The field of speech processing and translation is rapidly evolving, with a focus on improving efficiency, accuracy, and robustness. Recent developments have led to the creation of novel models and techniques that can handle multiple language pairs, noisy environments, and long-form speech transcripts. One of the key areas of research is the development of speech foundation models that can serve as general-purpose representations for a wide range of speech-processing tasks. These models have shown impressive performance on various downstream tasks, including speech recognition, speech emotion recognition, and spoken language understanding. Another area of focus is the improvement of noise robustness in speech recognition systems, with techniques such as variance-invariance-covariance regularization and dynamic pruning being explored. Additionally, there is a growing interest in developing efficient and accurate speech translation systems that can operate in real-time, with approaches such as simultaneous translation and alignment-based streaming machine translation being investigated. Noteworthy papers include: VARAN, which proposes a framework for dynamically tailoring layer aggregation to individual inputs, resulting in superior performance on automatic speech recognition and speech emotion recognition tasks. HuBERT-VIC, which introduces a noise-robust speech foundation model with variance-invariance-covariance regularization objectives, achieving significant improvements in noise robustness. CarelessWhisper, which presents a method for turning a transformer encoder-decoder model into a low-latency streaming model, outperforming existing non-fine-tuned streaming approaches in most cases.