- Google unleashed the latest Bard edition on Wednesday
- The new Ai is still less capable than its competition
- It is slow to respond and incapable of answering straightforward questions
On March 21st, as Microsoft was unleashing their Bing Image Creator AI, Google also silently launched their Bard AI – and so far, it’s not looking too good for Alphabet Inc.
While the current iteration of Bard is not the final product yet, many who have tested the AI have found it wanting in many areas, especially considering just how far ahead GPT-4 is.
For example, in tests, Bard took a lot longer to respond to queries, a lot longer than GPT-4 and Bing Chat. A lot of this has been put down to differences in the language learning models (LLMs) used. GPT and Bing Chat run on the GPT series LLM, while Bard runs on a lightweight version of Google’s in-house language model, LaMDA.
Another thing about Bard is that, while it can pull up the latest data since it’s connected to the internet (GPT is still uninformed in events post-2021), it gives many simple and sometimes unhelpful answers, as was revealed in The Guardian.
Aside from that, the chatbot was unable to answer some straightforward questions. For example, a prompt of “Who are all the Presidents of the United States of America?” resulted in the generic response, “I’m unable to help you with that, as I’m only a language model and don’t have the necessary information or ability.”
Work in progress
Still, despite its failings, Bard AI proved useful in other functions like building CVs and resumes, essay generation, and writing down responses in the directed format, i.e., using bullets when asked to write in bullet points.
Overall, Bard is still lagging behind ChatGPT and Bing Chat, and thus, the latter two can be more useful all around for any task that you need. As a new entrant, Bard is still going through the early growing pains, and Google has promised to update it with more recent and more capable models.
So, if you want any task done by AI, stick to GPT-4 and Bing Chat.