Even in the relentless news and noise of early 2023, you've probably heard of ChatGPT, the GPT-3 (Tech 3.5) based chatbot that is uncannily capable of simulating human-like responses. For example, consider this question and answer (created by GPT-3):
Is ChatGPT a search engine?
No, ChatGPT is not a search engine. It is an artificial intelligence chatbot that uses natural language processing to understand and respond to user input.
It's not bad, and it sounds like something a real human would write.
GPT (Generative Pretraining Transformer) is a deep learning system that generates text based on the prediction of speech patterns after being pretrained on a large number of documents. For this article, I'm using OpenAI Playground with the GPT-3 text-davinci-003 model, which can be programmed to produce shorter responses and is more reliable than ChatGPT Beta.
The current models are truly impressive, including the analysis of complex questions:
How would you explain ChatGPT to a 5 year old?
ChatGPT is like a robot friend who can talk to you and answer your questions. It's like having a friend who knows a lot about different things and can help you learn new things.
GPT can even mimic human creativity and make us question our own talents:
Write a slogan for a Martian restaurant run by Edgar Allen Poe-esque lizardmen. "Lizards of Mars, dine in splendor here where dreams unfold and palaces indulge!"
By now you've probably seen dozens of examples. So what does all of this have to do with Google search and our jobs as search engine marketers?
Google and the era of natural language
In the early days of research, search engines were designed to index short words and phrases and were unable to process natural language. If you are looking for a new laptop, you can search for terms such as "laptop" or "buy a laptop". I sometimes joke that if you walk into Best Buy and start yelling "Laptop!" security will probably escort you out. However, humans adapt to early research and simplify our language.
As search became more sophisticated, we learned that we could use more complex queries, and especially with the advent of voice search, we experimented with something closer to natural language. As Google adapted, something interesting happened: our queries became questions, which made Google realize that sometimes we don't want a link to a resource, we want answer.
Google has come a long way as an "answer engine", including creating the Knowledge Graph and displaying web results as direct answers, called snippets. For example, if you search for "how does Google search work?", you might get a featured snippet (essentially a matched answer) like this:
This particular example is a bit clunky, but it shows how the algorithm works. Google is looking for snippets of documents that might answer the question. In this case, it is obvious that the excised passage is not really a complete answer.
Now consider the same question that GPT-3 was asked:
How does Google search work?
Google Search works by crawling the web and indexing websites. When a user searches for something, Google looks through its index and returns the most relevant results. It uses a combination of algorithms and other techniques to determine the relevance of the results. These algorithms take into account factors such as the words used in the query, the freshness of the content, the popularity of the website, and the location of the user.
I can criticize this answer, but overall it's pretty good and much more complete than a snippet from a website. The GPT-3 engine basically aggregated multiple sources. It's easy to see how this could make Google nervous.
Trust and the attribution problem
On the surface, GPT-3 (and ChatGPT in particular) can return quite human-like responses. Ethical issues aside (including plagiarism and bias, which I'll discuss briefly later), there are serious practical issues.
Google's "answer" has one major advantage: attribution. We can just go to the source page. This not only gives us more information, but also more context, including details about the source itself. These details can help us trust (or in some cases distrust) the answer.
For example, consider another response returned by GPT-3 for the previous question:
This answer also seems plausible, except that most SEOs would cringe at the mention of keyword density (emphasis added by me), which is widely considered a myth. It's entirely possible that this information came from an entirely different source, but without attribution we cannot validate it.
Although this is a complex problem (machine-generated text does not use fonts the same way a human does, or it may use dozens or hundreds of fonts), there are hybrid approaches that reduce the number of fonts and allow mapping.
Although this feature is in beta and obviously does not work at Google scale, Neeva tries to provide primary sources. I deliberately chose a ridiculous example because the stakes are low. Imagine a question straight out of what Google calls the YMYL (your money, your life) space, such as: B. "Is Bitcoin or Dogecoin better?" :
Read Also : How AI is Changing the Fashion Industry?
White Clover Markets
Even in the relentless news and noise of early 2023, you've probably heard of ChatGPT, the GPT-3 (Tech 3.5) based chatbot that is uncannily capable of simulating human-like responses. For example, consider this question and answer (created by GPT-3):
It's not bad, and it sounds like something a real human would write.
GPT (Generative Pretraining Transformer) is a deep learning system that generates text based on the prediction of speech patterns after being pretrained on a large number of documents. For this article, I'm using OpenAI Playground with the GPT-3 text-davinci-003 model, which can be programmed to produce shorter responses and is more reliable than ChatGPT Beta.
The current models are truly impressive, including the analysis of complex questions:
How would you explain ChatGPT to a 5 year old?
ChatGPT is like a robot friend who can talk to you and answer your questions. It's like having a friend who knows a lot about different things and can help you learn new things.
GPT can even mimic human creativity and make us question our own talents:
Write a slogan for a Martian restaurant run by Edgar Allen Poe-esque lizardmen. "Lizards of Mars, dine in splendor here where dreams unfold and palaces indulge!"
By now you've probably seen dozens of examples. So what does all of this have to do with Google search and our jobs as search engine marketers?
Google and the era of natural language
In the early days of research, search engines were designed to index short words and phrases and were unable to process natural language. If you are looking for a new laptop, you can search for terms such as "laptop" or "buy a laptop". I sometimes joke that if you walk into Best Buy and start yelling "Laptop!" security will probably escort you out. However, humans adapt to early research and simplify our language.
As search became more sophisticated, we learned that we could use more complex queries, and especially with the advent of voice search, we experimented with something closer to natural language. As Google adapted, something interesting happened: our queries became questions, which made Google realize that sometimes we don't want a link to a resource, we want answer.
Google has come a long way as an "answer engine", including creating the Knowledge Graph and displaying web results as direct answers, called snippets. For example, if you search for "how does Google search work?", you might get a featured snippet (essentially a matched answer) like this:
This particular example is a bit clunky, but it shows how the algorithm works. Google is looking for snippets of documents that might answer the question. In this case, it is obvious that the excised passage is not really a complete answer.
Now consider the same question that GPT-3 was asked:
I can criticize this answer, but overall it's pretty good and much more complete than a snippet from a website. The GPT-3 engine basically aggregated multiple sources. It's easy to see how this could make Google nervous.
Trust and the attribution problem
On the surface, GPT-3 (and ChatGPT in particular) can return quite human-like responses. Ethical issues aside (including plagiarism and bias, which I'll discuss briefly later), there are serious practical issues.
Google's "answer" has one major advantage: attribution. We can just go to the source page. This not only gives us more information, but also more context, including details about the source itself. These details can help us trust (or in some cases distrust) the answer.
For example, consider another response returned by GPT-3 for the previous question:
This answer also seems plausible, except that most SEOs would cringe at the mention of keyword density (emphasis added by me), which is widely considered a myth. It's entirely possible that this information came from an entirely different source, but without attribution we cannot validate it.
Although this is a complex problem (machine-generated text does not use fonts the same way a human does, or it may use dozens or hundreds of fonts), there are hybrid approaches that reduce the number of fonts and allow mapping.
Although this feature is in beta and obviously does not work at Google scale, Neeva tries to provide primary sources. I deliberately chose a ridiculous example because the stakes are low. Imagine a question straight out of what Google calls the YMYL (your money, your life) space, such as: B. "Is Bitcoin or Dogecoin better?" :
Read Also : How AI is Changing the Fashion Industry?