Microsoft Bing And The Future Of Search
Generative Search With GPT-4
The word ‘Google’ has made its way into the English language as synonymous with the word ‘search.’ For many, Google is the main gateway interface to the web.
Unfortunately, filled with ads, sponsors, and unknown ranking models, Google hasn’t always been the ideal search experience.
That makes it ripe for a contending search model.
Since the arrival of ChatGPT, there has been a lot of speculation on how something like ChatGPT can be a replacement for Google.
The driving factor is the conversational simple-to-use interface that responds with humanlike answers, as opposed to supplying a list of web links to click through.
What makes ChatGPT powerful is that it can respond to questions, summarize long texts, complete sentences, and perform many other natural language tasks.
But is it reasonable to assume that ChatGPT, which is still flawed with potential misinformation, along with the energy costs used to run a synthesis engine like ChatGPT, could really be used by a billion people to access valuable information?
Here Comes Bing
According to a report by Semafor:
“Microsoft’s second-place search engine Bing is poised to incorporate a faster and richer version of ChatGPT, known as GPT-4, into its product in the coming weeks, marking rapid progress in the booming field of generative AI and a long-awaited challenge to Google’s dominance of search.”
There currently isn’t any info from Microsoft or OpenAI on this yet, so we can only speculate on what the user experience would be like, or what this means for the future of search.
which raises the interesting question of how would this work.
What Do We Use Internet Search For?
The main reason that people ‘search the web’ is to look for information.
If we use the mental model of a library, imagine entering a library looking for information. You can ask the librarian a question about what info you are looking for.
Now imagine that the librarian has also read all the books related to the topic you are asking about.
You have the option of getting 2 types of responses:
- The librarian can guide you to the books that provide the answer you are looking for.
- The librarian can tell you in her/his own words what the answer is, having read the books.
This is the difference between link-based search and generative search.
Both of these responses are dependent on what type of answer I am looking for. Both are valid forms of responses.
The challenge to link-based search is that you would have to filter through all the sources. The challenge to generative search is that the answer could be biased based on the interpretation of the librarian (Large Language Model).
It’s likely that in a lot of cases, generative search is what most people really want, and then if necessary, move to link-based search when more accurate detailed information with citations is necessary.
This is why a conversation-style ChatGPT can be a viable alternative and poses a threat to Google search.
The Ideal Search Engine
In my opinion, the ideal search engine would work as follows:
- It starts with a generative search, a natural language response to the question I am asking. All information must be 100% factual. If it’s an opinionated response, it must state so, and may include various opinions.
- After the initial natural language response, I can ask for citations and link references. This would point to web links, videos, or media files that support the response.
Final Thoughts
It remains to be seen if Bing can pull this off.
If Bing can put together a search engine with the power of generative search (based on real factual info) combined with the optional link-based search, it will be insanely popular.
Of course, that doesn’t in any way suggest that Google doesn’t also have a generative search model planned. It will be interesting to see how this all progresses.
If you liked this article, throw out some Medium love… claps, comment, and be sure to follow.
Reference: The librarian analogy was inspired by a podcast episode from ‘Another podcast.’