Bing Chat for search and research

ChatGPT launched in late 2022 and the popularity just sky-rocketed after Microsoft investment announcement. But has it delivered on it's promise. Or was is just a hype? Well, every new technology has a hype-cycle. And every hype-cycle that I have lived through (Internet, .dotCom, et al.) have changed my work or life in some way even after the hype plateaued. Sometimes profound, sometimes not so much. But changed it has.?

In my personal opinion, ChatGPT was never aimed at replacing Google. The mission of ChatGPT was to unleash the potential of a technology called Large Language Models via it's Open AI Framework. Multiple applications have thus spawned that claim to replace traditional tools and even people. Most of these are in the area of Marketing, Content writing, creatives and music. I unfortunately have no work in those areas. But for me the most impactful side-effect of ChatGPT has been the Bing Chat, which uses Open AI to take search experience to the next level. And for me, it has been an absolute blessing.?

So then one big question in the article is, why is the Bing user base still at a measly three something percent? Over more than two decades of dominance in the search industry, Google has become a verb. Which means it's going to take decades to displace Google Search.? But that's not the focus of this article. I want to share how I use Bing Chat to improve my work efficiency every day. Following is an example of how I use Bing Chat, especially for Research.?

A few days ago, a Microsoft customer I work with sent me a mail with the following chart:

It’s the Rate Limit chart from the Metrics section of the Azure Monitor for Azure Open AI. They wanted to know what does the Y Axis exactly represent? They claimed that have been very carefully managing their Prompts within the Token Limits of the Model and were not sure what these peaks were about.?

All Azure Open AI models have a Quota for the number of Tokens that you can submit per query. So I was sure that the chart represents the Token Rates. But, I wanted be sure. So I researched.?

I opened Bing Chat and asked what I wanted to know.

Turns out, it's not just the Tokens Per Minute (TPM), it's also about Requests Per Minute (RPM). Huh! It looks like it's a combination of RPM and TPM. Observe how it also tells me where it is getting specific lines of information from with those little superscripts of 1 and 2. Looks like it's getting information from Stack Overflow. I love Stack Overflow, but since I am replying to a customer I want to make sure it is from Microsoft official documents. So, I type another query.

Voila! Everything I need to know. Including details like 6 RPM per 1000 TPM, which is crucial piece of information. And it is from the Microsoft Documentation. Which means the customer should also monitor their number of requests they make and not just Token Limits. They might hit the Rate Limit with too frequent requests even if they are within the Token Limit.

?If I run the same query on Google…

I am sure I can find the answer in the first couple of links. But it's reading through them and putting pieces of information together. More clicks and more time.

And just before writing this article, I also tried the same query on Google Bard...

Looks like it completely missed TPM related information. And most importantly, I don't know where it's fetching its information from.

In conclusion, ChatGPT and Open AI spawned Bing Chat that changed the way I search and research, forever.

?Are you using Bing Chat the same way? If you have better tricks, please comment.

?