By Peter Garnry, Head of Equity Strategy at Saxo
Peter Garnry |
Microsoft has recently invested $10bn in OpenAI, the company behind the hyped ChatGPT, and is keen to roll the technology out across its various product. Before we answer the question of whether OpenAI’s ChatGPT could be the end of Alphabet as we know it we must explain some things about ChatGPT.
It is not revolutionary to produce a chatbot which can reply to questions from humans, and we already have AI technology integrated in many different applications . So what has made ChatGPT so popular in such a short time?
It is one of the first AI technologies which is easy to understand and to interact with for the general public. The difference from other chatbots is ChatGPT’s ability to respond in a language which is difficult to distinguish from a chat with a real human being. And it is trained on a much larger dataset with massive computational power, making it far superior to other chatbots.
The cost of offering such a service to the general public does however not come without a cost. This may also be one of the reasons why some of the earlier chatbots have not been presented to the general public. The estimated cost of running ChatGPT is around $100,000 per day Each word generated on ChatGPT is estimated to cost $0.0003, so it’s not a cheap exercise for ChatGPT.
Will ChatGPT become a replacement for Alphabet – or even the internet as Microsoft ’s CEO said on the Q4 2022 earnings conference call? ChatGPT is a machine learning model trained to generate text on a given input, where Alphabet is a search engine designed to find information on the internet through relationships of links between websites. Whereas Alphabet is able to find information which has been posted recently, ChatGPT needs to be re-trained on new data containing this information, which can be a computationally heavy process.
Our interaction with the current version of ChatGPT, clearly demonstrates limitations in the input data: