The talk at the FOSDEM conference in 2023 was about Apache Beam, a framework for data processing that specializes in streaming analytics. Apache Beam provides a unified computing model for batch and streaming, allowing pipelines written for batch processing to be easily adapted for streaming. It also offers multi-language support and can run on different platforms and runners, making it highly portable. The talk touched on various topics such as windowing, watermarking, stateful functions, machine learning inference, and cross-language transformations. The speaker emphasized the advantages of using Apache Beam in terms of portability and the ability to easily switch between different runners. The talk also provided links to additional resources and recommended further reading on Apache Beam.