Raphael Thys
Behind the Scenes of the new Microsoft Bing: What You Need to Know
Source: OFFICIAL LEWEB PHOTOS, CC BY 2.0, via Wikimedia Commons
I have gotten access to Microsoft Bing empowered with GPT-4.
Honestly, it is fantastic.
Everything I wanted to research during my consulting days came in seconds. So what would have taken a junior analyst an hour or so is shortened to 10 seconds.
An example is shown below.
“Give me market share data of the search industry and put it into tabular format,” and the below pops out. Needless to say, my default search engine is now Microsoft Bing rather than Google.
Source: Bing Search Results
And from the data above, it is evident that Microsoft Bing only makes a fraction of what Google Search does.
Source: Bing Search Results
Satya (Microsoft CEO) is now trying to take a bite out of the lion’s pie. He wants to challenge Google’s core business. And boy, it has been making headlines.
You can see the recent surge in Microsoft Bing in Google Trends.
Source: Google Trends
Satya has been involved with Bing before in his tenure at Microsoft.
He joined Microsoft in 1992 and initially worked on developing Windows NT, a family of operating systems. He later worked on other projects, including developing the company’s online advertising platform and cloud computing offerings. In 2008, he became the senior vice president of research and development for the company’s Online Services Division, which included Bing.
While overseeing the Online Services Division, he played a crucial role in shaping Microsoft’s search strategy and driving innovation within the Bing team. In addition, he was known for focusing on data and analytics to improve the search engine’s relevance and performance.
And now, he is transforming the search experience with Microsoft Bing.
But if you look under the hood, you can see that Microsoft Bing + ChatGPT needs to make more business sense.
Concern #1: Cost
The first primary reason is the costs associated with operating the new Microsoft Bing.
According to David Friedberg in the all-in-podcast, it has been estimated that running the GPT-3 model for chat GPT takes about 30 cents of compute for each result, which is about an order of magnitude higher than the 2.5 cents per search that it costs Google to present the results of a search query.
This higher cost of Bing has been a significant challenge for Microsoft in terms of being able to compete effectively with Google. Bing needs to significantly reduce its costs and provide a more cost-effective search service to users to be competitive.
A 10x cost reduction is necessary to make Bing more competitive. Achieving this will require deep technical expertise in optimization and compute platform management. Bing needs to optimize its models to provide search results and efficiently use compute resources.
So there’s still quite a lot of work to go before this becomes truly economically competitive with Google. And that really matters because if you get to the scale of Google, you’re talking about spending eight to $20 billion a quarter just to run search results and display them. — All-In Podcast
Reducing costs will make Bing more competitive and enable it to invest in developing new features and capabilities that can better serve users. It can also make the search market more dynamic, giving users more choices and making the competition between Bing and Google more intense.
This can lead to better user search experiences and more innovation in the search industry.
Overall, the high cost of Bing is a significant challenge for Microsoft.
Concern #2: Business Model
Another concern is how publishers profit from this model.
Microsoft Bing’s business model is causing concerns among publishers, as they are unclear on how to monetize their content through the search engine.
In an interview with The Verge, Satya Nadella, CEO of Microsoft, did not clearly answer how publishers can get paid for their content.
The concern is that when users search for a topic on Bing, they are presented with a list of answers provided by the search engine, and they may not click on links to the publisher’s website to read more. In addition, as artificial intelligence (AI) continues to improve search engine capabilities, users may be looking for more direct answers, and publishers may struggle to monetize their content.
The traditional model of search engine monetization involves presenting paid links alongside organic search results, with users clicking on the ads that provide better answers than the organic results.
This model is not applicable when search engines provide direct answers, as users may not need to click on any links. However, one potential solution could be to charge an affiliate commission if the answer provided by the search engine contains a link.
However, this approach may raise concerns over the accuracy and fairness of the answers, as advertisers could potentially manipulate the results. Publishers will need to work closely with search engines to develop a new business model that allows them to monetize their content while maintaining the quality of search results.
Despite these challenges, it is essential to note that they are temporary, and Microsoft is already taking steps to address them.
Hardware and compute costs are coming down exponentially.
If users use the engine in a sticky manner, they will find a way to monetize somehow.
Satya Nadella’s leadership and the company’s technical expertise in optimization and compute platform management can position Bing as a viable alternative to Google.
Moreover, Google has always found a way to countermeasure its competitors and is likely to do so again. Microsoft Bing’s current market share is tiny compared to Google, and the search giant is unlikely to give up its dominant position without a fight.
Overall, the integration of GPT-4 into Microsoft Bing has excellent potential, and it is exciting to see how it will shape the search engine market in the future.
And the interesting question is: What would Sundar do now?
Subscribe to DDIntel Here.
Visit our website here: https://www.datadriveninvestor.com
Join our network here: https://datadriveninvestor.com/collaborate