Comparing search engines with reasoning engines

- Let's talk about some of the pros and cons of reasoning engines versus search engines. We know that when you enter a query into a traditional search engine like Google or Bing, it searches through its vast index of web content, and then provides you with a ranked list of webpages that contain relevant information. So it's a good resource to use when you'd like to read further about a subject across a collection of different sources, but not necessarily when you want to ask deeper questions. That's because a search engine doesn't truly understand your query. It just matches your keywords to relevant results. - Traditional search is much more linear, right? There's a linear request being asked against a data set to bubble up anything that's remotely similar. There's often some level of ranking that's done to determine, you know, how closely to your question did a answer get, and that's how that rank of, you know, what's on page one of a search engine versus page two. But very rarely does it actually think about what did you mean by what you asked? And this is often a frustration with us as users thinking about a search engine. I'm trying to find something and I'll often be like, "Well, no, that's not I meant to say," and I'll have to adjust it. - A reasoning engine, on the other hand, is designed to understand and interpret human language, which allows it to engage in actual conversations with you. So instead of simply retrieving webpages for you to read and extract information, it provides you with direct relevant responses and can also maintain context and understand the intent behind your questions. - Reasoning engines take in a completely different perspective when you're asking it questions. So what we've seen in the evolution of something like a GPT, for example, is that when we start to ask a questions rather than it linearly looking for a direct answer and ranking everything that's similar to our response, it now is taking in the context of that question related to each other, right? So maybe topics, locations, places. - Sometimes, especially when you're first starting out, it may not be immediately obvious which one to use. But often, just by going through the motions of asking your question, you'll get a sense of which direction you should go. - Honestly, the first thing I will do, sometimes, just out of habit, I will go into a typical search engine, but almost immediately, as soon as I start typing, I'm like, "Maybe I should use a different engine for that." The main reason for doing that though is that often when I'm asking a question, I'm very rarely asking a pointed, very specific, factual answer. I'm usually asking it for something that is more related to help me with this idea. Help me create something, help me ideate. And if I do it in a traditional search engine, I will get examples that then have another step that I have to go into and actually find my answer, which creates more work. So with a reasoning engine, a lot of that is surfaced in a way that I don't have those extra steps, makes that process much faster, but also just much more human-friendly to the creative process. - One thing to watch out for, however, is that even though it sometimes seems that reasoning engines are an all knowing source of truth, that's not necessarily always the case. - One caveat that I think is really important about these reasoning engines is that when they do produce content, it sounds incredibly convincing, that that system has perfectly understood your query, and it has given you back valid results. Consider that maybe, it's not actually giving you the correct result. So everything it gives you at this point, you need to treat with a little bit of caution. Now over time, those reasoning engines might actually get better and better and better, and they will be correct. - One of the disadvantages of reasoning engines that I tend to think about often and we've heard this term like hallucination, right? Like can it be wrong? Now, I often also then think about the amount of things that have been wrong on a typical search engine, right? And that I think the difference though is because the reasoning engine is formatting its response and that response is much more kind of human and natural language-driven. That natural language tends to make us feel like it may be more accurate than a list of resources that we have to kind of determine the accuracy on our own. - So just be mindful of this. Whether you use a search engine, which provides that list of ranked web content, or a reasoning engine, which provides that human-like contextual response, it's up to you to dig deeper to verify and validate your results.