Two weeks ago top government officials, corporates, media houses and other influential people from the world gathered in Davos for the annual meeting of World Economic Forum .Over the years there has been good and bad publicity of the event but given that we are at an inflection point of Artificial Intelligence, it is hard to ignore the event.
Focusing on the open discussions on the topic during the event resulted in a deep insight into what was happening and what was about to happen. Before going into details of some, there were three common insights. One is that regulation needs to be there to control the risks, second the benefits of it outweigh the risk and third for now it will augment employees and not replace them.
There was often the question asked about the power consumption as processing such a huge amount of data needed high computing. But the answers to it were more logical, as the data centres would consume a lot of power that might double the consumption overall but the output and its level of quality will be more productive on a per unit consumption.
Another interesting insight was that almost all major LLMs have or are about to finish consuming the data present on the web. This would mean that for new and unique data they will have to depend on collaborating with companies or institutions rich in their field.
Though most insights were gains these five are major ones that define the near and medium-term trajectory of AI. But do be aware that within such high computing levels, this can be equal to max 5 human years.
コメント