Microsoft
MicrosoftSome feedback about ChatGPT...
With LABS.GENERATIVEAI, we often get beginning comments like, "Here is the information you asked for:". And we get ending comments like, "Is there anything else you need?"
In formulas, we seldom want that kind of text to be returned. So we need an easy, reliable way to suppress that extraneous content so we ONLY get the results we ask for.
We also need a reliable way to limit the amount of text returned. To illustrate in Bing:
Prompt: In 30 to 40 words, explain why the Fed raised interest rates.
This gave me a 55-word result.
Prompt: Why did the Fed raise interest rates? Use no more than 35 words.
This gave me a 32-word result.
I've found that when I give OpenAI and Bing the same prompts, OpenAI usually gives me more and better-formatted information. So for example, when I ask how to do something in Excel, OpenAI gives me numbered steps and Bing gives me a big paragraph.
Several days ago, I asked both OpenAI and Bing to "Give me 25 reasons not to use Excel for BI." Bing gave me ten reasons that averaged 10 words each. And OpenAI gave me 22 reasons that averaged 28 words each.
And today, I repeated the question. OpenAI gave me 25 numbered reasons. And Bing gave me, "I’m sorry, but I couldn’t find 25 specific reasons not to use Excel for BI from the current web page context or my web search results. However, there are some reasons why one might prefer using a dedicated BI tool like Power BI over Excel for their business analytics needs. https://blog.bismart.com/en/excel-vs-power-bi-which-one-is-betterhttps://blog.bismart.com/en/excel-vs-power-bi-which-one-is-better. https://blog.bismart.com/en/excel-vs-power-bi-which-one-is-betterhttps://blog.bismart.com/en/excel-vs-power-bi-which-one-is-better. Would you like more information on this topic?"
Similarly, when I asked OpenAI to write a limerick about Excel, it gave me these five lines:
In cells of Excel I do dwell,
Where numbers and formulas do tell,
The data I process,
With ease and finesse,
And in tables and charts I excel!
But in Bing, I got a paragraph:
There once was a program named Excel, Whose cells and formulas did compel, With charts and with graphs, And pivot table laughs, It made number crunching quite swell!
Bing wrote a better limerick, but it should have been formatted.
Finally, there's apparently no way to archive an AI conversation in Bing. But in OpenAI, that's easy to do. (I can't find a way to export a conversation, however.)
In short, I've found that where recent information isn't needed, OpenAI gives me more useful results than Bing does.
