Today Google Bard gets a little smarter with added math and logic features. Google employee Jack Crouchik announced the change on Twitter, stating, “Bard will better understand and respond to prompts for multi-step word and math problems as coding is coming soon.”
Logical questions were a major flaw when Bard arrived decades ago. Some of the responses made Bard look particularly silly to early testers. In one example from last week, Bard repeatedly claimed: 1 plus 2 equals 4Today, Google’s state-of-the-art AI chatbot model can now correctly say the answer is 3. So there have been at least some changes.Alternatively, you can correctly list the months of the year with make a name Like “Maruari”.
However, Bard still stumbles over a very basic question of logic.how to geeks Chris Hoffman On the first day, I asked the bard, “Which is heavier, a 5-pound feather or a 1-pound dumbbell?” Google Bard responded with the absurd claim that there is no such thing as a five-pound feather. In a reply, ChatGPT said that 5 pounds of feathers and his 1 pound dumbbell are “the same amount, which means he weighs 5 pounds” and didn’t do much more.
Ron Amadeo
With today’s update, Google Bard now gives the same wrong answer as ChatGPT: “A 5-pound feather weighs the same as a 1-pound dumbbell.” error (all of which look very bad in facts and numbers). But it’s interesting given that Google has been accused (and denied) of using ChatGPT’s output to train bards.
Logic is a major gap in Bard’s functionality, plus it’s artificially constrained from trying to answer programming questions, so I’m happy to hear that these features are coming soon from Krawczyk. It’s famous for being able to write tons of code in any language and style you like, and sometimes the code works.
Krawczyk adds: “We’re always balancing new features and efficiency in Bard. This update is just one example of the many improvements he’s making to Bard each week.”
Weekly improvements would be great. Google has been overwhelmed by Wall Street for taking a slow approach in releasing AI, but the company appears to be taking a slow approach with his Bard. The first release is called “Experimental” and is isolated to a small site at bard.google.com, not part of Google Search. Also, this service is only available in the US and UK.
In an interview with The New York Times posted today, Google CEO Sundar Pichai acknowledged that Google is still ahead of its best AI, saying: Some of our more capable PaLM models bring more features, such as reasoning and coding, to better answer math questions, so we can see progress in the next week. .
