The field of internet congestion control and networking is witnessing significant developments, with a focus on improving efficiency, reducing latency, and enhancing overall performance. Researchers are exploring innovative approaches, such as machine learning-based congestion control and latency-oriented extensions to existing protocols, to address the challenges posed by evolving network environments and applications. Notably, the use of adaptive ML-based termination in internet speed tests and the evaluation of BBR as a default TCP congestion control protocol are contributing to a better understanding of the trade-offs between accuracy, efficiency, and deployability. Furthermore, the investigation of learning-based congestion control and its comparison to human-derived algorithms is shedding light on the strengths and limitations of these approaches. The exploration of new architectures, such as publish-subscribe variants of DNS, is also opening up possibilities for improved update traffic management and reduced query latencies. Some noteworthy papers in this area include: TURBOTEST, which introduces a systematic framework for speed test termination that achieves nearly 2-4x higher data savings than existing approaches. TCP ROCCET, which presents a latency-based extension of TCP CUBIC that reduces latency and bufferbloat in cellular networks. Learning-Based vs Human-Derived Congestion Control, which provides an in-depth experimental study of learning-based congestion control and its comparison to human-derived algorithms.