Custom GPTs by OpenAI: A Revolutionary Leap or a Passing Fancy?

The last two weeks marked a significant milestone for OpenAI, with the introduction of a powerful new feature to their widely popular ChatGPT: custom GPTs. This addition heralds a new era in AI, enabling users to fundamentally alter the AI chatbot’s functioning. It's not just about tweaking communication styles; it's about feeding specific datasets for a more personalized AI experience. But amidst the buzz, a crucial question arises: Is this a groundbreaking shift in generative AI, or merely a novel quirk?

This development comes during turbulent times for OpenAI. The departure of co-founder and former CEO Sam Altman, along with the demotion and resignation of former president Greg Brockman and the exit of several key team members, have left the company's future in a somewhat precarious position. These changes have stirred speculation and cast a shadow of uncertainty over OpenAI's role in the competitive AI landscape. The success of custom GPTs could well dictate the company's continued relevance.

At its core, the underlying technology of GPT (Generative Pre-trained Transformer) powers the current surge in AI advancements. ChatGPT, now operating on GPT-4, not only processes text but can analyze images and integrate with other OpenAI tools like DALL-E 3. Previously, customizing ChatGPT's responses involved intricate prompting and manual data uploads. Custom GPTs have simplified this process, making personalization more accessible to all paying subscribers.

For initial exploration, users can easily navigate the ChatGPT interface to the custom GPTs section. OpenAI's initial offerings in this domain are cautiously brand-safe, ranging from coloring book creators to tech support aids, and even an AI that mimics Gen Z internet lingo. These custom GPTs are set to eventually link to a public app store, offering a plethora of AI tools tailored to specific needs. Currently, enthusiasts can find these creations on platforms like X and Reddit, with hundreds of applications already available, from tax assistance to sophisticated weather apps.

How to build GPTs

The GPT creation process is super simple: just tell the system you want to build. Here's a step by step guide:

Initiating the Build: You start by telling the system what kind of GPT model you want to build. This is likely done through a user interface or a command line tool where you specify your requirements.
Interacting with the System: Much like chatting with a standard ChatGPT, you communicate with the system, presumably providing it with instructions, preferences, or the type of application you're aiming to create.
Uploading Data: During this interactive process, you have the option to upload data files. These files can contain specific information, training data, or any other relevant content that you want your custom GPT to learn from or reference.
Builder's Role: Once you provide the necessary information and data, the builder takes over. It processes your inputs and the uploaded data to construct the GPT model according to your specifications.
Knowledge Base Creation: The files you upload act as a knowledge base for the GPT. This approach aligns with OpenAI's version of Retrieval Augmentation Generation (RAG). RAG is a method that combines pre-trained language models with a retrievable database, allowing the model to pull in external information when needed.

This process simplifies the creation of custom GPT models, making it accessible to a broader range of users and applications. By allowing users to upload their own data, it opens up possibilities for highly specialized and tailored AI models that can serve niche or specific purposes.

The integration of APIs into custom GPT models adds another layer of functionality and customization. Here's a closer look at how this works:

API Integration: APIs (Application Programming Interfaces) allow different software applications to communicate with each other. Integrating APIs with a GPT model means that the GPT can call external services or data sources as part of its processing and response generation.
OpenAPI Description: To integrate an API, you need to provide an OpenAPI description of the services you want to call. The OpenAPI Specification (formerly known as Swagger Specification) is a standard, language-agnostic interface to RESTful APIs which allows both humans and computers to discover and understand the capabilities of the service without access to source code, additional documentation, or inspection of network traffic.
Automatic Calls by the GPTs: Once integrated, the GPT can automatically call these APIs during its operations. This means that it can fetch data, perform actions, or interact with other software systems autonomously, based on the conversation or the data processing it's performing.
Practical Application - IKANGAI Bot: As an example, you can check out the IKANGAI Bot, which utilizes this feature to retrieve the latest blog posts and return them in a newsletter format.A bot like this could use external APIs for various functionalities, like accessing real-time data, interacting with other software services, or enhancing its conversational capabilities.

The ability to integrate APIs greatly expands the potential applications of custom GPT models, allowing for more dynamic, responsive, and powerful AI tools that can interact with a wide range of digital ecosystems.

Where do I find GPTs?

The absence of an official discovery platform for these apps has led to community-driven efforts for better accessibility. Notable among them is Rowan Cheung's initiative, using a custom GPT to aid in finding new apps. Cheung likens the current state of ChatGPT to the first iPhone and the launch of custom GPTs to the early days of Apple's App Store. He cautions, however, that these tools are not without their limitations and should not replace professional advice in critical areas.

Experimentation with custom GPTs reveals their potential and current limitations. Incorporating specific content, such as a newsletter, can yield results that are almost, but not quite, on point. The real excitement lies in the integration of APIs, allowing custom GPTs to interact with third-party services. This feature has already led to innovative applications like graphic design assistance and website creation, though they are still in their nascent stages and not without issues like slow response times and bugs.

The next App Store

The upcoming launch of the GPT Store by OpenAI represents a significant advancement in the realm of AI and custom GPTs. This platform will fundamentally change how GPTs are shared, discovered, and utilized. Here's an overview of what to expect:

Public Sharing of Custom GPTs: Users already have the ability to create their own GPTs and share them with the public. This marks a shift from private or limited sharing to a more open and collaborative environment where creators can showcase their work to a broader audience.
Launch of the GPT Store: Scheduled for later this month, the GPT Store will feature a curated selection of GPTs created by verified builders. This initiative by OpenAI not only provides a centralized platform for discovering custom GPTs but also adds a layer of credibility and quality assurance through the verification of creators.
Searchability and Leaderboards: Once GPTs are included in the store, they become searchable, allowing users to easily find GPTs that meet their specific needs or interests. Additionally, the introduction of leaderboards will add a competitive and motivational aspect for creators, as their GPTs can climb ranks based on popularity or usage.
Spotlight on Exceptional GPTs: OpenAI plans to highlight the most useful and delightful GPTs in various categories like productivity, education, and entertainment. This spotlight feature not only acknowledges the creativity and effort of the creators but also assists users in discovering high-quality GPTs.
Monetization Opportunities: In the coming months, a monetization model will be introduced, allowing creators to earn revenue based on the usage of their GPTs. This is a significant development, as it opens up new economic opportunities for AI developers and enthusiasts. It incentivizes the creation of innovative, useful, and engaging GPTs, potentially leading to a diverse and rich ecosystem of AI tools.
Community and Ecosystem Growth: The GPT Store is poised to foster a vibrant community of AI creators and users. By providing a platform for sharing, discovering, and monetizing AI creations, OpenAI is encouraging a collaborative and dynamic ecosystem that could lead to rapid advancements and novel applications of AI technology.

Despite the challenges, the AI community is optimistic about the potential of custom GPTs. Their success now hinges on OpenAI's ability to refine and enhance the platform and on developers' creativity in building valuable and desirable applications. As OpenAI navigates its uncertain future, custom GPTs stand as a testament to the company's continued innovation in the field of AI.