The latest language model from OpenAI, GPT-5, does little more than enhance the user experience, opines MIT Technology Review’s Mat Honan, adding that he sees it as more of a refined product than a revolutionary breakthrough.
This is in line with the views of his colleagues at other publications, one of whom compared it to a new release of a smartphone with incremental improvements.
Furthermore, following its launch, OpenAI found itself confronted by a user revolt, with many customers seeking to revert to the previous GPT-4 after missing its personality.
Honan also questions the over-hyping of AI models, although concedes that he is a regular user of such programs.
Indeed, he confesses to having recently used ChatGPT to find out more about why whales slap their tails on the water, an act known as lobtailing, only to discover that the explanation offered by the model was not wholly accurate.
In his view, whilst AI has some “cool features”, lobtailing - or making over-hyped claims - is not one of them.