From: myfirstmillionpod
Natural language processing (NLP) is emerging as a significant mega-trend, potentially even larger than mobile technology in its impact on software usage [00:50:14]. The core idea behind this trend is to make software truly intuitive by allowing users to express their intent in natural language, rather than learning complex commands or interfaces [00:49:36].
The Vision for Intuitive Software
Historically, software companies have claimed their products are intuitive and easy to use, but the reality is that users must translate their intent into a series of clicks, drags, and swipes to make the software perform the desired action [00:49:05]. The future envisions a world where users can simply state what they want to achieve, and the software understands and executes it [00:49:57].
For instance, in a program like Photoshop, a user could simply say, “remove the background,” instead of navigating menus [00:49:37]. In a business context, such as within Hubspot, a user could ask, “how many new people signed up for our service hub product in the last 90 days?” without needing to build a report manually [00:49:40]. This shift is possible due to advancements in natural language understanding, requiring primarily a translation layer between human intent and software action [00:50:04].
This development could enable billions of people to use software that was previously inaccessible due to the learning curve of clicks and drags [00:50:38]. The business-to-business (B2B) software sector, particularly in areas like business intelligence and reporting, is considered ripe for this transformation [00:50:52]. While consumer-facing applications like Alexa have shown early hints of this, the B2B world is poised for a significant leap [00:50:59].
Current Applications and “Magic Tricks”
The concept of a “magic trick” in software refers to an experience so seamless and effective that it’s hard to unsee, rendering older software obsolete by comparison [00:51:12].
Examples of such magic tricks in natural language processing include:
- GrowthBot: A tool built for marketers to extract marketing data and answer questions from various sources via natural language [00:51:48]. It allowed users to ask questions like “what are the top three keywords that Uber buys on PPC?” or “when was this domain registered?” [00:52:03]. Though a bit early at its launch, it demonstrated the potential for intuitive querying [00:52:27].
- D-Build and GPT-3: An example involving GPT-3 where a user could describe a website they wanted, such as one with two columns, pictures of places to stay on the left, and prices on the right, and the system would generate the website [00:59:16]. This showcased the ability to “code without having to code” by simply describing the desired outcome [00:59:07]. The ability to request actions like adding a PayPal button that only takes a maximum of five dollars and directs payments to a specific address, then seeing the HTML and CSS generated, was considered a “magic trick” [00:59:36].
- GPT-3: This technology is described as the “closest thing to magic” due to its text generation capabilities based on a vast corpus of data [01:00:23]. While practical use cases were still developing, it was recognized as a technology that would go “from a zero to a one very very quickly” [01:01:03]. It embodies the “slowly then all of a sudden” phenomenon [01:01:14].
The Power of Natural Language
This new paradigm is seen as a “next level Google” [00:58:24]. Instead of receiving a list of pages that might contain the answer, users would ask a question and directly receive the answer from the software [00:58:29]. This shift suggests a fundamental change in how users interact with technology, moving towards more direct and conversational interfaces [00:50:00]. It points to a future where software is driven by AI to be truly intuitive.