ChatGPT Makes Searching the Web Great Again
As an Italian citizen, I feel obliged to follow our most important music festival: Sanremo Music Festival. However, a typical Sanremo night broadcasts on TV from 8:30PM to 2:00AM. That’s more than 5 hours. Moreover, music performances are continuously interrupted by some boring speeches and sometimes speeches are interrupted by boring music performances.
So, like an average Italian citizen, I don’t watch Sanremo from start to finish, but still I want to know leaderboards and, more generally, what happened that I missed. A great number of websites publish posts talking about Sanremo Music Festival. As you may imagine, they capitalize on the high traffic Sanremo news attracts so they usually bury the interesting information under tons of ads, pop-ups and links to other useless articles.
This year, however, I tried a new approach: each morning I asked ChatGPT — with the web search feature enabled — for the leaderboard of the previous Sanremo night. The result was fantastic. In a few seconds I knew the leaderboard. No more opening Instagram only to be bombarded with irrelevant content, and no more dodging pop-ups and ads while reading a news article on a website designed solely to maximize ad clicks. ChatGPT takes care of surfing the web for you, and then returns the response to your query in a well-formatted, ad-free way.
In general, ChatGPT with web search can mitigate the problem of the modern web: annoying cookie consents, pages overoptimized for SEO where relevant information is only a fraction of the overall content, pop-ups, advertisements, unnecessarily slow to load websites, and so on. Finally, we can focus on the content instead of employing 99% of our brain resources wrestling against all that mess.
So, I advocate that ChatGPT with web search (or any LLM equipped with web search functionality) can be a new way to interact with the web, a sort of new UI for the web. And it’s a UI that respects users’ sense of agency and time more than most websites do. In this sense, ChatGPT can be seen as a digital wellbeing tool.
-
Of course, there are downsides to using LLMs for web searching. I sometimes feel guilty when leveraging this approach, because it definitely requires a lot more electricity than a simple web search, hence contributing to worsening global warming. Moreover, if everyone were to adopt this approach, websites relying on revenue from ads wouldn’t exist anymore.
-
It’d be cool to be able to use this same approach for social media websites, extracting only the useful information we want without our attention being stolen by the multiple attention-capturing dark patterns these websites implement. Unfortunately for us, all social media platforms (Instagram, Facebook, TikTok, X) forbid web scraping and also do not provide APIs to access user content. Of course, they want us to use their apps :)
-
In general, I believe there is value in seeing LLMs as UI for complex content. We’ve already seen this in various applications, e.g., extracting information from a PDF, using natural language to query a Pandas dataframe, etc.
Hey, after three years or so, I finally posted a new blog post. That’s cool. I hope I’ll be more constant now and also I hope I’ll be able to share with you a post I’ve had in mind for about five years now. I’m quite lazy with writing unfortunately.
As always, I’d love to hear what you think about this post. Feel free to drop me an email and we can start talking there!
See you!
← back