What we can envision with Search is a new wave: Search 3.0. This is where people get answers delivered directly, without going through website links. Google, kind of, already has that option with “I’m feeling lucky”, but the source is a single website. If traditional search engines are the librarians who tell us where we can read a particular book or go to a particular genre location in the library, large language model chatbots could be that oversmart person who knows everything and smugly tells you the answer when you ask them something.
Search 2.0 had something called natural language and semantic search, which meant that instead of having to use exact words to find an answer, you could ask questions in a more natural manner, as if you were talking to a friend and the search engine would try to figure out what you’re trying to find, even if you’re inarticulate. The thing about large language models is that they may be emblematic of society.
There have been reports that some LLMs base their training data off platforms like Reddit, which may have a higher male user base, so many ideas and viewpoints may not showcase the whole picture. Another interesting aspect is Section 230 in the US, which absolves Big Tech from liability, because it’s believed that platforms would not be responsible for the opinions of publishers.
Would this change if search engines like Google have direct links to the content? Would search engines then be responsible for any biased or misleading results? Could there be new revenue models with AI-based online advertising? In 2023, we may want information quickly that’s accurate and trustworthy, results that have deeper insights and nuance.
Microsoft founder William Gates opined that the company that wins the AI race would have a huge advantage over competitors, being so powerful that users wouldn’t need search engines ever again. The now-dead Internet Explorer, after its popularity died, used to be opened, only so that people could download Google Chrome or Mozilla Firefox
