Google Employees Criticize CEO For 'Rushed, Botched' Announcement of GPT Competitor Bard

Feb 13, 2023 · bard ·

Google employees and some experts in the Artificial Intelligence (AI) field have criticized CEO Sundar Pichai for his rushed and botched announcement of a competitor to the popular OpenAI’s GPT language model.

The Google team revealed their new AI tool, called BARD, on February 12th. BARD is a language model trained using transfer learning technique, which allows it to generate content that is more human-like than models trained completely from scratch. It was reportedly developed in response to OpenAI’s GPT-3, a larger and more powerful language model, and was meant to be an alternative to its use in areas such as natural language processing and conversation generation.

However, upon further analysis, it was found that the performance of the BARD model is not on the same level as OpenAI’s GPT-3. Commentators noted how the BARD model had been put together very quickly, with minimal input from researchers, resulting in a model that is less robust and versatile than its rival.

Google employees were particularly critical of Pichai's decision to hastily launch the model without properly assessing its capabilities or consulting experts in the field. They argued that the rushed announcement was only meant to distract from OpenAI’s own success and gain publicity for the company.

Critics have also pointed out the fact that Google has failed to provide any public benchmarks that can be used to compare BARD's performance to other models. With no way to gauge its efficacy, it is difficult to assess the model or make informed decisions about its use.

In response to the criticism, Google has stated that they are committed to improving BARD and that it will continue to be developed. Despite this assurance, the incident serves as a reminder of the importance of testing AI products thoroughly before launching them publicly. Time spent refining and perfecting a product can save businesses time and money in the long run, and help ensure that AI tools are reliable and safe to use.