This archive report was first published on 20 August 2021.
On August 20, 2021, Google announced its latest innovation in search technology, MUM (Multitask Unified Model), at the Google I/O 2021 conference. The tech giant aims to revolutionize search with this next-gen algorithm, making it easier for users to find answers to complex questions.
According to Pandu Nayak, VP of Search at Google, MUM is trained on 75 unique languages and a wide range of tasks, enabling it to provide a more comprehensive understanding of information on the web. The model uses the Text-to-Text (T5) structure, which is reportedly more impressive than the BERT model introduced by Google in 2019.
One of the key features of MUM is its ability to understand complex queries, including those that require reasoning or thinking. For instance, a user searching for information on climbing Mount Adams might receive a response that includes information on the best gear to use, weather conditions, and other relevant details.
However, concerns have been raised about the potential centralization of web traffic on Google. Sara Watson, a senior examiner at Insider Intelligence, believes that MUM could lead to a decrease in traffic to other websites, potentially harming publishers. Google has denied these claims, stating that MUM is not designed to keep more web traffic for itself.
As Google continues to develop MUM, it remains to be seen how this next-gen algorithm will impact the way we search the web. With its potential to provide more accurate and faster search results, MUM could be a game-changer for search technology.
