaptive architecture for large language models (LLMs) that allows real-time adjustments of model weights based on specific tasks. This innovation aims to enhance the efficiency of LLMs, reducing the need for extensive fine-tuning. Concurrently, EvolutionaryScale announced the release of ESM3, a generative language model capable of reasoning over protein properties such as sequence, structure, and function. This model has been made available for free to researchers worldwide via a public beta API, as highlighted in a recent publication in *Science Magazine*. Additionally, the SHAPES framework has been introduced for evaluating generative models of protein structures, utilizing structural embeddings to quantify distributional similarity. These advancements signify a growing trend in the integration of AI with biological research, promising to enhance capabilities in protein engineering and related fields.